This Tuesday, before the 3 PM CT scorecard submission, stop putting the Seedwork generator's output in the cell named "New Designs Launched." Put in the number this query returns, with the two dates set to the Monday and the following Monday of the week you are reporting:
``sql -- G:\iris\igl-ops-sage\data\igl.db SELECT COUNT(*) FROM dim_products WHERE shopify_created_at >= '2026-08-24' AND shopify_created_at < '2026-08-31'; ``
Then keep the generator count — it is a real number, just a different one — on the "Inputs This Week" line of the "WeScale Scorecard Health" block already defined in wescale-kb/members/tim-dobyns.md (line 231), relabelled X designs generated / Y launched so both stay visible and nobody confuses them again. Finally, add a one-line note to your 2026-W31 and 2026-W32 scorecard rows saying those 480s were generator output, not launches.
One repair this depends on: 265 products created since 2026-07-20 — 193 of them in August — have no design_id filled in, so the design table itself can no longer answer "how many designs launched." Until that join is backfilled, count products, not designs.
The one number about IGL's design pipeline that an outsider actually reviews is measuring the machine's output instead of the store's, and the gap between the two is 480 against 6.
Sam Altman's founding-day story is about what you do when the customer feedback loop is not available. Eleven people in Greg Brockman's apartment, no product and no customer for four and a half years. They tried three substitutes and reported the results plainly: public leaderboards worked, external demos for someone the researchers wanted to impress worked, and fake deadlines did not. The distinction is not effort. It is whether an outsider can see the number and know what it means.
IGL has a real leaderboard — forty-plus members typing the same input metrics into one sheet every week, reviewed live by a coach. But the number IGL feeds it is not measuring the store. In the week of 2026-W32 (August 3rd to 9th) IGL reported 480 new designs; the Shopify product table gained 6 products that week. The week before, 480 reported against 123 products. The metric's own name in the scorecard database is "New Designs Launched," its benchmark is 15, and the cohort median those weeks was 25 and 22.
The proof that the cell is meant to hold the store number, not the generator number, is in IGL's own later rows. In scorecard week 47 (August 17th to 23rd) IGL reported 48, and the store gained exactly 48 products. In every earlier week where the design-to-product join is still intact — the weeks of June 8th, June 22nd, July 6th and July 13th — designs and products match one for one: 24 and 24, 48 and 48, 96 and 96, 72 and 72. So the same person, using the same cell, has already reported it correctly four times. The 480s are the anomaly.
Why this needs three shelves: the founder shelf supplies the rule that a leaderboard only works if an outsider can read it, and that a number only you can interpret degrades into a fake deadline you cannot lose on. The mastermind shelf supplies the actual ledger, the metric's real name, and Call #41's evidence that a coach pulls top-8 and bottom-8 stores straight off these rows. The brand shelf supplies the catalogue size — 799 designs in dim_designs, ever — which is what makes 960 designs across a fortnight impossible as a launch count. Any one alone is a number-chase; together they name a broken instrument.
| Scorecard week | Calendar week | Reported to the cohort | Products the store actually gained | Cohort median |
|---|---|---|---|---|
| 44 (call #41) | 2026-W31, Jul 27 – Aug 2 | 480 | 123 | 25 |
| 45 (call #42) | 2026-W32, Aug 3 – 9 | 480 | 6 | 22 |
| 46 (call #43) | 2026-W33, Aug 10 – 16 | 46 | 88 | 20 |
| 47 (call #44) | 2026-W34, Aug 17 – 23 | 48 | 48 — exact match | 35 |
Weeks where the design-to-product join still works, and designs equal products one for one: week of Jun 8 = 24/24 · Jun 22 = 48/48 · Jul 6 = 96/96 · Jul 13 = 72/72.
Current week to report on Tuesday: 2026-W35 (Aug 24 – 30) = 0 products created. The product table was refreshed 2026-08-29 23:03, so that zero is a real reading, not a stale one — check it against what you know you published before you submit it.
