Open wescale-kb/scorecard/scripts/08-import-portal-scorecards.py. Line 47 maps the WeScale portal's "CVR" column to conversion_rate and stores whatever to_number() returns, and to_number() deletes the % sign and keeps the number — so "1.5%" is stored as 1.5. The older importer, 02-import-scorecard.py, sends the same metric through clean_numeric_value(raw, "percent") in utils.py, which divides by 100 and stores 0.015. Exactly three of the twelve metrics that script imports are percent-type, and all three broke at the same moment: conversion_rate, net_profit_margin and email_pct_revenue. Route all three through clean_numeric_value, divide the rows already stored for weeks 43–47 (calendar 2026-W30 through 2026-W34) by 100, then re-run 03-compute-rankings.py, 04-correlation-analysis.py and 05-generate-dashboards.py — in that order, because the rankings and the correlations are both computed from the broken values.
Two more things in the same pass. Fix the two 480s: your New Designs Launched reads 48, 480, 480, 46, 48 for 2026-W30 through 2026-W34, and those two 480.0 entries put you at rank 2 and rank 1 in the whole cohort on what is almost certainly a stray decimal point. And submit your 2026-W35 row — thirteen other members reported that week and you did not, which is why all eleven metrics on your own benchmarking dashboard currently render blank.
The single number IGL is supposed to steer by is stored in two incompatible units either side of 2026-W30, so every chart that crosses that line is wrong and the correlation table built on top of it points the wrong way.
The seed episode's sharpest sentence is about decentralization: it is not the absence of control, it is the substitution of a single honest number for supervision. Henry Singleton ran 130 profit centres with fewer than 50 people at headquarters and no human-resources department, and could say "why bother them if they were doing their job" only because the Teledyne return — net income plus cash flow, divided by two — told him who was doing the job. He invented that particular formula to kill one specific trick: inflating a reported figure with something that is not cash. Autonomy without that number is not delegation, it is abdication.
IGL runs the same shape — one operator, an automated fleet, one weekly number — and the number is broken in both of the ways Singleton's formula was built to prevent. The unit split makes the honest metric unreadable across the boundary. The two 480s inflate the output metric with something that is not comparable to what the rest of the cohort is counting. Singleton signed every check on his ranch by hand in retirement and called it a discipline: "it's amazing how much you learn about the business — there's a reminder of each event or action behind every check." He removed himself from operations; he never removed himself from the evidence.
The agentic-academy lesson supplies the half of the argument that no business shelf could: an ordinary workflow can be inspected like code and behaves the same way every time, but an automated pipeline's performance has to be measured — and here the measuring instrument is the thing that is broken. Everything resting on it inherits the error: the conversion ladder in the CRO plan, every cohort-median comparison, the significance math.
founders-kb/episodes/431-how-henry-singleton-worked — the seed, fetched 2026-09-05. Key Lesson 2 (the Teledyne return, deliberately built to be un-gameable across dozens of autonomous units), the decentralization note ("autonomy without that metric is abdication"), and Key Lesson 10 (signing every check by hand as a way of staying attached to the evidence).wescale-kb/scorecard/scripts/08-import-portal-scorecards.py — the defect itself. Line 47 is literally ("CVR", "conversion_rate") with no percent handling; to_number() at line 102 strips % and returns the bare float. utils.py line 46 documents that clean_numeric_value "always returns the decimal form for percent-type metrics" — and it is called only from 02-import-scorecard.py, never from this one.wescale-kb/scorecard/wescale-scorecard.db — the split, cohort-wide rather than IGL-only. Current to 2026-W35 (week_number 48); all values are member self-reports, not measurements.agentic-academy-kb/lessons/10-evaluations — captured 2026-08-24. States the governing principle: a normal workflow can be inspected like code and always behaves the same way, but performance that cannot be inspected has to be measured against expected results.wescale-kb/scorecard/scripts/08-import-portal-scorecards.py. Line 47 maps the WeScale portal's "CVR" column to conversion_rate and stores whatever to_number() returns, and to_number() deletes the % sign and keeps the number — so "1.5%" is stored as 1.5. The older importer, 02-import-scorecard.py, sends the same metric through clean_numeric_value(raw, "percent") in utils.py, which divides by 100 and stores 0.015. Exactly three of the twelve metrics that script imports are percent-type, and all three broke at the same moment: conversion_rate, net_profit_margin and email_pct_revenue. Route all three through clean_numeric_value, divide the rows already stored for weeks 43–47 (calendar 2026-W30 through 2026-W34) by 100, then re-run 03-compute-rankings.py, 04-correlation-analysis.py and 05-generate-dashboards.py — in that order, because the rankings and the correlations are both computed from the broken values. Two more things in the same pass. Fix the two 480s: your New Designs Launched reads 48, 480, 480, 46, 48 for 2026-W30 through 2026-W34, and those two 480.0 entries put you at rank 2 and rank 1 in the whole cohort on what is almost certainly a stray decimal point. And submit your 2026-W35 row — thirteen other members reported that week and you did not, which is why all eleven metrics on your own benchmarking dashboard currently render blank. ## Why (the one-liner) The single number IGL is supposed to steer by is stored in two incompatible units either side of 2026-W30, so every chart that crosses that line is wrong and the correlation table built on top of it points the wrong way. ## The insight The seed episode's sharpest sentence is about decentralization: it is not the absence of control, it is the substitution of a single honest number for supervision. Henry Singleton ran 130 profit centres with fewer than 50 people at headquarters and no human-resources department, and could say "why bother them if they were doing their job" only because the Teledyne return — net income plus cash flow, divided by two — told him who was doing the job. He invented that particular formula to kill one specific trick: inflating a reported figure with something that is not cash. Autonomy without that number is not delegation, it is abdication. IGL runs the same shape — one operator, an automated fleet, one weekly number — and the number is broken in both of the ways Singleton's formula was built to prevent. The unit split makes the honest metric unreadable across the boundary. The two 480s inflate the output metric with something that is not comparable to what the rest of the cohort is counting. Singleton signed every check on his ranch by hand in retirement and called it a discipline: "it's amazing how much you learn about the business — there's a reminder of each event or action behind every check." He removed himself from operations; he never removed himself from the evidence. The agentic-academy lesson supplies the half of the argument that no business shelf could: an ordinary workflow can be inspected like code and behaves the same way every time, but an automated pipeline's performance has to be measured — and here the measuring instrument is the thing that is broken. Everything resting on it inherits the error: the conversion ladder in the CRO plan, every cohort-median comparison, the significance math. ## Evidence trail - founders-kb/episodes/431-how-henry-singleton-worked — the seed, fetched 2026-09-05. Key Lesson 2 (the Teledyne return, deliberately built to be un-gameable across dozens of autonomous units), the decentralization note ("autonomy without that metric is abdication"), and Key Lesson 10 (signing every check by hand as a way of staying attached to the evidence). - wescale-kb/scorecard/scripts/08-import-portal-scorecards.py — the defect itself. Line 47 is literally ("CVR", "conversion_rate") with no percent handling; to_number() at line 102 strips % and returns the bare float. utils.py line 46 documents that clean_numeric_value "always returns the decimal form for percent-type metrics" — and it is called only from 02-import-scorecard.py, never from this one. - wescale-kb/scorecard/wescale-scorecard.db — the split, cohort-wide rather than IGL-only. Current to 2026-W35 (week_number 48); all values are member self-reports, not measurements. - agentic-academy-kb/lessons/10-evaluations — captured 2026-08-24. States the governing principle: a normal workflow can be inspected like code and always behaves the same way, but performance that cannot be inspected has to be measured against expected results. ## Worked examples (the full result set, not a sample) All twelve metrics script 08 imports, checked for the break at the week-42/43 boundary. Three are percent-type. All three break; the nine non-percent metrics are unaffected. | metric | unit | week 42 cohort max | week 43 cohort max | broken? | |---|---|---|---|---| | conversion_rate | percent | 0.0827 | 4.6 | yes | | net_profit_margin | percent | 0.29 | 33.5 | yes | | email_pct_revenue | percent | 0.6157 | 64.7 | yes | | new_ad_creatives | count | 257 | 334 | no | | fb_ad_spend | currency | 53,811 | 36,964 | no | | campaigns_sent | count | 6 | 40 | no | | fb_roas | ratio | 3.28 | 2.87 | no | | new_designs | count | 95 | 350 | no | | total_sales | currency | 127,155.78 | 129,203 | no | | fb_cpc | currency | 1.15 | 1.10 | no | | aov | currency | 77.34 | 109 | no | | mer | ratio | 5.52 | 1.15 | no | IGL's own stored strings, weeks 2026-W27 → 2026-W34 (the weeks table flips schema_version from V3B to portal-v1 at week_number 43 = 2026-W30): | metric | W27 | W28 | W29 | W30 | W31 | W32 | W33 | W34 | |---|---|---|---|---|---|---|---|---| | conversion_rate | 0.0206 | 0.0242 | 0.0187 | 1.6% | 1.7% | 2.0% | 1.8% | 1.5% | | net_profit_margin | -0.3964 | -0.3216 | -0.4099 | -44.0% | -33.2% | -51.0% | -29.0% | -51.3% | | email_pct_revenue | 0.1717 | 0.1975 | 0.1240 | 23.1% | 20.2% | 21.6% | 20.9% | 13.5% | What the dashboard does with those. 05-generate-dashboards.py renders percent values as (val * 100).toFixed(1) + '%'. So a stored 1.5 prints as 150.0% conversion and a stored -44.0 prints as -4400.0% net margin. New Designs Launched, 2026-W30 → W34: 48, 480.0, 480.0, 46, 48 — and the rankings table records percentile 0.980 (rank 2 of 50) and 1.000 (rank 1 of 59) for the two 480 weeks. 2026-W35: 13 members have observations; IGL has zero. igl-benchmarking.html, rebuilt 2026-09-05, carries "igl_val": null for all 11 metrics in its gap table. ## Assay verdicts - novelty — PASS (4/5): no prior art. Nugget #023 argued about which conversion figure to trust; nugget #013 found a different scorecard defect (all-zero weeks not filtered out of the averages). This is a third, distinct, mechanically verified defect at a named line. - grounding — PASS (4/5): every claim reproduced from disk. One overstatement caught and corrected here: the benchmarking dashboard holds 11 metrics, not twenty-one. - actionability — PASS (4/5): named script, named line, named function, named metrics, named weeks, plus the missing-row task. Held at 4 because fixing the ruler does not by itself change a design, ad or email decision this week. ## Operator verification and extension Every claim above was independently re-derived from disk before publication, and three corrections were made to the hunt's version. 1. The metric list was guessed and is now measured. The hunt named conversion_rate, popup_rate, email_pct_revenue, cogs_pct and overhead_pct. Two of those five — popup_rate and cogs_pct/overhead_pct — are not among the twelve labels script 08 imports at all. Running the check across every imported metric returns exactly three affected: conversion_rate, net_profit_margin, email_pct_revenue. net_profit_margin is the one the hunt missed and the one that matters most — IGL's margin renders as -4400% rather than -44%. 2. The correlation table is corrupted, and that is the part with a wrong answer in it rather than a wrong picture. 04-correlation-analysis.py's get_member_timeseries selects every week for a member with no boundary filter, and 60 members have conversion data on both sides of the break, so 1,218 stored correlation rows involving conversion_rate were computed over series containing a 100x step. Demonstrated, not asserted: correlating each member's stored conversion series against week number gives r = +0.584 (member 29), +0.588 (member 26), +0.578 (member 18); repairing the units first gives -0.197, -0.131 and +0.156. Three members, three sign or magnitude reversals. The stored table currently says conversion rates are rising across the cohort. They are not. 3. Within a single week the rankings are still sound, and the card should not overclaim. Every portal-era member is stored in the same (wrong) unit, so IGL's conversion rank and percentile inside weeks 43–47 are unaffected. What is broken is anything crossing the boundary: time-series charts, the correlations, and any pooled multi-week average. 4. Where the standing conversion problem actually stands. In 2026-W34, 53 members reported conversion; the cohort median is 2.6% and IGL is 1.5% — 42% below median, against the "~27% below" figure the standing problem statement carries. One week, self-reported, so this is a reading rather than a verdict — but it is a worse reading, not a better one, and it is why fixing the instrument comes before arguing about the number. Thread: ep-431's Teledyne return and "autonomy without the metric is abdication" → the raw strings in wescale-scorecard.db → the V3B → portal-v1 schema flip at week 43 → 08-import-portal-scorecards.py line 47 missing the normalisation → igl-benchmarking.html rendering igl_val: null → agentic-academy lesson 10 on measuring what cannot be inspected.Before you mine a folder in the competitor best-seller corpus (competitors-kb/proven-winners/), count how many different stores are actually in it. One command does it: ls <folder> | sed 's/__.*//' | sort | uniq -c | sort -rn. If the top store supplies more than a third of the files, that folder is one shopkeeper's shelf, not proof that the theme sells — mine it for mechanism (how the joke is built, how the type is set) but never as "this theme is validated." Write that rule into competitors-kb/miner-methodology.md at the Discover step, and fix line 30 of competitors-kb/proven-winners/README.md, which states that FaithWise + Co supplies 5 of the 29 faith-gardening entries. It supplies 25.