founders-kb/interviews/035-sam-altman — the seed. The founding-day substitutes table (lines 202-204): public leaderboards worked, external demos for an eminent person worked, fake deadlines did not work. The doctrine for operating without a feedback loop.wescale-kb/scorecard/wescale-scorecard.db — the cohort's self-reported weekly ledger, file last written 2026-08-29. A record of what members typed, not a sales measurement. Verified by direct query: member 31 is Tim Dobyns / Inspired Garden Life; metric 19's display_name is literally "New Designs Launched" with benchmark_target 15.0; the values for scorecard weeks 44-47 are 480, 480, 46, 48 and the ad-creative counts are 14, 8, 21, 4.wescale-kb/calls/call-41 — call of 2026-08-06. The coach pulled the top 8 and bottom 8 stores by 30-day trend off these rows. Chris Hall, verbatim: "That's a new stat I'm actually tracking — not how many designs I make, but how many designs actually scale." Establishes that an outsider reads these rows and judges on them.wescale-kb/calls/call-43 — call of 2026-08-20. Records a member's passive generation loop at "325 designs / 137 creatives in one week" with only 10-15% of the daily 100 published. Shows the same cell holding different definitions across members. See the correction below on whose row that is.igl-brand-kb/audience/who-data-map — Status: CURRENT SoR, created 2026-08-29, one day old. Names dim_designs at 799 designs and fact_product_sales_daily at 685 sold products, and gives the database path. The lifetime catalogue figure that makes 960 launches in a fortnight impossible.G:\iris\igl-ops-sage\data\igl.db — queried directly by the operator, outside the knowledge bases. dim_designs = 799 rows. dim_products = 1,282 rows, fully refreshed 2026-08-29 23:03, so it is current. Products created per calendar week: W31 123, W32 6, W33 88, W34 48, W35 0.sql -- G:\iris\igl-ops-sage\data\igl.db SELECT COUNT(*) FROM dim_products WHERE shopify_created_at >= '2026-08-24' AND shopify_created_at < '2026-08-31'; ` Then keep the generator count — it is a real number, just a different one — on the "Inputs This Week" line of the "WeScale Scorecard Health" block already defined in wescale-kb/members/tim-dobyns.md (line 231), relabelled X designs generated / Y launched so both stay visible and nobody confuses them again. Finally, add a one-line note to your 2026-W31 and 2026-W32 scorecard rows saying those 480s were generator output, not launches. One repair this depends on: 265 products created since 2026-07-20 — 193 of them in August — have no design_id filled in, so the design table itself can no longer answer "how many designs launched." Until that join is backfilled, count products, not designs. ## Why (the one-liner) The one number about IGL's design pipeline that an outsider actually reviews is measuring the machine's output instead of the store's, and the gap between the two is 480 against 6. ## The insight Sam Altman's founding-day story is about what you do when the customer feedback loop is not available. Eleven people in Greg Brockman's apartment, no product and no customer for four and a half years. They tried three substitutes and reported the results plainly: public leaderboards worked, external demos for someone the researchers wanted to impress worked, and fake deadlines did not. The distinction is not effort. It is whether an outsider can see the number and know what it means. IGL has a real leaderboard — forty-plus members typing the same input metrics into one sheet every week, reviewed live by a coach. But the number IGL feeds it is not measuring the store. In the week of 2026-W32 (August 3rd to 9th) IGL reported 480 new designs; the Shopify product table gained 6 products that week. The week before, 480 reported against 123 products. The metric's own name in the scorecard database is "New Designs Launched," its benchmark is 15, and the cohort median those weeks was 25 and 22. The proof that the cell is meant to hold the store number, not the generator number, is in IGL's own later rows. In scorecard week 47 (August 17th to 23rd) IGL reported 48, and the store gained exactly 48 products. In every earlier week where the design-to-product join is still intact — the weeks of June 8th, June 22nd, July 6th and July 13th — designs and products match one for one: 24 and 24, 48 and 48, 96 and 96, 72 and 72. So the same person, using the same cell, has already reported it correctly four times. The 480s are the anomaly. Why this needs three shelves: the founder shelf supplies the rule that a leaderboard only works if an outsider can read it, and that a number only you can interpret degrades into a fake deadline you cannot lose on. The mastermind shelf supplies the actual ledger, the metric's real name, and Call #41's evidence that a coach pulls top-8 and bottom-8 stores straight off these rows. The brand shelf supplies the catalogue size — 799 designs in dim_designs, ever — which is what makes 960 designs across a fortnight impossible as a launch count. Any one alone is a number-chase; together they name a broken instrument. ## Evidence trail - founders-kb/interviews/035-sam-altman — the seed. The founding-day substitutes table (lines 202-204): public leaderboards worked, external demos for an eminent person worked, fake deadlines did not work. The doctrine for operating without a feedback loop. - wescale-kb/scorecard/wescale-scorecard.db — the cohort's self-reported weekly ledger, file last written 2026-08-29. A record of what members typed, not a sales measurement. Verified by direct query: member 31 is Tim Dobyns / Inspired Garden Life; metric 19's display_name is literally "New Designs Launched" with benchmark_target 15.0; the values for scorecard weeks 44-47 are 480, 480, 46, 48 and the ad-creative counts are 14, 8, 21, 4. - wescale-kb/calls/call-41 — call of 2026-08-06. The coach pulled the top 8 and bottom 8 stores