The corpus captures each store's top ~25 best-sellers, so a store that only sells one theme dumps its entire top 25 into one folder — which makes the narrowest folders look like the strongest demand signals when they are the weakest.
Daniel Ludwig had been burned twice by experts' charts. Specialists said a loaded 60,000-ton carrier could come down a river without grounding. Geologists said Grand Bahama coral would hold supertankers. Both were wrong. So before he built in Panama he bought a nickel bolt and a twenty-cent ball of string, knotted it every six feet, and spent a day in a rented motorboat re-taking every sounding on the chart himself. The doctrine is not distrust of experts. It is that a chart you are about to spend money against gets verified personally, cheaply, once.
The competitor best-seller corpus is IGL's chart for design mining. Sounding it takes one shell command, and the result inverts how the folders should be read. faith-gardening — the folder closest to IGL's own brand — is 25 of 29 files from a single store, faithwiseco.com. Not because faith-garden demand concentrates there, but because the capture rule takes each store's top ~25 and FaithWise + Co is a 100%-faith brand, so all 25 of its rows land in one folder while every general garden store contributes one or two faith designs out of its own top 25. The folder's shape is an artifact of store specialization, not of what buyers want. The trustworthy folders are the wide ones.
Why no single knowledge base reaches this: the corpus README states the concentration wrongly and would never catch itself. Kelli Sanders' member profile in the mastermind base explains why FaithWise's top rows sell, which makes the folder look more legitimate, not less. And the corpus carries no sales data at all — manifest.json has rank and price and no sales or revenue field — so "proven" means shelf position scraped on 2026-07-12, roughly eight weeks stale, never a measurement. The Ludwig frame is what turns three separate documents into one instruction: sound the chart before you build on it, and the sounding costs nothing.
founders-kb/episodes/068-daniel-ludwig-the-invisible-billionaire — primary, the seed. Ludwig's verification doctrine after two expert failures: the bolt-and-string harbor sounding in Panama, done personally at near-billionaire wealth because the chart was about to be spent against.competitors-kb/proven-winners/README — theme-aggregate, the only one used. Built 2026-07-12, no Status: field, ~8 weeks old. States the capture rule ("top ~25 best-sellers per store") that creates the artifact, and carries the false claim on line 30 that FaithWise supplies 5 of the 29.competitors-kb/proven-winners/manifest.json — primary. Field list contains rank and price and no sales or revenue field, which is what makes "proven" mean shelf position, not demand.wescale-kb/members/kelli-sanders — primary. Owner of faithwiseco.com; a faith-only brand with no other thematic overlay, which is exactly why her whole top 25 lands in one folder. Her figures ($42,029 in week 41; $482,128.42 year-to-date, self-reported 2026-05-18) are cohort-scorecard REFERENCE numbers for her store — not IGL numbers and not measurements of design demand.igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox export dated 2025-12-11 covering 2024-08-31 to 2025-12-09, no Status: field, ~9 months old. Across 537 active reviews: "sanctuary" 0, "sacred" 1, "spiritual" 3. IGL's buyers do not speak in the explicit-faith register.G:\iris\igl-ops-sage\data\igl.db — measurement, outside the knowledge bases. fact_product_sales_daily joined to bridge_collection_products, trailing 12 months to 2026-09-04. Supplies every figure in "The queryable measurement nobody ran" above.competitors-kb/proven-winners/), count how many different stores are actually in it. One command does it: ls <folder> | sed 's/__.*//' | sort | uniq -c | sort -rn. If the top store supplies more than a third of the files, that folder is one shopkeeper's shelf, not proof that the theme sells — mine it for mechanism (how the joke is built, how the type is set) but never as "this theme is validated." Write that rule into competitors-kb/miner-methodology.md at the Discover step, and fix line 30 of competitors-kb/proven-winners/README.md, which states that FaithWise + Co supplies 5 of the 29 faith-gardening entries. It supplies 25. ## Why (the one-liner) The corpus captures each store's top ~25 best-sellers, so a store that only sells one theme dumps its entire top 25 into one folder — which makes the narrowest folders look like the strongest demand signals when they are the weakest. ## The insight Daniel Ludwig had been burned twice by experts' charts. Specialists said a loaded 60,000-ton carrier could come down a river without grounding. Geologists said Grand Bahama coral would hold supertankers. Both were wrong. So before he built in Panama he bought a nickel bolt and a twenty-cent ball of string, knotted it every six feet, and spent a day in a rented motorboat re-taking every sounding on the chart himself. The doctrine is not distrust of experts. It is that a chart you are about to spend money against gets verified personally, cheaply, once. The competitor best-seller corpus is IGL's chart for design mining. Sounding it takes one shell command, and the result inverts how the folders should be read. faith-gardening — the folder closest to IGL's own brand — is 25 of 29 files from a single store, faithwiseco.com. Not because faith-garden demand concentrates there, but because the capture rule takes each store's top ~25 and FaithWise + Co is a 100%-faith brand, so all 25 of its rows land in one folder while every general garden store contributes one or two faith designs out of its own top 25. The folder's shape is an artifact of store specialization, not of what buyers want. The trustworthy folders are the wide ones. Why no single knowledge base reaches this: the corpus README states the concentration wrongly and would never catch itself. Kelli Sanders' member profile in the mastermind base explains why FaithWise's top rows sell, which makes the folder look more legitimate, not less. And the corpus carries no sales data at all — manifest.json has rank and price and no sales or revenue field — so "proven" means shelf position scraped on 2026-07-12, roughly eight weeks stale, never a measurement. The Ludwig frame is what turns three separate documents into one instruction: sound the chart before you build on it, and the sounding costs nothing. ## The full sounding (operator-run, all 17 theme folders) The hunt surfaced eight folders. Here is every one, re-counted from disk on 2026-09-05, image files only, _pool/ and _filtered/ excluded. | Folder | Files | Stores | Top store | n | Share | Verdict | |---|---:|---:|---|---:|---:|---| | parallel/auto-vintage | 26 | 2 | vintageaircooledrides-com | 25 | 96.2% | one shopkeeper's shelf | | parallel/ocean-coastal | 26 | 2 | dolcereverie-com | 25 | 96.2% | one shopkeeper's shelf | | parallel/spiritual-empower | 18 | 3 | lotus-moon-co | 16 | 88.9% | one shopkeeper's shelf | | garden/faith-gardening | 29 | 5 | faithwiseco-com | 25 | 86.2% | one shopkeeper's shelf | | garden/moss-mushroom | 17 | 4 | allbutfaded-io | 14 | 82.4% | one shopkeeper's shelf | | parallel/outdoor-adventure | 17 | 2 | summitsoul-shop | 12 | 70.6% | one shopkeeper's shelf | | parallel/food-italian | 36 | 3 | saucetown-co | 25 | 69.4% | one shopkeeper's shelf | | parallel/horse-heritage | 28 | 3 | trianglethreads-shop | 19 | 67.9% | one shopkeeper's shelf | | garden/slow-living | 2 | 2 | gardengirltees-com | 1 | 50.0% | too small to read either way | | garden/garden-wildlife | 71 | 15 | trailkeepers-store | 25 | 35.2% | fails, and the hunt missed it | | misc/misc-nongarden | 23 | 7 | wildpetal-store | 8 | 34.8% | fails, marginally | | garden/nature-outdoor | 57 | 6 | vanerianstreetwear-com | 18 | 31.6% | passes | | garden/veg-homestead | 60 | 12 | americanfarmcompany-com | 16 | 26.7% | passes | | garden/plant-humor | 57 | 11 | botanicalwear-com | 14 | 24.6% | passes | | garden/wildflowers-floral | 77 | 13 | flowerwise-dk | 15 | 19.5% | passes | | garden/pollinators | 43 | 12 | plantparent-co | 8 | 18.6% | passes | | garden/general-garden | 32 | 12 | gardenistee-com | 5 | 15.6% | passes | Three operator corrections to the hunt's own worked examples. 1. garden/garden-wildlife fails the rule at 35.2% and the hunt never counted it. This is the biggest miss, because garden-wildlife is IGL's own historical winner lane — the bird-feeder family that nuggets #026, #029 and #030 are all about. 25 of its 71 files come from one store, trailkeepers-store. It is a wide folder in absolute terms (15 stores) but one merchant still sets more than a third of it, so the "birds are proven" read needs the same caution as the faith read. 2. general-garden's biggest contributor is gardenistee-com with 5, not thecottagemeadow-com. The verdict is unchanged (5 of 32 = 15.6%, a genuine multi-store signal); the store name in the hunt's example was wrong. 3. The rule needs a floor. It fires on garden/slow-living, which holds two files. A folder under roughly ten files carries no signal in either direction — add "and the folder holds at least 10 files" to the rule, or it will keep flagging noise. ## The queryable measurement nobody ran The nugget above is about the competitor corpus, so it does not need IGL sales data to be true. But the obvious next question — does IGL's own faith lane actually sell? — has an answer sitting in the ops database (G:\iris\igl-ops-sage\data\igl.db, fact_product_sales_daily joined to bridge_collection_products), and no knowledge base holds it. Run over the trailing twelve months (2025-09-05 to 2026-09-04), the two faith collections Sacred Garden and Rooted in Faith hold 37 products, of which 10 sold nothing at all, and the rest sold 155 units for $5,006.47 — against 5,328 units and $176,999.14 for everything else. The faith lane is 2.8% of units and 2.8% of revenue, at $185 of revenue per selling product versus $308 elsewhere. And inside those collections, the register matters more than the shelf. The top five sellers are Beautiful in Their Differences (28 units), Dirt Therapist (24), Beware of Gardens (13), Find Peace in Your Garden (12) and Beware of Seed Savers (10) — 87 of the 155 units, and four of the five carry no explicit-faith word at all. Every plainly devotional title sits at the bottom: Prayer Takes Root 5, Rooted in His Word 3, Divine Partnership 3, Planted by Living Water 2, Sacred Truths 1, Cultivating Sacred Ground 0, Heaven's Orchard 0, Sacred Orchard 0. Counting titles by register, 114 of 155 units (74%) came from designs with no explicit-faith word in the name. This is a measurement, not a quotation, and it agrees with the customer-voice evidence in the thread: across 537 active reviews in IGL's own December 2025 export, "sanctuary" appears 0 times, "sacred" once, "spiritual" three times. Read together with the folder count, the same warning lands twice: the explicit-faith register that fills 86% of the faith-gardening folder is one competitor's business model, and IGL's own buyers have not bought it. That is directly relevant to the standing faith-transition question, and it is the strongest argument for keeping the corpus honest — the folder that looked most like IGL's future is the one folder that was never a demand signal. One thing this measurement does not say. It measures what IGL has sold, through the channels and ad spend IGL actually ran. It does not measure what IGL could sell if the faith lane were funded and advertised the way the humor lane was — which is exactly the trap nugget #029 caught two nights ago, where a lane's collapse turned out to be a defunding rather than a demand failure. Treat 2.8% as a statement about the current state, not a verdict on the faith transition. ## Evidence trail - founders-kb/episodes/068-daniel-ludwig-the-invisible-billionaire — primary, the seed. Ludwig's verification doctrine after two expert failures: the bolt-and-string harbor sounding in Panama, done personally at near-billionaire wealth because the chart was about to be spent against. - competitors-kb/proven-winners/README — theme-aggregate, the only one used. Built 2026-07-12, no Status: field, ~8 weeks old. States the capture rule ("top ~25 best-sellers per store") that creates the artifact, and carries the false claim on line 30 that FaithWise supplies 5 of the 29. - competitors-kb/proven-winners/manifest.json — primary. Field list contains rank and price and no sales or revenue field, which is what makes "proven" mean shelf position, not demand. - wescale-kb/members/kelli-sanders — primary. Owner of faithwiseco.com; a faith-only brand with no other thematic overlay, which is exactly why her whole top 25 lands in one folder. Her figures ($42,029 in week 41; $482,128.42 year-to-date, self-reported 2026-05-18) are cohort-scorecard REFERENCE numbers for her store — not IGL numbers and not measurements of design demand. - igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox export dated 2025-12-11 covering 2024-08-31 to 2025-12-09, no Status: field, ~9 months old. Across 537 active reviews: "sanctuary" 0, "sacred" 1, "spiritual" 3. IGL's buyers do not speak in the explicit-faith register. - G:\iris\igl-ops-sage\data\igl.db — measurement, outside the knowledge bases. fact_product_sales_daily joined to bridge_collection_products, trailing 12 months to 2026-09-04. Supplies every figure in "The queryable measurement nobody ran" above. ## Assay verdicts - novelty — pass, 4. Re-derived the count independently and confirmed README line 30 is wrong. Prior art sweep found nothing on per-folder store concentration: the three shipped corpus critiques attack different defects (#005 missing listing dates, #009 winner-only sampling, #010 IGL's own designs counted as rivals). Operator re-check: "store concentration", "one shopkeeper" and "25 of 29" return zero hits across all 13 knowledge bases and all 30 prior nuggets. Every prior use of the word "concentration" in the vault refers to IGL's own sales concentration by design, a different thing. - grounding — pass, 5. Every number re-derived from disk and matched exactly. Operator re-check: all 17 folders recounted; the eight cited examples are all correct, and see the three corrections above for what the hunt missed rather than got wrong. - actionability — pass, 4. Changes a named decision (which folders count as validated demand at the Discover step), and the move is one shell command, a stated threshold, a specific file and line to correct, and a specific folder to re-classify. Thread: episode 068, Ludwig sounds the Panama harbor himself → the corpus README claims FaithWise is 5 of 29 → the actual count is 25 of 29 → Kelli Sanders' profile explains why (faith-only store, whole top 25 lands in one folder) → ran the count across all 17 theme folders → IGL's customer reviews show its buyers never use that register → the ops database shows the faith lane is 2.8% of units.Create long-sleeve listings for Quietly Cheering for Moss (70 units in the last 90 days, IGL's #1) and Listen to the Birds (67 units, #2) this week — neither has one, and IGL's long-sleeve catalog is 48 listings built around designs that were selling in mid-2025, not the ones selling now. Then put Excuse Me, Bird Feeder's Empty Long Sleeve — which already exists and has quietly earned $9,507.66 on 240 units — into the next Klaviyo send and the September Meta rotation, written as a seasonal changeover, not as a new product. Pull the push at the end of February: long sleeves go from 11.3% of units in December to 0.6–1.3% every month from March through August. Leave Template #19 in igl-brand-kb/voice/email-templates.md exactly as it is — see correction 3 below.