by 30-day trend off these rows. Chris Hall, verbatim: "That's a new stat I'm actually tracking — not how many designs I make, but how many designs actually scale." Establishes that an outsider reads these rows and judges on them. - wescale-kb/calls/call-43 — call of 2026-08-20. Records a member's passive generation loop at "325 designs / 137 creatives in one week" with only 10-15% of the daily 100 published. Shows the same cell holding different definitions across members. See the correction below on whose row that is. - igl-brand-kb/audience/who-data-map — Status: CURRENT SoR, created 2026-08-29, one day old. Names dim_designs at 799 designs and fact_product_sales_daily at 685 sold products, and gives the database path. The lifetime catalogue figure that makes 960 launches in a fortnight impossible. - G:\iris\igl-ops-sage\data\igl.db — queried directly by the operator, outside the knowledge bases. dim_designs = 799 rows. dim_products = 1,282 rows, fully refreshed 2026-08-29 23:03, so it is current. Products created per calendar week: W31 123, W32 6, W33 88, W34 48, W35 0. ## Worked examples | Scorecard week | Calendar week | Reported to the cohort | Products the store actually gained | Cohort median | |---|---|---|---|---| | 44 (call #41) | 2026-W31, Jul 27 – Aug 2 | 480 | 123 | 25 | | 45 (call #42) | 2026-W32, Aug 3 – 9 | 480 | 6 | 22 | | 46 (call #43) | 2026-W33, Aug 10 – 16 | 46 | 88 | 20 | | 47 (call #44) | 2026-W34, Aug 17 – 23 | 48 | 48 — exact match | 35 | Weeks where the design-to-product join still works, and designs equal products one for one: week of Jun 8 = 24/24 · Jun 22 = 48/48 · Jul 6 = 96/96 · Jul 13 = 72/72. Current week to report on Tuesday: 2026-W35 (Aug 24 – 30) = 0 products created. The product table was refreshed 2026-08-29 23:03, so that zero is a real reading, not a stale one — check it against what you know you published before you submit it. ## Assay verdicts - novelty — PASS (4/5). No prior art for the claim. The 480 figure surfaced once before as a killed candidate (nuggets/logs/2026-08-28.md:168) attached to a different move about ad counts, never as a metric-definition defect. Advice-layer check clean: tim-dobyns.md prescribes submitting the scorecard and tracking design hit rate, never what unit the number is in. - grounding — PASS (4/5), after operator correction. Every scorecard figure reproduced by direct query. Two claims were wrong and are corrected below. - actionability — PASS (5/5). Hangs on a cadence Tim already runs every Tuesday, changes one cell, and the replacement number is a two-line query whose answer is already computed above. Thread: Altman's substitutes table (leaderboards work, fake deadlines don't) → the scorecard database rows for member 31, weeks 44-47 → the metric's own name, "New Designs Launched" → call-41, where a coach reads these rows to rank stores → call-43, where members report generator output in the same cell → who-data-map's 799 lifetime designs → the operator's own query of dim_products`, which supplies the number that should have been in the cell all along.Step 0 (5 minutes, do this first). Find out what Loox actually does to a review when its Shopify product is deleted. Nothing in any of the thirteen knowledge bases says — check Loox's own help pages, or delete one worthless test product that has a review and watch what happens to it. Everything below assumes the answer is bad; if it turns out Loox keeps them and re-attaches on re-import, you have spent five minutes and can purge freely.
Then, before deleting anything: export the current Loox reviews CSV and build a do-not-delete list of every product handle carrying at least one review. Reviews attach to product handles, so a handle that disappears takes its social proof with it — and IGL's proof is unusually concentrated. In the 11 December 2025 export, 314 of 537 active reviews sit on ten handles and 215 sit on one. That leaves roughly 223 reviews spread thinly across the tail, which is exactly the part of the catalog a low-sales sweep aims at. For any handle you still want gone, re-import its reviews against the surviving parent design in the Loox CSV importer first, then delete.
Hard-protect excuse-me-bird-feeders-empty-long-sleeve-3330043 (9 reviews). It is the long-sleeve spin-off of the 215-review flagship, and a "no sale in 90 days" sweep is precisely the thing that kills a slow-moving spin-off of a winner.
Three separate people in the mastermind worked out the mechanics of purging a catalog to free Shopify variant slots, and not one of them said what the purge does to the reviews those products are carrying.
Arnault's argument is that the brand, not the product, is the compounding asset: a fashion house is worth what it is worth because of accumulated equity, and the current product line is only that equity's present expression. Reviews are the one piece of accumulated equity IGL owns in a form that is countable and visible to a stranger — a shopper who has never heard of the brand reads them and decides.
The mastermind's prune instructions are technically correct and silent on this. Chris Heckman's Call #17 teaching is explicit: archived designs still count against the Shopify variant ceiling, so you have to delete rather than archive. Tim asked in that same chat whether archiving was enough, and Chris confirmed it was not — naming the cost as losing the product's sales history. Reviews were never mentioned. Alex Wheeler then removed every design that sold under three times, about 60% of his catalog, and Steve Brigden answered with a "last call" campaign to monetize the tail before deleting it. Three people, three careful passes at the sequence, one blind spot shared by all of them.
It is a live job, not a hypothetical: Call #34 on 2026-06-18 lists "prune 90-day-unsold" as a standing answer to the variant ceiling, and the Wheeler and Brigden posts are both dated 2026-07-29.