Claude Hopkins multiplied Bissell's business by offering the same sweeper in twelve wood finishes he chose himself, and IGL's own sales database says the same design in a different garment is worth about 21% more per unit — but only in the four months that start now.
The second Bissell breakthrough in Hopkins' autobiography is quieter than the Christmas one and more useful here. He knew wood, so he convinced Bissell to release twelve finishes and let the buyer pick: "It multiplied the use of carpet sweepers and it gave Bissell sweepers the practical monopoly which they maintain to this day" (founders-kb/episodes/170-claude-hopkins-my-life-in-advertising.md:628). The mechanism stripped of its era: the same design in more variants is nearly free growth. The part everyone skips is that Hopkins picked the variants himself. He did not survey for them.
The mastermind proved the second half of that sentence the hard way, and the receipts sit in a raw chat log rather than in any framework file. Call #11's chat has Sal Presti — "I had tons of requests for vnecks... vnecks account for 2% of the sales" — Eric Corry refusing to list them at all, and Peter Heckman generalising it: "Sometimes what the customer asks for is not what they really want." Against that, three members independently put long sleeves at 10%, 30–40%, and "much, much better."
IGL's own database says the cohort is right about the ranking and wrong about the reason, and the reason is the whole finding. Store-wide, long sleeves run 8.2% of units in September, 9.8% in November and 11.3% in December, then fall to 0.6–1.3% every month from March through August. That is a cold-weather curve, not the sun-protection story a single review suggests. It also means the window is a calendar event Tim can schedule, and it opens the month this card was written. The economics inside it beat the tee: $39.83 per long-sleeve unit against $32.82 per tee over the trailing year, same design, same customer.
And here is the gap no shelf holds alone. The 48 long-sleeve listings are tomato-heavy and were built for the designs winning a year ago. IGL's two best-selling shirts of the last 90 days — Quietly Cheering for Moss and Listen to the Birds — have no long-sleeve version at all, heading into the only season where long sleeves matter. The knowledge bases could see Hopkins' mechanism and the cohort's ranking; only the sales database could see that the variant catalog had drifted off the winners.
founders-kb/episodes/170-claude-hopkins-my-life-in-advertising — line 628, the twelve Bissell wood finishes and the "practical monopoly" result; and the fact Hopkins chose them rather than polling. Primary; biography episode.wescale-kb/raw-data/chats/call-11 — lines 208–276, the full v-neck-versus-long-sleeve exchange: Sal Presti 2%, Bonnie 5% v-neck / 10% long sleeve at Black Friday, Stace 5% v-neck / 30–40% long sleeve, Shona Todd 32% v-neck, Eric Corry refusing, Peter Heckman's line. Primary raw chat; the enriched call file dates it "Late December 2025 / Early January 2026." Members' self-reported shares of their own stores, not audited, not IGL's.igl-brand-kb/community/customer-reviews-q4-2025 — export dated 11 December 2025, no declared status. Review-rank table: the bird-feeder tee at 215 of 551 reviews, its long-sleeve sibling at 9. Jennifer Fox's long-sleeve review naming sun protection. Review counts, not units.igl-brand-kb/voice/email-templates — Template #19 (line 924, added 2026-05-30), classified REPLY_ONLY: "We add designs to V-neck on request, so treat every inquiry as both a sale opportunity and a roadmap signal."wescale-kb/calls/call-31 — dated 2026-05-28, Lesson 2: "iterate on designs first, always — if a design takes off, the next 10 ads are variations of that design concept." The seed. It never considers garment variants of a proven design.mylibrary-kb/books/animal-vegetable-miracle — the seasonal-calendar frame: demand that is a season rather than a preference is scheduled, not tested.igl-brand-kb/voice/email-templates.md exactly as it is — see correction 3 below. ## Why (the one-liner) Claude Hopkins multiplied Bissell's business by offering the same sweeper in twelve wood finishes he chose himself, and IGL's own sales database says the same design in a different garment is worth about 21% more per unit — but only in the four months that start now. ## The insight The second Bissell breakthrough in Hopkins' autobiography is quieter than the Christmas one and more useful here. He knew wood, so he convinced Bissell to release twelve finishes and let the buyer pick: "It multiplied the use of carpet sweepers and it gave Bissell sweepers the practical monopoly which they maintain to this day" (founders-kb/episodes/170-claude-hopkins-my-life-in-advertising.md:628). The mechanism stripped of its era: the same design in more variants is nearly free growth. The part everyone skips is that Hopkins picked the variants himself. He did not survey for them. The mastermind proved the second half of that sentence the hard way, and the receipts sit in a raw chat log rather than in any framework file. Call #11's chat has Sal Presti — "I had tons of requests for vnecks... vnecks account for 2% of the sales" — Eric Corry refusing to list them at all, and Peter Heckman generalising it: "Sometimes what the customer asks for is not what they really want." Against that, three members independently put long sleeves at 10%, 30–40%, and "much, much better." IGL's own database says the cohort is right about the ranking and wrong about the reason, and the reason is the whole finding. Store-wide, long sleeves run 8.2% of units in September, 9.8% in November and 11.3% in December, then fall to 0.6–1.3% every month from March through August. That is a cold-weather curve, not the sun-protection story a single review suggests. It also means the window is a calendar event Tim can schedule, and it opens the month this card was written. The economics inside it beat the tee: $39.83 per long-sleeve unit against $32.82 per tee over the trailing year, same design, same customer. And here is the gap no shelf holds alone. The 48 long-sleeve listings are tomato-heavy and were built for the designs winning a year ago. IGL's two best-selling shirts of the last 90 days — Quietly Cheering for Moss and Listen to the Birds — have no long-sleeve version at all, heading into the only season where long sleeves matter. The knowledge bases could see Hopkins' mechanism and the cohort's ranking; only the sales database could see that the variant catalog had drifted off the winners. ## Worked examples — proven designs with no long-sleeve listing Trailing 90 days to 2026-09-03, active tee listings, checked against all 48 long-sleeve listings: | Units (90d) | Revenue | Last sold | Design | |---|---|---|---| | 70 | $2,312.17 | 2026-08-28 | Quietly Cheering for Moss | | 67 | $2,272.82 | 2026-09-02 | Listen to the Birds | | 26 | $856.70 | 2026-08-25 | Johnny Appleseed | | 26 | $856.70 | 2026-08-25 | Fruit of Spirit Seed Packets | | 25 | $833.64 | 2026-09-03 | Quietly Rooting for Moss | | 23 | $801.62 | 2026-08-26 | My Chickens Are My Therapists | | 22 | $738.03 | 2026-08-25 | Know Your Tomatoes | The three top-16 designs that do already have a long-sleeve sibling are Life's Too Short Tomatoes, Excuse Me Bird Feeder's Empty, and Quietly Plotting Tomato Harvest. ## Evidence trail - founders-kb/episodes/170-claude-hopkins-my-life-in-advertising — line 628, the twelve Bissell wood finishes and the "practical monopoly" result; and the fact Hopkins chose them rather than polling. Primary; biography episode. - wescale-kb/raw-data/chats/call-11 — lines 208–276, the full v-neck-versus-long-sleeve exchange: Sal Presti 2%, Bonnie 5% v-neck / 10% long sleeve at Black Friday, Stace 5% v-neck / 30–40% long sleeve, Shona Todd 32% v-neck, Eric Corry refusing, Peter Heckman's line. Primary raw chat; the enriched call file dates it "Late December 2025 / Early January 2026." Members' self-reported shares of their own stores, not audited, not IGL's. - igl-brand-kb/community/customer-reviews-q4-2025 — export dated 11 December 2025, no declared status. Review-rank table: the bird-feeder tee at 215 of 551 reviews, its long-sleeve sibling at 9. Jennifer Fox's long-sleeve review naming sun protection. Review counts, not units. - igl-brand-kb/voice/email-templates — Template #19 (line 924, added 2026-05-30), classified REPLY_ONLY: "We add designs to V-neck on request, so treat every inquiry as both a sale opportunity and a roadmap signal." - wescale-kb/calls/call-31 — dated 2026-05-28, Lesson 2: "iterate on designs first, always — if a design takes off, the next 10 ads are variations of that design concept." The seed. It never considers garment variants of a proven design. - mylibrary-kb/books/animal-vegetable-miracle — the seasonal-calendar frame: demand that is a season rather than a preference is scheduled, not tested. ## Operator corrections (five, applied before shipping) 1. The "unpushed listing with 9 reviews" premise is wrong, and it is a review-count artifact. The candidate read rank 5 / 9 reviews in a nine-month-old Loox export as neglect. fact_product_sales_daily says the bird-feeder long sleeve has sold 240 units for $9,507.66 since 2025-05-08 and last sold 2026-08-26. It is the design family's second-biggest earner, behind only the tee ($125,720 on 4,018 units). The move survives; "nobody is pushing it" does not. 2. "Sun protection, not warmth" is contradicted by IGL's own seasonality. The candidate built a whole reframe — long sleeves are a different product for gardeners because they work outdoors in sun — on one review. The monthly curve is unambiguously cold-weather. The sun angle may still sell a shirt; it is not why the category moves, and copy built on it would miss the reason people actually buy in November. 3. Killing the v-neck pipeline is the wrong call and the data argues the other way. IGL's v-necks have earned $7,394 across 36 selling products, and in every month from March to August 2026 v-necks outsold long sleeves (August: 16 units to 6). Sal's 2% carries a qualifier the candidate dropped — "That 2% is for that store" — and Bonnie's own note in the same chat says "Grandmas are requesting v-necks in my store… Comments: crewnecks too tight and make their necks look short." IGL's core buyer is 55–75 and 68–70% female. Template #19 stays as written. 4. The economic argument the candidate never made is the strongest one. Long sleeve earns $39.83 per unit against the tee's $32.82 over the trailing 365 days — roughly 21% more on artwork that already exists and advertising already paid for. 5. The concentration risk worth naming. In December 2025, 100 of IGL's 130 long-sleeve units were the bird feeder — 77% of the season on one design; across September–December 2025 it was 160 of ~223. That is why the two missing listings above are the move rather than a nice-to-have. ## Assay verdicts - novelty — PASS (4): verified unpublished; the call-11 chat log is unmined and no prior nugget touches garment-variant seasonality. Operator re-check: "Essential Long Sleeves", "$39.83" and "11.3%" return zero hits across all thirteen knowledge bases and all 29 prior nuggets. - grounding — PASS (5): judged the strongest-grounded candidate in the batch, and every chat figure and attribution was re-verified by the operator against lines 208–276. Two of its premises were nonetheless wrong (corrections 1 and 2) — both claims about IGL's business state, both resolvable only outside the knowledge bases. - actionability — PASS (5): changes a standing written policy and a dated calendar decision. The operator corrections narrowed it: the policy change was dropped, the calendar move sharpened, and two named listings replaced a general instruction. Thread: call-31 lesson 2 "iterate the winner's design first" → ep-170 Hopkins' twelve Bissell wood finishes → raw-data/chats/call-11 the numbers members actually posted → customer-reviews-q4-2025 (215-review winner, its 9-review long sleeve) → email-templates Template #19 → (operator) fact_product_sales_daily for the seasonality, the price gap, and the two missing listingsBefore you set Q4 budget, open igl-brand-kb/audience/profiles/05-carol-garden-witness.md section 8 and replace the line "fading fast (11.9% of last-120d; lane −22.5% in 90 days vs store −12.4%)" with the same-weeks-last-year pair and the ad spend printed beside it: June–August 2025 the bird-feeder lane sold $23,912 on $6,592 of Meta spend, which was 22.9% of the whole ad account; June–August 2026 it sold $1,129 on $68, which was 0.24%. Then, in Meta Ads Manager before Thanksgiving (26 Nov 2026), you un-pause the three existing bird-feeder ads that last delivered on 2026-01-15 rather than building new creative for that lane — PostID_Ad_001_Excuse Me in campaign 40_MANUAL_CBO_Scaling_WW_11.13..2025_Dup, and its two copies in 40_MANUAL_CBO_Scaling_US_10.12.2025 and 10_CBO_Scaling_US_SQUIRRELS_11.13.2025 — because they are PostID ads carrying the shirt's accumulated 1,232 ad comments, and they ran at 2.00x, 1.83x and 2.22x Meta-reported return in the last season they were funded. Finally, add one column to the Tuesday WeScale scorecard you already submit: each lane's share of Meta spend printed next to its revenue, so no lane is ever again recorded as fading in a month nobody bought it impressions.