No single shelf reaches this. The founders episode supplies the reason accumulated equity outranks current inventory; the mastermind supplies the delete-don't-archive mechanic and the 60% benchmark; only IGL's own review export shows the concentration that turns routine hygiene into a self-inflicted wound.
founders-kb/episodes/296-bernard-arnault-richest-man-in-the-world — Seed. The premise that a brand is a compounding machine and the product only its expression; "A powerful brand is magic" (line 52), luxury as "the only area in which it is possible to make luxury profit margins" (line 16). Frames why destroying accumulated proof to reclaim variant slots is the wrong trade.wescale-kb/calls/call-17 — Chris Heckman's Shopify variant teaching, lines 125-129: purge products with no sale in 90 days and fewer than 5 lifetime sales; archived designs still count, so delete rather than archive; Tim asked in chat and Chris confirmed deletion removes the product's sales history. Date is imprecise — the frontmatter reads "Late Cohort (Weeks 17-19)", not a calendar date. Coaching, not a commitment. Reviews are never mentioned, which is the finding.wescale-kb/members/alex-wheeler — Slack post dated 2026-07-29 (line 158): "I recently removed every design that sold less than 3 times from my store, about 60%." The threshold and scale IGL would be copying.wescale-kb/members/steve-brigden — Slack post dated 2026-07-29 (lines 141-143): build a collection of every never-sold product, run a "last call" campaign, then re-imagine the survivors. Shows the sequence was worked carefully and still skipped reviews.wescale-kb/calls/call-34 — Dated 2026-06-18, line 32: the 1,000-variant pressure item, "prune 90-day-unsold, or narrow to top 2-3 colors," with catalog pruning inside members' 72-hour commitments. Evidence the prune is near-term, not hypothetical.igl-brand-kb/community/customer-reviews-q4-2025 — Loox export dated 11 December 2025, ~8.5 months old, no Status field, covering 2024-08-31 to 2025-12-09. 551 total / 537 active reviews, 51 with photos, 24 story-rich. Top-ten table sums to 314, with 215 on excuse-me-the-bird-feeders-empty-again. Supplies every handle above. Review counts, not sales.excuse-me-bird-feeders-empty-long-sleeve-3330043 (9 reviews). It is the long-sleeve spin-off of the 215-review flagship, and a "no sale in 90 days" sweep is precisely the thing that kills a slow-moving spin-off of a winner. ## Why (the one-liner) Three separate people in the mastermind worked out the mechanics of purging a catalog to free Shopify variant slots, and not one of them said what the purge does to the reviews those products are carrying. ## The insight Arnault's argument is that the brand, not the product, is the compounding asset: a fashion house is worth what it is worth because of accumulated equity, and the current product line is only that equity's present expression. Reviews are the one piece of accumulated equity IGL owns in a form that is countable and visible to a stranger — a shopper who has never heard of the brand reads them and decides. The mastermind's prune instructions are technically correct and silent on this. Chris Heckman's Call #17 teaching is explicit: archived designs still count against the Shopify variant ceiling, so you have to delete rather than archive. Tim asked in that same chat whether archiving was enough, and Chris confirmed it was not — naming the cost as losing the product's sales history. Reviews were never mentioned. Alex Wheeler then removed every design that sold under three times, about 60% of his catalog, and Steve Brigden answered with a "last call" campaign to monetize the tail before deleting it. Three people, three careful passes at the sequence, one blind spot shared by all of them. It is a live job, not a hypothetical: Call #34 on 2026-06-18 lists "prune 90-day-unsold" as a standing answer to the variant ceiling, and the Wheeler and Brigden posts are both dated 2026-07-29. No single shelf reaches this. The founders episode supplies the reason accumulated equity outranks current inventory; the mastermind supplies the delete-don't-archive mechanic and the 60% benchmark; only IGL's own review export shows the concentration that turns routine hygiene into a self-inflicted wound. ## Operator corrections applied Four things the hunt and all three judges let through, corrected here: 1. The load-bearing mechanism is unverified. "Deleting a product deletes its reviews" is stated nowhere in any of the thirteen knowledge bases — it is plausible platform behavior, not a documented fact. It has been demoted from the headline into a Step 0 check. The finding that survives regardless is the silence: the cohort's purge instructions never raise the question at all. 2. Review count is not sales count. The candidate argued that nine named handles "sit below the top-10 review table, which is exactly the fewer-than-5-lifetime-sales band." That does not follow. A product with a review had at least one sale, and ordinary review-capture rates mean it probably had several. Those handles are a look-up list, not established prune candidates. 3. Every product handle in the export is truncated (each ends in ...). Eight of the nine carry their numeric Shopify product ID, so they resolve. colorful-vintage-floral-geometric-patter... (line 183) does not — it has no ID and cannot be resolved from this file. Look it up in Shopify. 4. The export is 8.5 months old. Dated 2025-12-11, covering 2024-08-31 to 2025-12-09, with no Status field. Every count here is a quotation from a dated document, never a measurement. There is no product-level sales data anywhere in the thirteen knowledge bases, so nothing here can tell you which products are actually in the prune band — only Shopify can. ## The handles to look up before any sweep These carry story-rich or quoted reviews and are not in the top-ten table, so the export cannot tell you how many reviews each holds beyond "at least one, fewer than six": | Handle | Why it matters | |---|---| | excuse-me-bird-feeders-empty-long-sleeve-3330043 | 9 reviews, rank 5; long-sleeve spin-off of the 215-review flagship | | seed-hoarder-definition-racerback-3338355 | Lydia N.'s 5-star — the best fabric/fit testimonial in the file ("not constantly pulling on the straps") | | waters-edge-bird-lovers-t-shirt-3678494 | Janet F.'s repeat-buyer review ("maybe 10 or so" shirts owned); also Leslie S. | | easily-distracted-by-tomatoes-tank-top-3338942 | Story-rich review | | tomato-lovers-collective-v-neck-3334630 | Story-rich review | | tomato-season-way-of-life-long-sleeve-3330035 | Story-rich review | | tomato-season-way-of-life-t-shirt-2478073 | Story-rich review | | aphid-assassin-t-shirt-2242689 | Story-rich review | | heirloom-heat-t-shirt-3698796 | becca c.'s 5-star | | colorful-vintage-floral-geometric-patter... | Joseph T.'s 5-star — handle truncated with no product ID; resolve in Shopify | ## Evidence trail - founders-kb/episodes/296-bernard-arnault-richest-man-in-the-world — Seed. The premise that a brand is a compounding machine and the product only its expression; "A powerful brand is magic" (line 52), luxury as "the only area in which it is possible to make luxury profit margins" (line 16). Frames why destroying accumulated proof to reclaim variant slots is the wrong trade. - wescale-kb/calls/call-17 — Chris Heckman's Shopify variant teaching, lines 125-129: purge products with no sale in 90 days and fewer than 5 lifetime sales; archived designs still count, so delete rather than archive; Tim asked in chat and Chris confirmed deletion removes the product's sales history. Date is imprecise — the frontmatter reads "Late Cohort (Weeks 17-19)", not a calendar date. Coaching, not a commitment. Reviews are never mentioned, which is the finding. - wescale-kb/members/alex-wheeler — Slack post dated 2026-07-29 (line 158): "I recently removed every design that sold less than 3 times from my store, about 60%." The threshold and scale IGL would be copying. - wescale-kb/members/steve-brigden — Slack post dated 2026-07-29 (lines 141-143): build a collection of every never-sold product, run a "last call" campaign, then re-imagine the survivors. Shows the sequence was worked carefully and still skipped reviews. - wescale-kb/calls/call-34 — Dated 2026-06-18, line 32: the 1,000-variant pressure item, "prune 90-day-unsold, or narrow to top 2-3 colors," with catalog pruning inside members' 72-hour commitments. Evidence the prune is near-term, not hypothetical. - igl-brand-kb/community/customer-reviews-q4-2025 — Loox export dated 11 December 2025, ~8.5 months old, no Status field, covering 2024-08-31 to 2025-12-09. 551 total / 537 active reviews, 51 with photos, 24 story-rich. Top-ten table sums to 314, with 215 on excuse-me-the-bird-feeders-empty-again. Supplies every handle above. Review counts, not sales. ## Assay verdicts - novelty: PASS · 4/5 — grepped every knowledge base, CATALOG.md, all 23 prior catalog files and every night log for prune/purge/delete crossed with review/Loox; zero hits. A real constraint bolted onto an assigned build. - grounding: PASS · 5/5 — strongest in the batch; the key evidence turned out stronger than the candidate claimed (call-17.md:129 carries Tim's own chat question). Independently re-verified by the operator: top-ten sum = 314 ✅, 537 − 314 = 223 ✅, all quotations verbatim ✅, one handle unresolvable ⚠️, deletion mechanism undocumented ⚠️. - actionability: PASS · 5/5 — inserts one cheap, reversible step in front of an irreversible action the cohort is actively recommending. - Summed score 14/15. Tonight's second 3-of-3 survivor scored 12/15 and is logged as a SEED: line, not catalogued. Thread: ep-296 Arnault (brand equity compounds; the product is only its expression) → call-17 Chris Heckman "delete, not archive" plus Tim's own chat question → alex-wheeler 2026-07-29 pruned 60% of his catalog → steve-brigden 2026-07-29 last-call-before-delete → call-34 2026-06-18 prune is a live 72-hour commitment → igl-brand-kb customer-reviews-q4-2025 (215 of 537 reviews on one handle; ~223 in the tail a sweep targets)Before you touch any of the 22 force-ranked fixes in wescale-kb/reports/igl-cro-plan-2026-08-06.md, run that plan's own Stage 0 once — but record five numbers over one fixed 30-day Shopify window instead of the four it asks for: add-to-cart %, reach-checkout %, Shopify's reported conversion rate, orders ÷ sessions, and the share of orders paid through Shop Pay / Apple Pay / PayPal express. The fifth number is there because wescale-kb/members/samantha-wright.md:84 already documents the mechanism that makes two honest conversion figures disagree: Shopify counts sessions that completed the checkout step, and an express checkout can complete an order without ever registering that step.