Spitznagel's rule that a measurement window shorter than the payoff cycle turns a working strategy into an apparent loss led to the year-over-year test the dossiers never ran — and running it showed the opposite of what the dossier concluded: the lane that is 49.5% of all-time revenue was not tested and found wanting, it was defunded and then written up as dying.
The candidate that won tonight's assay proposed the right instrument and predicted the wrong reading. Its claim was that IGL's five WHO dossiers judge each buyer lane on a 90- or 120-day window, that this is shorter than the annual cycle gardeners live on, and that Carol's lane is therefore being recorded as decaying during her off-season — her own dossier says "her season OPENS at Thanksgiving." So: measure the same weeks last year instead. That instrument is correct and the dossiers genuinely lack it.
Run it and the verdict inverts. The same summer window a year apart is not a gentle seasonal dip: the bird-feeder lane went from $23,912 to $1,129, −95.3%, far worse than the −22.5% the dossier reports. Year-over-year would have confirmed the decay verdict, not overturned it. What overturns it is the number that sits in neither the dossier nor any knowledge base: over those same months the lane's share of the Meta ad account fell from 22.9% to 0.24% while total account spend stayed flat (−3%, $28,803 to $27,926). Sales did not fall away from a lane that was being advertised. Advertising fell away from the lane, and the sales followed it out.
The reason this matters in September rather than in the abstract is that the lane was not a poor performer at the moment it was cut. In the same summer months when it was funded it returned 1.54x Meta-reported revenue against 1.35x for everything else in the account; in December 2025, its own season, it took $13,882 of spend and returned 2.01x against 1.53x for the rest, and did $25,537 in revenue, up 32% on December 2024. It was the best-returning thing in the account, and its budget went to $948 in January, $79 in March, $4 in July. Over the same stretch the account's overall return fell from 1.39x to 0.97x. The rest of the store did grow — everything except this lane went from $53,792 to $102,679 January-through-August, +90.9% — which is the real and good news the single store-wide "−9.6%" hides. Both halves are invisible in any number that aggregates the lanes together.
No single shelf produces this. The founders shelf supplies the principle: Spitznagel's Universa bleeds 0.5% a month on options that expire worthless, sometimes for three years, and would read as a failure on every window shorter than its payoff — "being patient now in order to be strategically impatient later." The brand shelf supplies the artifact: five dossiers written on 2026-08-29, all marked CURRENT SoR, all judging on 90 or 120 days, one of them calling the company's largest lane "fading fast" with the caveat "partly seasonal: her season OPENS at Thanksgiving" left unresolved. The mastermind shelf supplies both the proof that the longer lens changes verdicts (Eric Corry, call #15: "Jan 2025 for sloths $981 in sales. Jan 2026 $84K," in a month brands write off) and the named failure mode (call #25's limiting-beliefs round: Chris Hall's "I can't overcome seasonality"). Tim's library supplies the agrarian version — F. H. King, 1911: "a durable system is governed by accumulated stocks and recurring inflows and outflows, not one season's output alone." And the ops database, which is not a knowledge base at all, supplies the fact that overturns the conclusion all four were converging on.
This also corrects a nugget shipped two nights ago. #026 (2026-08-31) measured the same collapse from the units side and diagnosed it as creative failure — "IGL is not dangerously dependent on one shirt, it lost its one winner and has not produced another." The winner was not lost. Its impressions were, from 1,002,719 in a summer to 5,005.
founders-kb/episodes/070-mark-spitznagel-the-dao-of-capital — the seed. Universa pays 0.5%/month on options that expire worthless, waiting up to three years for the payoff; "being patient now in order to be strategically impatient later"; the conifer that is slow first and accelerates later. Establishes that a measurement window shorter than the payoff cycle turns a working strategy into an apparent loss.igl-brand-kb/audience/profiles/05-carol-garden-witness — status: CURRENT SoR, created 2026-08-29 (5 days old). Lines 110-112: "$176k, 49.5% of all-time, one SKU = 36% of company units — and is fading fast (11.9% of last-120d; lane −22.5% in 90 days vs store −12.4%, partly seasonal: her season OPENS at Thanksgiving)." Line 32 confirms the annual cycle. These are quoted figures from a document, not a measurement — and it is the document this nugget corrects.igl-brand-kb/audience/profiles/02-donna-seed-keeper — status: CURRENT SoR, 2026-08-29, lines 108-109: "Flagship fatigued (Quietly Plotting: $21k lifetime → $444/90d)". A tomato-harvest-plotting phrase judged on a summer window, with the prescribed response being to replace it. The same window defect, second lane.wescale-kb/calls/call-15 — 2026-01-29. Eric Corry, verbatim: "Jan 2025 for sloths $981 in sales. Jan 2026 $84K," which the notes frame as landing "in January, which many brands treat as a dead month." The cohort's own proof that a year-over-year lens overturns a verdict on a written-off month.wescale-kb/calls/call-25 — 2026-04-09, the limiting-beliefs round: Chris Hall's "I can't overcome seasonality," Alex's "my snowsports brand can't scale through summer." Seasonality misreads are a named, recurring cohort failure that nobody has connected to a measurement-window fix.mylibrary-kb/books/farmers-of-forty-centuries — F. H. King, 1911, rights-verified public domain, enriched 2026-08-20. Framework 11: "a durable system is governed by accumulated stocks and recurring inflows and outflows, not one season's output alone," plus "One snapshot is not a growing season." The rule already sits on Tim's shelf in agricultural form and has never been pointed at the dossiers.igl-brand-kb/seasonal/six-season-calendar-operational — theme-aggregate. CURRENT SoR, updated 2026-03-03. Supplies each lane's real season window (Advent opens 29 Nov 2026, Thanksgiving 26 Nov; Epiphany 6 Jan – 17 Feb), which is the mechanism the dossier's season column would use.fact_product_sales_daily, fact_meta_daily and bridge_product_ads in G:\iris\igl-ops-sage\data\igl.db (operator queries run 2026-09-03; sales table covers 2024-07-30 to 2026-09-02, ad table 2024-07-22 to 2026-09-02). Every figure in this card that is not a document quotation comes from these three tables.igl-brand-kb/audience/profiles/05-carol-garden-witness.md section 8 and replace the line "fading fast (11.9% of last-120d; lane −22.5% in 90 days vs store −12.4%)" with the same-weeks-last-year pair and the ad spend printed beside it: June–August 2025 the bird-feeder lane sold $23,912 on $6,592 of Meta spend, which was 22.9% of the whole ad account; June–August 2026 it sold $1,129 on $68, which was 0.24%. Then, in Meta Ads Manager before Thanksgiving (26 Nov 2026), you un-pause the three existing bird-feeder ads that last delivered on 2026-01-15 rather than building new creative for that lane — PostID_Ad_001_Excuse Me in campaign 40_MANUAL_CBO_Scaling_WW_11.13..2025_Dup, and its two copies in 40_MANUAL_CBO_Scaling_US_10.12.2025 and 10_CBO_Scaling_US_SQUIRRELS_11.13.2025 — because they are PostID ads carrying the shirt's accumulated 1,232 ad comments, and they ran at 2.00x, 1.83x and 2.22x Meta-reported return in the last season they were funded. Finally, add one column to the Tuesday WeScale scorecard you already submit: each lane's share of Meta spend printed next to its revenue, so no lane is ever again recorded as fading in a month nobody bought it impressions. ## Why (the one-liner) Spitznagel's rule that a measurement window shorter than the payoff cycle turns a working strategy into an apparent loss led to the year-over-year test the dossiers never ran — and running it showed the opposite of what the dossier concluded: the lane that is 49.5% of all-time revenue was not tested and found wanting, it was defunded and then written up as dying. ## The insight The candidate that won tonight's assay proposed the right instrument and predicted the wrong reading. Its claim was that IGL's five WHO dossiers judge each buyer lane on a 90- or 120-day window, that this is shorter than the annual cycle gardeners live on, and that Carol's lane is therefore being recorded as decaying during her off-season — her own dossier says "her season OPENS at Thanksgiving." So: measure the same weeks last year instead. That instrument is correct and the dossiers genuinely lack it. Run it and the verdict inverts. The same summer window a year apart is not a gentle seasonal dip: the bird-feeder lane went from $23,912 to $1,129, −95.3%, far worse than the −22.5% the dossier reports. Year-over-year would have confirmed the decay verdict, not overturned it. What overturns it is the number that sits in neither the dossier nor any knowledge base: over those same months the lane's share of the Meta ad account fell from 22.9% to 0.24% while total account spend stayed flat (−3%, $28,803 to $27,926). Sales did not fall away from a lane that was being advertised. Advertising fell away from the lane, and the sales followed it out. The reason this matters in September rather than in the abstract is that the lane was not a poor performer at the moment it was cut. In the same summer months when it was funded it returned 1.54x Meta-reported revenue against 1.35x for everything else in the account; in December 2025, its own season, it took $13,882 of spend and returned 2.01x against 1.53x for the rest, and did $25,537 in revenue, up 32% on December 2024. It was the best-returning thing in the account, and its budget went to $948 in January, $79 in March, $4 in July. Over the same stretch the account's overall return fell from 1.39x to 0.97x. The rest of the store did grow — everything except this lane went from $53,792 to $102,679 January-through-August, +90.9% — which is the real and good news the single store-wide "−9.6%" hides. Both halves are invisible in any number that aggregates the lanes together. No single shelf produces this. The founders shelf supplies the principle: Spitznagel's Universa bleeds 0.5% a month on options that expire worthless, sometimes for three years, and would read as a failure on every window shorter than its payoff — "being patient now in order to be strategically impatient later." The brand shelf supplies the artifact: five dossiers written on 2026-08-29, all marked CURRENT SoR, all judging on 90 or 120 days, one of them calling the company's largest lane "fading fast" with the caveat "partly seasonal: her season OPENS at Thanksgiving" left unresolved. The mastermind shelf supplies both the proof that the longer lens changes verdicts (Eric Corry, call #15: "Jan 2025 for sloths $981 in sales. Jan 2026 $84K," in a month brands write off) and the named failure mode (call #25's limiting-beliefs round: Chris Hall's "I can't overcome seasonality"). Tim's library supplies the agrarian version — F. H. King, 1911: "a durable system is governed by accumulated stocks and recurring inflows and outflows, not one season's output alone." And the ops database, which is not a knowledge base at all, supplies the fact that overturns the conclusion all four were converging on. This also corrects a nugget shipped two nights ago. #026 (2026-08-31) measured the same collapse from the units side and diagnosed it as creative failure — "IGL is not dangerously dependent on one shirt, it lost its one winner and has not produced another." The winner was not lost. Its impressions were, from 1,002,719 in a summer to 5,005. ## Evidence trail - founders-kb/episodes/070-mark-spitznagel-the-dao-of-capital — the seed. Universa pays 0.5%/month on options that expire worthless, waiting up to three years for the payoff; "being patient now in order to be strategically impatient later"; the conifer that is slow first and accelerates later. Establishes that a measurement window shorter than the payoff cycle turns a working strategy into an apparent loss. - igl-brand-kb/audience/profiles/05-carol-garden-witness — status: CURRENT SoR, created 2026-08-29 (5 days old). Lines 110-112: "$176k, 49.5% of all-time, one SKU = 36% of company units — and is fading fast (11.9% of last-120d; lane −22.5% in 90 days vs store −12.4%, partly seasonal: her season OPENS at Thanksgiving)." Line 32 confirms the annual cycle. These are quoted figures from a document, not a measurement — and it is the document this nugget corrects. - igl-brand-kb/audience/profiles/02-donna-seed-keeper — status: CURRENT SoR, 2026-08-29, lines 108-109: "Flagship fatigued (Quietly Plotting: $21k lifetime → $444/90d)". A tomato-harvest-plotting phrase judged on a summer window, with the prescribed response being to replace it. The same window defect, second lane. - wescale-kb/calls/call-15 — 2026-01-29. Eric Corry, verbatim: "Jan 2025 for sloths $981 in sales. Jan 2026 $84K," which the notes frame as landing "in January, which many brands treat as a dead month." The cohort's own proof that a year-over-year lens overturns a verdict on a written-off month. - wescale-kb/calls/call-25 — 2026-04-09, the limiting-beliefs round: Chris Hall's "I can't overcome seasonality," Alex's "my snowsports brand can't scale through summer." Seasonality misreads are a named, recurring cohort failure that nobody has connected to a measurement-window fix. - mylibrary-kb/books/farmers-of-forty-centuries — F. H. King, 1911, rights-verified public domain, enriched 2026-08-20. Framework 11: "a durable system is governed by accumulated stocks and recurring inflows and outflows, not one season's output alone," plus "One snapshot is not a growing season." The rule already sits on Tim's shelf in agricultural form and has never been pointed at the dossiers. - igl-brand-kb/seasonal/six-season-calendar-operational — theme-aggregate. CURRENT SoR, updated 2026-03-03. Supplies each lane's real season window (Advent opens 29 Nov 2026, Thanksgiving 26 Nov; Epiphany 6 Jan – 17 Feb), which is the mechanism the dossier's season column would use. - Outside the knowledge bases — fact_product_sales_daily, fact_meta_daily and bridge_product_ads in G:\iris\igl-ops-sage\data\igl.db (operator queries run 2026-09-03; sales table covers 2024-07-30 to 2026-09-02, ad table 2024-07-22 to 2026-09-02). Every figure in this card that is not a document quotation comes from these three tables. ## Assay verdicts - novelty — PASS, 3/5. "No nugget has touched the measurement window on the WHO dossiers, and no cohort document offers a window longer than P30 (call-20's audit SOP)." Deducted because "Carol's own dossier already prints the caveat 'partly seasonal', so the hunt is finishing a sentence the source started." Operator note: the correction below lands past where the source's sentence stops — the spend-share collapse appears in no knowledge base, and independent greps for "PostID", "share of spend", "13,882" and "22.9%" across all thirteen bases and all 28 prior nuggets return zero hits. - grounding — PASS, 5/5. Every quote re-opened and verbatim: Carol lines 110-112, Donna 108-109, call-20 line 205 ("across P7, P14, P30 timeframes" and stops there), call-15 line 304, call-25 line 22. - actionability — PASS, 5/5. "Names the specific decision it changes and it is a live, imminent one: do not move Q4 budget away from the lane that is 49.5% of all-time revenue on the strength of a June-through-August window." - Summed strength 13/15 — top of three 3-of-3 survivors in round 1. ## Operator corrections applied before shipping 1. The move's hedge was unnecessary and was deleted. The candidate wrote "if the table does not reach back that far, pull the Shopify order export for Nov 2025-Feb 2026 once and keep it." fact_product_sales_daily covers 2024-07-30 to 2026-09-02 — over two years. The year-over-year comparison is directly runnable, so the nugget runs it instead of asking Tim to. 2. Running the proposed test reverses the candidate's conclusion, and this is the decisive correction. The candidate's stated reason for protecting Carol's Q4 budget was that a summer window misreads a seasonal trough as decay. Measured: same weeks a year apart, the lane is −95.3%, materially worse than the −22.5% the dossier reports. Its instrument was right and its reading was wrong. The finding that actually protects the budget is the 99% spend cut, which the candidate never looked at. 3. A causal claim was scoped down to what the data supports. This card does not assert that restoring spend restores revenue. It asserts the narrower and fully evidenced thing: the lane was not tested at low demand, it was defunded, so "fading" is not a conclusion the data licenses. The named action is a funded retest before Carol's season, not a prediction. 4. All return-on-ad-spend figures are labelled Meta-reported, not Shopify-verified. IGL's attribution is known to be partial, so these are the platform's numbers and are used only to compare the lane against the rest of the same account over the same days — a comparison that survives an attribution gap applied evenly. 5. Nugget #026's diagnosis is corrected rather than quietly contradicted, and the contradiction is stated in the body so the catalog does not carry two opposed readings of the same collapse. 6. Sixth night running, the decisive correction came from igl.db and nowhere else — and this time in its newest form: the candidate did not declare the measurement impossible (#028's failure) or quote a stale document (#007's), it proposed a live, correct query and then reasoned about the answer without running it. Thread: ep-070 Spitznagel (0.5%/month bleed, three-year payoff, "patient now to be strategically impatient later") → the measurement window is the time horizon → Carol dossier 2026-08-29, "fading fast, −22.5% in 90 days… her season OPENS at Thanksgiving" → the other four dossiers, all 90/120-day verdicts → call-20's audit SOP stops at P30 → call-15 Eric Corry $981 → $84K year-over-year January → call-25 "I can't overcome seasonality" as a named limiting belief → King, Farmers of Forty Centuries, "one snapshot is not a growing season" → six-season calendar supplies the real season windows → operator: run it against igl.db, and the lane is −95.3%, not −22.5% → check the spend and it fell from 22.9% of the account to 0.24%Step 0 (10 minutes, before any copy). Your European orders are not a plan, they are a running operation: 135 of your last 1,609 Printify orders — 8.4%, about one in twelve — printed at OPT OnDemand in Prague, and every one of those 135 was charged VAT (value-added tax, Europe's sales tax) averaging $3.96 an order. All-time that is 189 orders and $739.36 of VAT you have already paid. Ask Jakob and Petra in the mastermind how they stopped paying it: they run Wild Spirit, a nature-and-birds shirt brand selling into the EU, did $1M in ten months at 25-30% net margin, and say plainly "we don't pay VAT on Printify because we are registered with our company in the EU." Same niche, same problem, already solved, one Slack message away.
Then the copy, on one product family only. In Shopify Markets, add market-specific product-page copy for the European Union market reading "Printed and shipped inside the EU — no customs or import fees," the wording already approved in igl-brand-kb/voice/email-templates.md section 21 Variant A. Put it only on the Gildan 64000 Softstyle "Essential Tee" (Printify blueprint 145). That is the one product family the Prague routing is proven on — 334 of the 337 line items OPT OnDemand has ever printed for you are blueprint 145. A store-wide line would promise "no customs fees" on hoodies and Comfort Colors products that have never once routed to Prague, and Dan Glavin warns in wescale-kb/members/dan-glavin.md:211 that Printify's Global Fulfilment "must be activated, it is not automatic, and it is product by product."
Then, and only then, the test. One European ad set at your normal $25-30/day test budget with the no-duties line as the hook, held to the kill rules already written in wescale-kb/planning/igl-ads-execution-plan-2026-07-08.md: kill at ~$55-60 spend with no purchase, or sustained return on ad spend below 1.7.
IGL has quietly been running a European business for a year — one order in twelve, printed inside the EU and therefore duty-free for the customer — and the only place that fact is written down is a customer-service reply that fires after a shopper thinks to email and ask.
Kerkorian could not afford the surplus military planes that were obviously mispriced, so he did something more useful than buying them: he worked out why the profit was sitting there uncaptured. The answer was range — the planes could not cross the water. He gutted them, installed extra fuel tanks, flew them over himself, and took the spread. He did not find a cheap asset; he found the single constraint keeping everyone else away, and removed exactly that.
For a United States shirt brand selling into Europe, the constraint is customs. Transatlantic apparel arrives with a duty bill at the door, which is why so few small American shops advertise there. IGL removed that constraint without ever treating it as an asset: EU orders print at OPT OnDemand in Prague, so the package never crosses a customs border. What makes this a finding rather than a nice theory is that it is measured, not configured. The knowledge bases say IGL can print in Prague. The operations database says it does, 135 times in six months, every one of them carrying Czech VAT at about 21% of cost — which is what an order genuinely fulfilled inside the EU looks like, and what a US-fulfilled order does not.
The proof it is real and not theoretical is in the template's own origin note: a Belgian customer, Hatice Bakindi, wrote in on 2026-06-11 asking about import taxes. The first draft of the reply hedged with "duties might apply." Checking the actual provider location flipped the answer entirely in the customer's favor. That is a completed, Tim-approved answer to a question nobody is being invited to ask — reachable only by the small fraction of shoppers motivated enough to email a foreign brand before buying instead of closing the tab.
And it carries Kerkorian's other requirement, the one he insisted on before sizing any bet: the downside is already capped. The provider is live, the copy is written and approved, the customs logic is documented with a verification rule, and the fulfillment leg costs nothing to keep if the ads fail. The only new spend is one test ad set. No single knowledge base could produce this — the founders shelf supplies the fuel-tank move, the brand shelf holds the fuel tank IGL already installed, the mastermind shelf supplies both the member who solved the VAT half and the discipline for how to test it, and only the operations database could say how big it already is.
founders-kb/episodes/065-kirk-kerkorian-greatest-deal-maker — the seed practice (line 72): study the planes enough to understand that limited range is why the profit sits uncaptured, install extra fuel tanks, pilot them over yourself. Plus the Douglas Aircraft buyback clause (line 170) as the downside-capping habit.igl-brand-kb/voice/email-templates — Template #21, classified REPLY_ONLY (line 52): the whole no-duties argument, the Session-407 verify-provider-location-first hard rule (lines 1009-1013), and the Hatice Bakindi origin note. Version 3.4, added 2026-06-11, "Approved by Tim S407."wescale-kb/members/jakob-and-petra — Wild Spirit, nature/birds, Austria, selling across the EU: $1M in ten months, 25-30% net margin, and the VAT answer in their own words (lines 19, 38, 168, 171). The hunt missed this file; the operator added it.wescale-kb/members/dan-glavin — line 211, dated 2026-08-06: Australia stood up in two days, "so easy to stand up," ~2.2 return on ad spend, sells when the US is dead. Same line: "Holding the UK and Europe until VAT is sorted" and Global Fulfilment is per product.wescale-kb/members/alex-wheeler — 2026-08-05: opened all markets by email alone, first international order within hours, and the mechanism this move needs — "markets are not shipping zones," per-market theme edits override the store default one-way.wescale-kb/planning/igl-ads-execution-plan-2026-07-08 — (theme-aggregate) the $25-30/day test budget and the kill rules (~$55-60 with no purchase, or sustained return on ad spend below 1.7). Its $54.42 average order value and 2.06% conversion come from a single travel week (2026-W26) and the document says so — read them as one week, not a standing measurement.G:\iris\igl-ops-sage\data\igl.db (fact_printify_orders, fact_printify_line_items) for the 8.4% share, the VAT totals and the blueprint routing; igl-production/shopify-uploader/knowledge/blank-registry.md section 5 for EU to OPT OnDemand (id 30, Prague CZ) and section 7 for the parked regional-catalog decision.igl-brand-kb/voice/email-templates.md section 21 Variant A. Put it only on the Gildan 64000 Softstyle "Essential Tee" (Printify blueprint 145). That is the one product family the Prague routing is proven on — 334 of the 337 line items OPT OnDemand has ever printed for you are blueprint 145. A store-wide line would promise "no customs fees" on hoodies and Comfort Colors products that have never once routed to Prague, and Dan Glavin warns in wescale-kb/members/dan-glavin.md:211 that Printify's Global Fulfilment "must be activated, it is not automatic, and it is product by product." Then, and only then, the test. One European ad set at your normal $25-30/day test budget with the no-duties line as the hook, held to the kill rules already written in wescale-kb/planning/igl-ads-execution-plan-2026-07-08.md: kill at ~$55-60 spend with no purchase, or sustained return on ad spend below 1.7. ## Why (the one-liner) IGL has quietly been running a European business for a year — one order in twelve, printed inside the EU and therefore duty-free for the customer — and the only place that fact is written down is a customer-service reply that fires after a shopper thinks to email and ask. ## The insight Kerkorian could not afford the surplus military planes that were obviously mispriced, so he did something more useful than buying them: he worked out why the profit was sitting there uncaptured. The answer was range — the planes could not cross the water. He gutted them, installed extra fuel tanks, flew them over himself, and took the spread. He did not find a cheap asset; he found the single constraint keeping everyone else away, and removed exactly that. For a United States shirt brand selling into Europe, the constraint is customs. Transatlantic apparel arrives with a duty bill at the door, which is why so few small American shops advertise there. IGL removed that constraint without ever treating it as an asset: EU orders print at OPT OnDemand in Prague, so the package never crosses a customs border. What makes this a finding rather than a nice theory is that it is measured, not configured. The knowledge bases say IGL can print in Prague. The operations database says it does, 135 times in six months, every one of them carrying Czech VAT at about 21% of cost — which is what an order genuinely fulfilled inside the EU looks like, and what a US-fulfilled order does not. The proof it is real and not theoretical is in the template's own origin note: a Belgian customer, Hatice Bakindi, wrote in on 2026-06-11 asking about import taxes. The first draft of the reply hedged with "duties might apply." Checking the actual provider location flipped the answer entirely in the customer's favor. That is a completed, Tim-approved answer to a question nobody is being invited to ask — reachable only by the small fraction of shoppers motivated enough to email a foreign brand before buying instead of closing the tab. And it carries Kerkorian's other requirement, the one he insisted on before sizing any bet: the downside is already capped. The provider is live, the copy is written and approved, the customs logic is documented with a verification rule, and the fulfillment leg costs nothing to keep if the ads fail. The only new spend is one test ad set. No single knowledge base could produce this — the founders shelf supplies the fuel-tank move, the brand shelf holds the fuel tank IGL already installed, the mastermind shelf supplies both the member who solved the VAT half and the discipline for how to test it, and only the operations database could say how big it already is. ## Evidence