Then write the result into the plan's Open items line (line 208), which today says IGL's add-to-cart and reach-checkout percentages "are not in any knowledge base." They are — in two files the plan never cites. And re-run the payback table on lines 184-185 with the measured number in place of 1.6, because that table is what justifies every budget decision downstream of it.
Time: 30 minutes in Shopify analytics, plus one edit to the plan. No tool purchase, no staff.
The plan that governs all of IGL's conversion work is priced off a conversion rate quoted in a mastermind call on 2026-05-14, while a July 8 file — and a second copy of it — records Tim saying the number out loud as 2-2.25%, and the plan cites neither.
There are two conversion rates for IGL sitting in wescale-kb, roughly eight weeks apart, in files that never reference each other.
The 1.6% comes from Call #29, 2026-05-14 (wescale-kb/calls/call-29.md:134: "IGL's conversion is ~1.6% vs. a 2.2% target"), and is restated twice in wescale-kb/members/tim-dobyns.md (lines 71 and 83). It is the number the CRO plan's payback table is built on: +0.5pp (1.6 → 2.1%) | ~$3,700 | 33× and +1.0pp (1.6 → 2.6%) | ~$7,500 | 67× (lines 184-185). The plan declares itself Status: CURRENT SoR for IGL conversion work.
The 2-2.25% is Tim's own spoken estimate on 2026-07-08, recorded in two places: wescale-kb/reviews/jo-s-1on1-tim-2026-07-08.md:63 ("Tim's stated funnel problem: conversion rate ~2-2.25%, add-to-cart only 4-5%") and wescale-kb/planning/igl-ads-jo-s-insights-2026-07-08.md:80 ("IGL funnel today: CR ~2-2.25%, ATC 4-5%"). Grepping the August plan for 1on1, 2-2.25 or 4-5% returns nothing.
These are not necessarily contradictory — and that is the actual problem. Eight weeks separate them; the store may simply have improved. Neither number carries a measurement window, neither is a live figure, and nobody wrote down which one is current. So the plan is not built on a wrong number, it is built on a number of unrecorded vintage — and it says in its own closing line that the supporting figures do not exist, while two files hold them.
The founder material supplies the diagnostic frame. founders-kb/interviews/024-ed-catmull.md:34 names the failure mechanically: "you go to your first conclusion and you stop thinking… you've got to keep peeling away the layers." And Catmull's other observation on line 54 is the twist that fits here exactly — "there was nothing that Steve ever said that had not been said by somebody else before." The correction was already in the room, spoken by Tim himself, four weeks before the plan was written. The Howard Hughes episode is the seed and the general form: founders-kb/episodes/059-howard-hughes-the-definitive-biography.md:34 is Senra on the widest gap he has ever found between a founder's mythology and his reality. For a solo operator that gap is not reputational, it is numerical — the number you believe about your own store versus the number that is true.
Why no single shelf could produce it: founders-kb supplies the frame, wescale-kb holds both figures in files that never meet plus the technical mechanism that explains a divergence, and igl-brand-kb's review corpus shows IGL has precisely the multi-order buyer population that a session-denominated metric mismeasures.
founders-kb/episodes/059-howard-hughes-the-definitive-biography — primary, the seed. Line 34: the widest myth-vs-reality gap Senra has covered. Episode note, no status header; a founder record, not a business record.founders-kb/interviews/024-ed-catmull — primary, interview dated 2026-06-14. Line 34 ("you go to your first conclusion and you stop thinking"); line 54 (the same note only lands when it comes from outside).wescale-kb/calls/call-29 — primary, call dated 2026-05-14, no status header. Line 134 is the origin of the 1.6% figure: "IGL's conversion is ~1.6% vs. a 2.2% target." Three months old.wescale-kb/reviews/jo-s-1on1-tim-2026-07-08 — primary, dated 2026-07-08 in frontmatter, no status header. Line 63 records "conversion rate ~2-2.25%, add-to-cart only 4-5%" — a verbal estimate with no period attached. Corroborated at wescale-kb/planning/igl-ads-jo-s-insights-2026-07-08.md:80.wescale-kb/members/samantha-wright — primary, member profile, metric rows individually dated. Line 84 carries the mechanism: Shopify conversion rate = sessions that completed checkout ÷ total sessions; express checkouts can complete an order without registering the step; bots inflate the denominator.igl-brand-kb/community/customer-reviews-q4-2025 — primary, Loox export dated 2025-12-11, 551 reviews spanning 2024-08-31 to 2025-12-09, no status header. Frances Lamont: "I have bought 6 shirts so far (3 orders total)"; Janet F.: "maybe 10 or so." The multi-order population a session-based rate mismeasures.wescale-kb/reports/igl-cro-plan-2026-08-06 — theme-aggregate, built 2026-08-06, Status: CURRENT SoR. Payback table lines 184-185; Stage 0 threshold table lines 82-87; Open items line 208.wescale-kb/reports/igl-cro-plan-2026-08-06.md, run that plan's own Stage 0 once — but record five numbers over one fixed 30-day Shopify window instead of the four it asks for: add-to-cart %, reach-checkout %, Shopify's reported conversion rate, orders ÷ sessions, and the share of orders paid through Shop Pay / Apple Pay / PayPal express. The fifth number is there because wescale-kb/members/samantha-wright.md:84 already documents the mechanism that makes two honest conversion figures disagree: Shopify counts sessions that completed the checkout step, and an express checkout can complete an order without ever registering that step. Then write the result into the plan's Open items line (line 208), which today says IGL's add-to-cart and reach-checkout percentages "are not in any knowledge base." They are — in two files the plan never cites. And re-run the payback table on lines 184-185 with the measured number in place of 1.6, because that table is what justifies every budget decision downstream of it. Time: 30 minutes in Shopify analytics, plus one edit to the plan. No tool purchase, no staff. ## Why (the one-liner) The plan that governs all of IGL's conversion work is priced off a conversion rate quoted in a mastermind call on 2026-05-14, while a July 8 file — and a second copy of it — records Tim saying the number out loud as 2-2.25%, and the plan cites neither. ## The insight There are two conversion rates for IGL sitting in wescale-kb, roughly eight weeks apart, in files that never reference each other. The 1.6% comes from Call #29, 2026-05-14 (wescale-kb/calls/call-29.md:134: "IGL's conversion is ~1.6% vs. a 2.2% target"), and is restated twice in wescale-kb/members/tim-dobyns.md (lines 71 and 83). It is the number the CRO plan's payback table is built on: +0.5pp (1.6 → 2.1%) | ~$3,700 | 33× and +1.0pp (1.6 → 2.6%) | ~$7,500 | 67× (lines 184-185). The plan declares itself Status: CURRENT SoR for IGL conversion work. The 2-2.25% is Tim's own spoken estimate on 2026-07-08, recorded in two places: wescale-kb/reviews/jo-s-1on1-tim-2026-07-08.md:63 ("Tim's stated funnel problem: conversion rate ~2-2.25%, add-to-cart only 4-5%") and wescale-kb/planning/igl-ads-jo-s-insights-2026-07-08.md:80 ("IGL funnel today: CR ~2-2.25%, ATC 4-5%"). Grepping the August plan for 1on1, 2-2.25 or 4-5% returns nothing. These are not necessarily contradictory — and that is the actual problem. Eight weeks separate them; the store may simply have improved. Neither number carries a measurement window, neither is a live figure, and nobody wrote down which one is current. So the plan is not built on a wrong number, it is built on a number of unrecorded vintage — and it says in its own closing line that the supporting figures do not exist, while two files hold them. The founder material supplies the diagnostic frame. founders-kb/interviews/024-ed-catmull.md:34 names the failure mechanically: "you go to your first conclusion and you stop thinking… you've got to keep peeling away the layers." And Catmull's other observation on line 54 is the twist that fits here exactly — "there was nothing that Steve ever said that had not been said by somebody else before." The correction was already in the room, spoken by Tim himself, four weeks before the plan was written. The Howard Hughes episode is the seed and the general form: founders-kb/episodes/059-howard-hughes-the-definitive-biography.md:34 is Senra on the widest gap he has ever found between a founder's mythology and his reality. For a solo operator that gap is not reputational, it is numerical — the number you believe about your own store versus the number that is true. Why no single shelf could produce it: founders-kb supplies the frame, wescale-kb holds both figures in files that never meet plus the technical mechanism that explains a divergence, and igl-brand-kb's review corpus shows IGL has precisely the multi-order buyer population that a session-denominated metric mismeasures. ## Evidence trail - founders-kb/episodes/059-howard-hughes-the-definitive-biography — primary, the seed. Line 