trail - founders-kb/episodes/065-kirk-kerkorian-greatest-deal-maker — the seed practice (line 72): study the planes enough to understand that limited range is why the profit sits uncaptured, install extra fuel tanks, pilot them over yourself. Plus the Douglas Aircraft buyback clause (line 170) as the downside-capping habit. - igl-brand-kb/voice/email-templates — Template #21, classified REPLY_ONLY (line 52): the whole no-duties argument, the Session-407 verify-provider-location-first hard rule (lines 1009-1013), and the Hatice Bakindi origin note. Version 3.4, added 2026-06-11, "Approved by Tim S407." - wescale-kb/members/jakob-and-petra — Wild Spirit, nature/birds, Austria, selling across the EU: $1M in ten months, 25-30% net margin, and the VAT answer in their own words (lines 19, 38, 168, 171). The hunt missed this file; the operator added it. - wescale-kb/members/dan-glavin — line 211, dated 2026-08-06: Australia stood up in two days, "so easy to stand up," ~2.2 return on ad spend, sells when the US is dead. Same line: "Holding the UK and Europe until VAT is sorted" and Global Fulfilment is per product. - wescale-kb/members/alex-wheeler — 2026-08-05: opened all markets by email alone, first international order within hours, and the mechanism this move needs — "markets are not shipping zones," per-market theme edits override the store default one-way. - wescale-kb/planning/igl-ads-execution-plan-2026-07-08 — (theme-aggregate) the $25-30/day test budget and the kill rules (~$55-60 with no purchase, or sustained return on ad spend below 1.7). Its $54.42 average order value and 2.06% conversion come from a single travel week (2026-W26) and the document says so — read them as one week, not a standing measurement. - Outside the shelves: G:\iris\igl-ops-sage\data\igl.db (fact_printify_orders, fact_printify_line_items) for the 8.4% share, the VAT totals and the blueprint routing; igl-production/shopify-uploader/knowledge/blank-registry.md section 5 for EU to OPT OnDemand (id 30, Prague CZ) and section 7 for the parked regional-catalog decision. ## Assay verdicts - novelty — PASS (4/5): Grepped "customs", "import dut", "duties" across the bases and the whole nugget catalog; every hit is the reply-only template itself or margin-side VAT warnings. Nothing proposes surfacing the fact as storefront copy. Docked one point for not addressing the VAT constraint named in a file it cited. - grounding — PASS (4/5): All quotations verified verbatim. Caught a two-day slip on Dan Glavin's log date (2026-08-06, not 2026-08-08). Correctly refused to assert any IGL sales figure by country. - actionability — PASS (4/5): One line of copy, wording pre-approved, one capped test, with its own verification gate. Held at 4 because it opens a new lane rather than changing a decision already on the desk. Summed score 12/15. Two other candidates also passed 3-of-3 tonight; they are recorded as SEED lines in the nightly log, not in this catalog. ## Operator corrections (5, two decisive) 1. DECISIVE — the size was never measured, and it is measurable. The candidate said the bases "contain no IGL sales or traffic data by country at all," which is true of the bases and irrelevant: fact_printify_orders carries print_provider, and provider 30 is OPT OnDemand. 135 of 1,609 orders in the last 180 days (8.4%), 189 all-time back to 2025-09-12, most recently 2026-08-27. 100% of them carry tax averaging 21.1% of cost ($7.13 on $33.85, $2.33 on $10.98) against about $1 and only 94% of orders for the US providers — the signature of genuine EU fulfilment. This turns "you could" into "you already are," a different and much stronger finding. Fourth consecutive night the decisive correction came from the operations database and nowhere else. 2. DECISIVE — the move ignored VAT while citing the file that names it as the blocker. The candidate used dan-glavin.md:211 as demand evidence; the same line says he is holding Europe until VAT is sorted. And IGL is paying it: $739.36 to date. The correction adds Jakob and Petra — same niche, EU-based, $1M in ten months — whose stated reason for not paying it is an EU company registration. Advertising harder into a lane carrying an unaddressed ~21% supplier-VAT surcharge would have scaled a known cost, silently. 3. The store-wide product-page line was wrong at catalog scale. OPT OnDemand has only ever printed blueprint 145 (334 line items) and blueprint 80 (3). Scoped the copy to the Essential Tee. Corroborated by Dan Glavin's note that Global Fulfilment is per product, not automatic. 4. Blueprint identity corrected: blueprint 145 is the Gildan 64000 Softstyle, IGL's "Essential Tee" (blank-registry.md line 26), not the Gildan 5000 that Jakob and Petra run. 5. Better demand evidence, and it answers IGL's own parked question. Alex Wheeler opened all markets by email alone and got an international order within hours — closer to this move than Dan Glavin's Australia. His "per-market theme edits override the store default one-way" is precisely the Shopify-Markets capability verification that blank-registry.md section 7 lists as a prerequisite for the parked regional-catalog decision. The cohort already answered a question IGL wrote down as open. Thread: founders ep-065, find the one constraint holding the profit and remove exactly that → igl-brand-kb template 21, a Belgian customer asks about duties and the true answer is zero because IGL prints in Prague → the operations database says that is 8.4% of orders, all VAT-charged → wescale-kb jakob-and-petra hold the VAT answer, dan-glavin holds the warning, alex-wheeler holds the mechanism → igl-ads-execution-plan supplies the $25-30/day capped testBuild the left-chest template Erik Stefans described on 2026-07-31 — Printify's embroidery template as the placement reference, wired into your upload flow so the direct-to-garment version lands in the right spot automatically. Then pick what feeds it from live sales, not from the review file. One query against fact_product_sales_daily in G:\iris\igl-ops-sage\data\igl.db, filtered to dim_products.product_type = 'Essential Tees' and shopify_status = 'active', sales since 2026-06-03, gives you 298 shirts that sold at all and the top 25 below. Count the words in each design's phrase and route by Tim's own Half-Second Test, which allows an image alone or an image plus 3-5 words:
``sql -- G:\iris\igl-ops-sage\data\igl.db — the left-chest worklist, trailing 90 days SELECT f.product_title, SUM(f.quantity) AS units, MAX(f.date) AS last_sale FROM fact_product_sales_daily f JOIN dim_products p ON p.shopify_product_id = f.shopify_product_id WHERE f.date >= date('now','-90 day') AND p.product_type = 'Essential Tees' AND p.shopify_status = 'active' GROUP BY 1 ORDER BY units DESC LIMIT 25; ``
The mastermind's highest-rated free play of the month has a routing rule hidden inside Tim's own design test, and the shortlist every brand document would hand you is drawn from a nine-month-old review file whose top pick has not sold a shirt since February.
Kalanick and Graves are the two poles of the same question, and left-chest recycling sits exactly between them. Kalanick creates problems on purpose — new cities, new industries, new companies inside the company — and holds that "the only constraint on our imagination is management capacity" (interview 034, lines 14 and 162). Graves went the other way for thirty years: one product, no menu additions, deliberately eating the "veto vote" that pushed McDonald's to stack items, because "variety drives more sales… it's actually frequency" (interview 004, line 176). Recycling a design into a pocket version is a Graves move wearing Kalanick clothes — no new concepts at all, just more depth on what already sells. That is why the mastermind's weekly run log ranks it first for IGL: "zero new concepts, just a repositioning template in the upload flow."
The non-obvious part is that the play has a routing rule nobody wrote down, and the rule is already in IGL's own files. Tim's Half-Second Test, added to buyer-portraits.md on 2026-08-29 and marked CURRENT SoR, requires a club member to read a design in about 0.2 seconds — "fewer than 5-7 words — often an image alone, or an image plus 3-5 words." Carol's decision card in the same file specifies a font "big enough to read across a yard." A design shrunk to a 3-inch chest patch either still clears those bars or it does not, and the word count of the design's own phrase decides it before anyone opens a mockup. Erik's post already contains both branches — shrink the whole thing, or lift one element and drop the text — but not the rule for choosing.
The correction that matters is which shirts go on the list. Every brand document that ranks IGL's designs ranks them by the December 2025 review export or by Eric Corry's January 2026 audit of 2025 full-year sales, and both describe a store that no longer exists. In that review table, Vibrant Birds Garden is the third-most-reviewed product; in the sales table its T-shirt last sold on 2026-02-03. The store's current number one and two — Quietly Cheering for Moss at 76 units and Listen to the Birds at 68 — do not appear in the review table at all. The audit's "top 2 designs = 55% of all units" made this look like a narrow shortlist play against a handful of survivors; in the trailing 90 days it takes 32 products to reach half the units and 298 shirts sold at least one. The eligible pool is roughly ten times what the documents imply, which is the difference between a fiddly afternoon and a genuine quarter of free inventory.
One thing the sales data settles that no amount of reading would: the corpus records that "the word 'quietly' is wrong — six separate customers rejected it on the moss design alone" (buyer-portraits.md:237). Quietly Cheering for Moss is nonetheless the best-selling shirt in the store over the last 90 days, and Quietly Rooting for Moss is seventh. Both are four words. Both shrink.
founders-kb/interviews/034-travis-kalanick-transcript — Seed. Problem creation as the growth engine ("the only constraint on our imagination is management capacity", lines 14 and 162), and the specialised-machine argument: a humanoid for varied low-volume work, a purpose-built machine for a thousand pancakes an hour (lines 142-144). Prior-art note: the short notes twin of this interview was mined on an earlier night; the transcript was read here at higher resolution.founders-kb/interviews/004-todd-graves-transcript — The opposite founder. The veto-vote argument he refused (lines 18, 20, 56), "variety drives more sales — it's actually frequency" (line 176), and "death by a thousand cuts, then your food one day is not cravable" (line 30).wescale-kb/members/erik-stefans — The play and its numbers, Slack post dated 2026-07-31 (lines 103-111): 42 left-chest versions live, $29,182 in 60 days at about 8% of store revenue; a clean 29-design batch launched one day in late May doing $18,815 in 69 days (~$649 per design); +23% additive on the 25 whose full-front original was still selling; the lift-one-element alternative; and the Printify embroidery-template placement trick.igl-brand-kb/community/customer-reviews-q4-2025 — Export dated 2025-12-11, covering 2024-08-31 to 2025-12-09, 551 reviews. The Top Reviewed Products table (lines 246-255) that a review-ranked shortlist would be built from: bird feeder 215, Quietly Plotting 29, Vibrant Birds Garden 14, Cardinal Birds Graphic 7. Cited here as the source that is wrong for this purpose, not as support.wescale-kb/members/eric-corry — The 28-29 January 2026 IGL consulting review: Tier 1 = 123 products at 5+ sales; "Top 2 designs = 55% of all units sold (bird feeder alone = 49%)"; 423 products with sales in 2025 on 6,002 units. Explicitly a 2025 full-year figure — see the operator corrections.igl-brand-kb/audience/buyer-portraits (theme-aggregate — the one permitted) — status: CURRENT SoR, created 2026-08-29. The Half-Second Test in Tim's words at line 431 ("~0.2 seconds… fewer than 5-7 words — often an image alone, or an image plus 3-5 words"), Carol's font card at line 405 ("big enough to read across a yard"), the "quietly" rejection at line 237, and the Quietly Plotting collapse at line 178. Also cited, not counted toward the thread gate: wescale-kb/planning/weekly-run-log.md:102 — "erik-stefans' left-chest recycling is the highest return-on-effort idea of the week for IGL." (An aggregate; the candidate's attribution to the run log was correct and the grounding judge's contrary note was mistaken.) And G:\iris\igl-ops-sage\data\igl.db, tables fact_product_sales_daily and dim_products, which sits outside all thirteen knowledge bases.fact_product_sales_daily in G:\iris\igl-ops-sage\data\igl.db, filtered to dim_products.product_type = 'Essential Tees' and shopify_status = 'active', sales since 2026-06-03, gives you 298 shirts that sold at all and the top 25 below. Count the words in each design's phrase and route by Tim's own Half-Second Test, which allows an image alone or an image plus 3-5 words: - Four words or fewer → shrink as-is, no decision needed. That is 18 of the top 25, starting with Quietly Cheering for Moss (76 units), Listen to the Birds (68), Life's Too Short Tomatoes (43), Johnny Appleseed (26), Quietly Rooting for Moss (24). - Exactly five words → eye-check each one at 3 inches before committing. That is the other 7, including "Excuse Me, Bird Feeder's Empty" (28 units, and the only one that is a full line of dialogue). If a five-worder fails the check, use Erik's second method from the same post — lift one element out and drop the text. For the bird feeder that is the squirrel character alone. - Do not put Vibrant Birds Garden on the list. Its T-shirt last sold 2026-02-03, nearly seven months ago, and Erik's +23% lift appeared only on designs whose full-front original was still selling. ``sql -- G:\iris\igl-ops-sage\data\igl.db — the left-chest worklist, trailing 90 days SELECT f.product_title, SUM(f.quantity) AS units, MAX(f.date) AS last_sale FROM fact_product_sales_daily f JOIN dim_products p ON p.shopify_product_id = f.shopify_product_id WHERE f.date >= date('now','-90 day') AND p.product_type = 'Essential Tees' AND p.shopify_status = 'active' GROUP BY 1 ORDER BY units DESC LIMIT 25; ` ## Why (the one-liner) The mastermind's highest-rated free play of the month has a routing rule hidden inside Tim's own design test, and the shortlist every brand document would hand you is drawn from a nine-month-old review file whose top pick has not sold a shirt since February. ## The insight Kalanick and Graves are the two poles of the same question, and left-chest recycling sits exactly between them. Kalanick creates problems on purpose — new cities, new industries, new companies inside the company — and holds that "the only constraint on our imagination is management capacity" (interview 034, lines 14 and 162). Graves went the other way for thirty years: one product, no menu additions, deliberately eating the "veto vote" that pushed McDonald's to stack items, because "variety drives more sales… it's actually frequency" (interview 004, line 176). Recycling a design into a pocket version is a Graves move wearing Kalanick clothes — no new concepts at all, just more depth on what already sells. That is why the mastermind's weekly run log ranks it first for IGL: "zero new concepts, just a repositioning template in the upload flow." The non-obvious part is that the play has a routing rule nobody wrote down, and the rule is already in IGL's own files. Tim's Half-Second Test, added to buyer-portraits.md on 2026-08-29 and marked CURRENT SoR, requires a club member to read a design in about 0.2 seconds — "fewer than 5-7 words — often an image alone, or an image plus 3-5 words." Carol's decision card in the same file specifies a font "big enough to read across a yard." A design shrunk to a 3-inch chest patch either still clears those bars or it does not, and the word count of the design's own phrase decides it before anyone opens a mockup. Erik's post already contains both branches — shrink the whole thing, or lift one element and drop the text — but not the rule for choosing. The correction that matters is which shirts go on the list. Every brand document that ranks IGL's designs ranks them by the December 2025 review export or by Eric Corry's January 2026 audit of 2025 full-year sales, and both describe a store that no longer exists. In that review table, Vibrant Birds Garden is the third-most-reviewed product; in the sales table its T-shirt last sold on 2026-02-03. The store's current number one and two — Quietly Cheering for Moss at 76 units and Listen to the Birds at 68 — do not appear in the review table at all. The audit's "top 2 designs = 55% of all units" made this look like a narrow shortlist play against a handful of survivors; in the trailing 90 days it takes 32 products to reach half the units and 298 shirts sold at least one. The eligible pool is roughly ten times what the documents imply, which is the difference between a fiddly afternoon and a genuine quarter of free inventory. One thing the sales data settles that no amount of reading would: the corpus records that "the word 'quietly' is wrong — six separate customers rejected it on the moss design alone" (buyer-portraits.md:237). Quietly Cheering for Moss is nonetheless the best-selling shirt in the store over the last 90 days, and Quietly Rooting for Moss is seventh. Both are four words. Both shrink. ## Worked examples — the top 25 live-selling tees, routed | Units (90d) | Words | Last sale | Design | Route | |---|---|---|---|---| | 76 | 4 | 2026-08-28 | Quietly Cheering for Moss | shrink as-is | | 68 | 4 | 2026-08-30 | Listen to the Birds | shrink as-is | | 43 | 4 | 2026-08-26 | Life's Too Short Tomatoes | shrink as-is | | 28 | 5 | 2026-08-31 | Excuse Me, Bird Feeder's Empty | eye-check; lift the squirrel if it fails | | 26 | 2 | 2026-08-25 | Johnny Appleseed | shrink as-is | | 26 | 5 | 2026-08-25 | Fruit of Spirit Seed Packets | eye-check | | 24 | 4 | 2026-08-28 | Quietly Rooting for Moss | shrink as-is | | 23 | 5 | 2026-08-26 | My Chickens Are My Therapists | eye-check | | 22 | 3 | 2026-08-25 | Know Your Tomatoes | shrink as-is | | 18 | 4 | 2026-08-23 | Support Your Local Farmacy | shrink as-is | | 18 | 4 | 2026-08-02 | One Spirit Many Colors | shrink as-is | | 18 | 3 | 2026-08-30 | I Like Silence | shrink as-is | | 15 | 2 | 2026-08-14 | Tomato Love | shrink as-is | | 14 | 5 | 2026-08-26 | Rooting for the Weird Ones | eye-check | | 13 | 4 | 2026-08-26 | Quietly Plotting Tomato Harvest | shrink as-is (but see note) | | 13 | 5 | 2026-08-26 | A Lot of Freaking Plants | eye-check | | 13 | 4 | 2026-08-23 | More Wildflowers Less Hurry | shrink as-is | | 12 | 3 | 2026-07-25 | Cardinal Birds Graphic | shrink as-is | | 12 | 2 | 2026-08-27 | Chaos Gardener | shrink as-is | | 10 | 5 | 2026-08-11 | Excuse to Avoid Social Interaction | eye-check | | 10 | 2 | 2026-08-25 | Be Kind | shrink as-is | | 10 | 5 | 2026-08-14 | Lock Your Car Doors Zucchini | eye-check | | 9 | 3 | 2026-08-18 | Tomato Lovers Collective | shrink as-is | | 9 | 4 | 2026-08-22 | I Trust The Forest | shrink as-is | | 9 | 3 | 2026-08-24 | Sorry I'm Late | shrink as-is | Excluded on inspection: "Roots of Renewal — Join Us and Plant Even More Trees!" sold 27 units in the window and ranks fifth on raw units, but dim_products records it as product_type = sprout-additional-trees, shopify_status = draft. It is the tree-planting add-on, not a shirt. It is not a left-chest candidate. Note on Quietly Plotting: it passes the word test at four words, but buyer-portraits.md:178 records that the flagship "Quietly Plotting" has collapsed to $444 in 90 days from a $21k lifetime, and the sales table agrees (13 units). It is eligible, not a priority. ## Evidence trail - founders-kb/interviews/034-travis-kalanick-transcript — Seed. Problem creation as the growth engine ("the only constraint on our imagination is management capacity", lines 14 and 162), and the specialised-machine argument: a humanoid for varied low-volume work, a purpose-built machine for a thousand pancakes an hour (lines 142-144). Prior-art note: the short notes twin of this interview was mined on an earlier night; the transcript was read here at higher resolution. - founders-kb/interviews/004-todd-graves-transcript — The opposite founder. The veto-vote argument he refused (lines 18, 20, 56), "variety drives more sales — it's actually frequency" (line 176), and "death by a thousand cuts, then your food one day is not cravable" (line 30). - wescale-kb/members/erik-stefans — The play and its numbers, Slack post dated 2026-07-31 (lines 103-111): 42 left-chest versions live, $29,182 in 60 days at about 8% of store revenue; a clean 29-design batch launched one day in late May doing $18,815 in 69 days (~$649 per design); +23% additive on the 25 whose full-front original was still selling; the lift-one-element alternative; and the Printify embroidery-template placement trick. - igl-brand-kb/community/customer-reviews-q4-2025 — Export dated 2025-12-11, covering 2024-08-31 to 2025-12-09, 551 reviews. The Top Reviewed Products table (lines 246-255) that a review-ranked shortlist would be built from: bird feeder 215, Quietly Plotting 29, Vibrant Birds Garden 14, Cardinal Birds Graphic 7. Cited here as the source that is wrong for this purpose, not as support. - wescale-kb/members/eric-corry — The 28-29 January 2026 IGL consulting review: Tier 1 = 123 products at 5+ sales; "Top 2 designs = 55% of all units sold (bird feeder alone = 49%)"; 423 products with sales in 2025 on 6,002 units. Explicitly a 2025 full-year figure — see the operator corrections. - igl-brand-kb/audience/buyer-portraits (theme-aggregate — the one permitted) — status: CURRENT SoR, created 2026-08-29. The Half-Second Test in Tim's words at line 431 ("~0.2 seconds… fewer than 5-7 words — often an image alone, or an image plus 3-5 words"), Carol's font card at line 405 ("big enough to read across a yard"), the "quietly" rejection at line 237, and the Quietly Plotting collapse at line 178. Also cited, not counted toward the thread gate: wescale-kb/planning/weekly-run-log.md:102 — "erik-stefans' left-chest recycling is the highest return-on-effort idea of the week for IGL." (An aggregate; the candidate's attribution to the run log was correct and the grounding judge's contrary note was mistaken.) And G:\iris\igl-ops-sage\data\igl.db, tables fact_product_sales_daily and dim_products, which sits outside all thirteen knowledge bases. ## Assay verdicts - novelty — PASS, 4 of 5. Verified the advice-layer match: the run log assigns the play but no file constrains it. "left-chest" returns zero hits across CATALOG.md and all 26 catalog cards. A genuine narrowing, and the operator corrections add a second novel element the judge did not see — that the eligible list contradicts the brand's own ranking documents. - grounding — PASS, 4 of 5. Every quotation re-opened and verified verbatim by the operator: Erik's six figures, all ten review counts, both Kalanick passages, all four Graves passages, and all four buyer-portraits lines. One judge ding was itself wrong (see evidence trail). The score stands, but two of the candidate's premises were document-true and world-false — corrected below. - consequence — PASS, 4 of 5. Startable this week on a template Tim was already told to build, solo, no capital. Names the specific decision it changes: which designs feed the template first. ## Operator corrections applied before shipping Five corrections; the two decisive ones are again PREMISES rather than misquotations — the seventh night running. 1. The shrink-as-is shortlist was dead on arrival. The candidate's first named pick, vibrant-birds-garden-t-shirt, was chosen because it ranks third in the December 2025 review table. Its T-shirt last sold 2026-02-03; only its long-sleeve sibling has moved since (1 unit, last 2026-06-24). Erik's +23% was measured only on designs whose original was still selling, so the play does not apply to it at all. Removed and replaced with the live list. Cardinal Birds Graphic (12 units, last 2026-07-25) and Tomato Lovers Collective (18 across six cuts) survive the same check and stay. 2. The "shortlist not catalog-wide" argument runs backwards. The candidate reasoned from Eric Corry's "Tier 1 = 123 products, top 2 = 55% of units" that the eligible pool is small and lopsided. That is a January 2026 audit of 2025 full-year data — the exact figure nugget #026 measured as expired one night ago. In the trailing 90 days it takes 32 products to reach half of 1,220 units and 298 shirts sold at least once. The pool is roughly ten times larger than the candidate claimed, and that widening is what makes the play worth doing. 3. A better test the hunt missed, and it is mechanical. The candidate routed by "illustration-led versus text joke," a judgment call. Tim's own Half-Second Test states a word budget, so the routing can be decided by counting words in the design's phrase. Run against the live top 25, every design is 2-5 words — none exceeds the budget outright. So the candidate's framing ("your two most important designs are jokes made of words that break at pocket size") is false as stated; the honest split is 18 clear shrinks at four words or fewer and 7 boundary cases at exactly five. 4. One recommendation was inverted. The candidate told Tim to leave listen-to-birds-not-news alone entirely. That product sold 8 units in 90 days. The store's second-best-selling shirt is "Listen to the Birds T-Shirt" at 68 units, a different product the candidate never mentions — invisible to it because it does not appear in the review table. 5. A non-shirt was about to enter the list. "Roots of Renewal — Join Us and Plant Even More Trees!" ranks fifth on raw units (27) in the window. dim_products records it as sprout-additional-trees, status draft` — the tree-planting add-on. Filtered out by product type. Thread: kalanick interview (create problems; purpose-built beats generalist) → graves interview (one item, thirty years, eat the veto vote) → erik-stefans 2026-07-31 (left-chest recycling, +23% additive, or lift one element) → weekly-run-log ranks it #1 for IGL → customer-reviews-q4-2025 top-products table (the shortlist a reader would build) → buyer-portraits CURRENT SoR (the Half-Second Test that routes it) → igl.db (the shortlist that is actually true)Three things, all on Tuesday, all on machinery you already run.
1. Correct the record. igl-brand-kb/core/strategic-synthesis-v4.md is marked Status: CURRENT SoR and its Section 12 still reads "Excuse Me, Bird Feeder's Empty -- 2,965 units (49% of all sales)" and "Top 2 designs: 55% of all units sold." Those numbers are labelled 2025 Full Year and they are correct for 2025. They are also the numbers every plan, agent and strategy document in this workspace reads as the shape of the business. Add one line directly beneath that block: "2026 YTD (through Aug 30): top design 9.3% of units, top 2 = 18.3%, 30 designs needed to reach half of units in the trailing 90 days." Until that line exists, every downstream plan is describing a company that stopped existing in January.
2. Add two numbers to the Tuesday scorecard block. The ## WeScale Scorecard Health block in wescale-kb/members/tim-dobyns.md already has an "Inputs This Week" line and already feeds your 3 PM CT Tuesday submission. Add top-design share and designs-to-half beside it. Both come from one query against the ops database you already keep — no hand-reading of Shopify admin, and no dependence on the design_id join that has been broken since mid-July:
``sql -- trailing 90 days, G:\iris\igl-ops-sage\data\igl.db SELECT product_title, SUM(quantity) AS units FROM fact_product_sales_daily WHERE date >= date('now','-90 day') GROUP BY 1 ORDER BY units DESC; -- top-design share = units of row 1 / total -- designs-to-half = how many rows it takes to reach 50% of total ``
3. Change Tuesday's question. The cohort's whole Scale-vs-Fix routine, the Ad Creative Skill workflow from Call #40 (take your top 5 designs, generate 20 variations each), and Alex Wheeler's adapt-the-inside-jokes play all assume there is a winner to work from. At 6.8% of trailing-90-day units, you do not currently have one. So the Fix question this quarter is not "am I too concentrated" — it is "why has nothing replaced the bird feeder." Before Harvest/Fall opens September 27, take the top three of the trailing 90 days by name — Quietly Cheering for Moss (84 units), Listen to the Birds (70), Life's Too Short Tomatoes (46) — and run the Call #40 variation workflow on those three specifically, instead of spreading the week's 24 new concepts across unrelated ideas.
Kingsolver's turkey-monoculture warning sent the hunt looking for dangerous concentration at IGL, and the ops database says the concentration ended eight months ago — nobody wrote it down, and the strategy file of record still describes the old shape.
Barbara Kingsolver's Chapter 6 fact is that of the 400 million turkeys Americans eat each year, more than 99% are one breed, the Broad-Breasted White, which cannot walk, fly, or reproduce naturally. Chapter 3 prices the arrangement: nonhybrid vegetable varieties in seed catalogs fell from about 5,000 in 1981 to 600 by 1998, with the Irish Potato Famine as the standing warning. Stripped of agriculture, the mechanism is that a portfolio can look productive while being one organism wide, and the narrowness stays invisible until the organism fails.
Read against IGL's System of Record, that lands hard. Section 12 of strategic-synthesis-v4.md reports 2025 as 6,002 units with one shirt at 2,965 of them — 49% — the top two at 55%, and 423 of a 1,200-plus catalog selling at all. The customer-review export agrees from the other side: 215 of 551 reviews sit on that one design. And the mastermind has a live casualty on the shelf: at Call #40 on 2026-07-30, Thomas Knapp came off a record $200,000 month and fell to barely $1,000 in a day, root cause recorded as an entire account riding on five to ten static ads built around a single design concept. Chouinard supplies the founder-shelf version: "Nature doesn't like empires. It doesn't like accumulation in one place. It doesn't like monoculture."
Then the measurement reverses it. fact_product_sales_daily in the ops database reproduces the 2025 book almost exactly — 6,018 units, bird feeder 2,949, 49.0%, 431 products with sales, 122 with five or more — which is the proof that the table and the System of Record are counting the same thing. Run the same query forward and the concentration is gone. Top-design share by month: 68% in January 2025, 49% in May, 20% in November, 56% in the December spike, then 21%, 12%, 20%, 16%, 27%, 15%, 7.0% in July 2026 and 3.4% in August. The bird feeder shirt and all its variants went from 68% of a month's units to 3.0%. The number of designs needed to make up half of all units went from 2 across 2025 to 30 across the trailing 90 days and 34 in August. A 40%-concentration alarm would have fired every single month of 2025 and not once since January.
So the diagnosis inverts, and the inverted version is the useful one. This is not a brand that is dangerously dependent on one shirt; it is a brand that lost its one winner and has not produced another, while total volume fell 14.6% year over year on the same eight-month window (3,758 units Jan–Aug 2025 against 3,211 Jan–Aug 2026). The long tail did not out-compete the bird feeder — it absorbed a smaller business. That is a different problem, with a different fix, and it is the opposite of the one the evidence trail was assembled to warn about.
The one document that called it is the wiki-link note this Librarian wove onto line 559 of the synthesis file: Design Bank Theory read "top 2 designs = 55% of all units" as "what a bank account looks like before the base broadens." The base broadened. Nothing in thirteen knowledge bases records that it happened, and the file carrying the note still leads with the 2025 numbers.
mylibrary-kb/books/animal-vegetable-miracle — primary. Ch.6, the >99% / 400M Broad-Breasted White turkey monoculture; Ch.3, nonhybrid varieties ~5,000 (1981) → 600 (1998) and the Irish Potato Famine as warning. Supplies the mechanism the night went looking for.igl-brand-kb/core/strategic-synthesis-v4 — theme-aggregate. Status: CURRENT SoR, last updated February 11 2026. Section 12 is explicitly headed "Product Performance (2025 Full Year)": 6,002 units, bird feeder 2,965 = 49%, top 2 = 55%, 423 of 1,200+ products with sales, 123 with 5+ sales, 136 with exactly one. Quoted honestly for 2025 — and this is the file the correction targets, because it is read as current.igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox export dated December 11 2025; Top Reviewed Products table, row 1 excuse-me-the-bird-feeders-empty-again at 215 of 551 total reviews, row 2 at 29. Reviews, not sales — the concentration confirmed from customer behaviour.wescale-kb/calls/call-40-insights — primary. Dated 2026-07-30. Thomas Knapp, record $200K month to barely $1,000 in a day, 2.37 ROAS trailing 7 days, "the entire account rode on 5-10 static ads built around a single design concept." Also the Ad Creative Skill workflow (top designs → 20 variations) that step 3 of the move redirects.founders-kb/episodes/060-yvon-chouinard-weve-learned-patagonias-first-40-years — primary. Line 270, verbatim: "Nature doesn't like empires. It doesn't like accumulation in one place. It doesn't like monoculture."G:\iris\igl-ops-sage\data\igl.db, table fact_product_sales_daily — operator measurement, outside the thirteen knowledge bases. 5,921 rows, 2024-07-30 to 2026-08-30. Reproduces the 2025 System-of-Record figures to within 0.3% and then refutes their currency. This is the only source in reach that holds per-product units; the knowledge bases hold none.igl-brand-kb/core/strategic-synthesis-v4.md is marked Status: CURRENT SoR and its Section 12 still reads "Excuse Me, Bird Feeder's Empty -- 2,965 units (49% of all sales)" and "Top 2 designs: 55% of all units sold." Those numbers are labelled 2025 Full Year and they are correct for 2025. They are also the numbers every plan, agent and strategy document in this workspace reads as the shape of the business. Add one line directly beneath that block: "2026 YTD (through Aug 30): top design 9.3% of units, top 2 = 18.3%, 30 designs needed to reach half of units in the trailing 90 days." Until that line exists, every downstream plan is describing a company that stopped existing in January. 2. Add two numbers to the Tuesday scorecard block. The ## WeScale Scorecard Health block in wescale-kb/members/tim-dobyns.md already has an "Inputs This Week" line and already feeds your 3 PM CT Tuesday submission. Add top-design share and designs-to-half beside it. Both come from one query against the ops database you already keep — no hand-reading of Shopify admin, and no dependence on the design_id join that has been broken since mid-July: ``sql -- trailing 90 days, G:\iris\igl-ops-sage\data\igl.db SELECT product_title, SUM(quantity) AS units FROM fact_product_sales_daily WHERE date >= date('now','-90 day') GROUP BY 1 ORDER BY units DESC; -- top-design share = units of row 1 / total -- designs-to-half = how many rows it takes to reach 50% of total ` 3. Change Tuesday's question. The cohort's whole Scale-vs-Fix routine, the Ad Creative Skill workflow from Call #40 (take your top 5 designs, generate 20 variations each), and Alex Wheeler's adapt-the-inside-jokes play all assume there is a winner to work from. At 6.8% of trailing-90-day units, you do not currently have one. So the Fix question this quarter is not "am I too concentrated" — it is "why has nothing replaced the bird feeder." Before Harvest/Fall opens September 27, take the top three of the trailing 90 days by name — Quietly Cheering for Moss (84 units), Listen to the Birds (70), Life's Too Short Tomatoes (46) — and run the Call #40 variation workflow on those three specifically, instead of spreading the week's 24 new concepts across unrelated ideas. ## Why (the one-liner) Kingsolver's turkey-monoculture warning sent the hunt looking for dangerous concentration at IGL, and the ops database says the concentration ended eight months ago — nobody wrote it down, and the strategy file of record still describes the old shape. ## The insight Barbara Kingsolver's Chapter 6 fact is that of the 400 million turkeys Americans eat each year, more than 99% are one breed, the Broad-Breasted White, which cannot walk, fly, or reproduce naturally. Chapter 3 prices the arrangement: nonhybrid vegetable varieties in seed catalogs fell from about 5,000 in 1981 to 600 by 1998, with the Irish Potato Famine as the standing warning. Stripped of agriculture, the mechanism is that a portfolio can look productive while being one organism wide, and the narrowness stays invisible until the organism fails. Read against IGL's System of Record, that lands hard. Section 12 of strategic-synthesis-v4.md reports 2025 as 6,002 units with one shirt at 2,965 of them — 49% — the top two at 55%, and 423 of a 1,200-plus catalog selling at all. The customer-review export agrees from the other side: 215 of 551 reviews sit on that one design. And the mastermind has a live casualty on the shelf: at Call #40 on 2026-07-30, Thomas Knapp came off a record $200,000 month and fell to barely $1,000 in a day, root cause recorded as an entire account riding on five to ten static ads built around a single design concept. Chouinard supplies the founder-shelf version: "Nature doesn't like empires. It doesn't like accumulation in one place. It doesn't like monoculture." Then the measurement reverses it. fact_product_sales_daily in the ops database reproduces the 2025 book almost exactly — 6,018 units, bird feeder 2,949, 49.0%, 431 products with sales, 122 with five or more — which is the proof that the table and the System of Record are counting the same thing. Run the same query forward and the concentration is gone. Top-design share by month: 68% in January 2025, 49% in May, 20% in November, 56% in the December spike, then 21%, 12%, 20%, 16%, 27%, 15%, 7.0% in July 2026 and 3.4% in August. The bird feeder shirt and all its variants went from 68% of a month's units to 3.0%. The number of designs needed to make up half of all units went from 2 across 2025 to 30 across the trailing 90 days and 34 in August. A 40%-concentration alarm would have fired every single month of 2025 and not once since January. So the diagnosis inverts, and the inverted version is the useful one. This is not a brand that is dangerously dependent on one shirt; it is a brand that lost its one winner and has not produced another, while total volume fell 14.6% year over year on the same eight-month window (3,758 units Jan–Aug 2025 against 3,211 Jan–Aug 2026). The long tail did not out-compete the bird feeder — it absorbed a smaller business. That is a different problem, with a different fix, and it is the opposite of the one the evidence trail was assembled to warn about. The one document that called it is the wiki-link note this Librarian wove onto line 559 of the synthesis file: Design Bank Theory read "top 2 designs = 55% of all units" as "what a bank account looks like before the base broadens." The base broadened. Nothing in thirteen knowledge bases records that it happened, and the file carrying the note still leads with the 2025 numbers. ## Evidence trail - mylibrary-kb/books/animal-vegetable-miracle — primary. Ch.6, the >99% / 400M Broad-Breasted White turkey monoculture; Ch.3, nonhybrid varieties ~5,000 (1981) → 600 (1998) and the Irish Potato Famine as warning. Supplies the mechanism the night went looking for. - igl-brand-kb/core/strategic-synthesis-v4 — theme-aggregate. Status: CURRENT SoR, last updated February 11 2026. Section 12 is explicitly headed "Product Performance (2025 Full Year)": 6,002 units, bird feeder 2,965 = 49%, top 2 = 55%, 423 of 1,200+ products with sales, 123 with 5+ sales, 136 with exactly one. Quoted honestly for 2025 — and this is the file the correction targets, because it is read as current. - igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox export dated December 11 2025; Top Reviewed Products table, row 1 excuse-me-the-bird-feeders-empty-again at 215 of 551 total reviews, row 2 at 29. Reviews, not sales — the concentration confirmed from customer behaviour. - wescale-kb/calls/call-40-insights — primary. Dated 2026-07-30. Thomas Knapp, record $200K month to barely $1,000 in a day, 2.37 ROAS trailing 7 days, "the entire account rode on 5-10 static ads built around a single design concept." Also the Ad Creative Skill workflow (top designs → 20 variations) that step 3 of the move redirects. - founders-kb/episodes/060-yvon-chouinard-weve-learned-patagonias-first-40-years — primary. Line 270, verbatim: "Nature doesn't like empires. It doesn't like accumulation in one place. It doesn't like monoculture." - G:\iris\igl-ops-sage\data\igl.db, table fact_product_sales_daily — operator measurement, outside the thirteen knowledge bases. 5,921 rows, 2024-07-30 to 2026-08-30. Reproduces the 2025 System-of-Record figures to within 0.3% and then refutes their currency. This is the only source in reach that holds per-product units; the knowledge bases hold none. ## Assay verdicts - novelty — pass, 3. Judge's own reasoning: the 49% concentration fact is prior art (nuggets #007 and #010, plus a near-identical candidate killed on 2026-08-12), and only the instrument survived. Operator note: after correction the finding is no longer the concentration but its disappearance, which nothing in the vault or the thirteen bases records — the correction raises novelty rather than lowering it. - grounding — pass, 5. Every quotation and every 2025 figure verified verbatim against its source, with periods and CURRENT SoR status carried. Operator note: the judge verified document-truth correctly and, exactly as on nugget #007, could not see that a correctly-quoted 2025 number had stopped describing 2026. - actionability — pass, 4. Two defined metrics with numerator, denominator, window and source, hung on an existing Tuesday block, plus a threshold that flips an existing decision. Judge flagged the 40% gate as arbitrary and "breed the #2 and #3 designs" as the softest sentence — both were removed in correction: the gate would never fire against live data, and the vague verb is replaced by three named designs and a named existing workflow. Thread: animal-vegetable-miracle Ch.6 turkey monoculture + Ch.3 seed-variety collapse → igl-brand-kb strategic-synthesis-v4 §12 (49% of 2025 units on one design) → igl-brand-kb customer-reviews-q4-2025 (215 of 551 reviews, same design) → wescale-kb call-40-insights (Knapp: $200K month to $1,000/day on one design concept) → founders-kb ep.060 Chouinard "nature doesn't like monoculture" → operator query of fact_product_sales_daily`, which reverses the conclusion