34: the widest myth-vs-reality gap Senra has covered. Episode note, no status header; a founder record, not a business record. - founders-kb/interviews/024-ed-catmull — primary, interview dated 2026-06-14. Line 34 ("you go to your first conclusion and you stop thinking"); line 54 (the same note only lands when it comes from outside). - wescale-kb/calls/call-29 — primary, call dated 2026-05-14, no status header. Line 134 is the origin of the 1.6% figure: "IGL's conversion is ~1.6% vs. a 2.2% target." Three months old. - wescale-kb/reviews/jo-s-1on1-tim-2026-07-08 — primary, dated 2026-07-08 in frontmatter, no status header. Line 63 records "conversion rate ~2-2.25%, add-to-cart only 4-5%" — a verbal estimate with no period attached. Corroborated at wescale-kb/planning/igl-ads-jo-s-insights-2026-07-08.md:80. - wescale-kb/members/samantha-wright — primary, member profile, metric rows individually dated. Line 84 carries the mechanism: Shopify conversion rate = sessions that completed checkout ÷ total sessions; express checkouts can complete an order without registering the step; bots inflate the denominator. - igl-brand-kb/community/customer-reviews-q4-2025 — primary, Loox export dated 2025-12-11, 551 reviews spanning 2024-08-31 to 2025-12-09, no status header. Frances Lamont: "I have bought 6 shirts so far (3 orders total)"; Janet F.: "maybe 10 or so." The multi-order population a session-based rate mismeasures. - wescale-kb/reports/igl-cro-plan-2026-08-06 — theme-aggregate, built 2026-08-06, Status: CURRENT SoR. Payback table lines 184-185; Stage 0 threshold table lines 82-87; Open items line 208. ## Assay verdicts - Novelty — PASS (4). Verified the contradiction independently and found it stronger than claimed: two files hold the "missing" number, not one. No nugget covers it; #013 is pop-up rates, a different metric. The express-checkout mechanism in samantha-wright.md has never been applied to IGL. - Grounding — PASS (5). Reproduced every quote and line number; confirmed by its own grep that the two documents never reference each other. - Consequence — PASS (5). Changes the ranking of 22 fixes and the budget attached to them. Five named numbers, one fixed window, no staff or capital. Honest that neither figure is a measurement it can make. ## Operator corrections applied before shipping Four. The first two change what the nugget claims. 1. The 1.6% is neither the plan's own invention nor simultaneous with the July figure. The hunter and all three judges treated the two numbers as a same-period contradiction. It traces to Call #29, 2026-05-14 (call-29.md:134), restated at tim-dobyns.md:71,83 — roughly eight weeks before Tim's 2-2.25%. The honest claim is not "your plan contradicts you," it is "your plan is priced off a May figure you may have superseded in July, and nobody recorded which is current." Headline rewritten accordingly. 2. "1.6% reads as a whole-funnel leak" is wrong by the plan's own table. Stage 0 sets the sessions-converted red flag at under 1.5%; 1.6% clears it. Likewise 4-5% add-to-cart clears the under-4% red flag. Neither figure trips a red flag — both sit below target (2.5-3% conversion, 6-8% add-to-cart). The candidate and the consequence judge both framed the fork as two different diagnostic categories; it is a difference of degree. The sharper signal in Tim's own July numbers is the add-to-cart gap, which points at the product page. 3. The payback shrinks but the plan's spending verdict survives. At a 2.1% baseline, +0.5pp is a 24% relative lift on ~$11,950/month ≈ $2,850, not the table's ~$3,700 — roughly a quarter less. Still 25× the $112/month tool cost, so the plan's "the tool cost is never the issue" conclusion holds. What moves is the ranking of the free fixes, not the buy/don't-buy call. 4. Two different conversion targets are in circulation. Call #29 and tim-dobyns.md use a 2.2% target; the CRO plan's Stage 0 uses the cohort's 2.5-3%. Worth settling in the same 30 minutes. Honest limit: no knowledge base here holds live Shopify or product-level sales data. Both figures are quotations from documents, not measurements. The reconciliation has to happen in Shopify. Thread: ep-059 Howard Hughes (myth vs reality) → Catmull interview #24 (peel the layers) → reviews/jo-s-1on1-tim-2026-07-08 (Tim's own 2-2.25% / 4-5%) → reports/igl-cro-plan-2026-08-06 (CURRENT SoR, priced off 1.6, declares those numbers missing) → calls/call-29 (where the 1.6 actually came from, 2026-05-14) → members/samantha-wright (how the session denominator breaks) → igl-brand-kb community/customer-reviews-q4-2025 (the multi-order buyers)Step 0, five minutes, before anything else: open Printify and confirm which blank IGL actually prints on today, because two of your own files disagree — a December 2025 customer review praises "high quality Gildan soft shirts," while the March 2026 Happy Plant People teardown lists IGL's blanks as "Comfort Colors (via Printify)." The Call #32 action item only makes sense if you are not already there.
Then: cross the Call #32 action item — "Put your top 10–20 best sellers on Comfort Colors as a separate premium SKU" (wescale-kb/calls/call-32.md:103, restated in your own to-do list at wescale-kb/members/tim-dobyns.md:80) — off the list, and run the same premium-second-SKU experiment on a lightweight blank instead: Bella Canvas 3001 or Gildan 6400. That is the exact pair Jo Conlon (Happy Plant People, the garden-niche member closest to IGL) sells as "Lightweight" beside her Comfort Colors "Heavyweight," priced the same, customer chooses. You already have a Bella Canvas relationship — the 3500 V-neck is in your own email templates.
Before you list anything: have Derylin order one current-blank tee and one candidate blank in her own size from Printify, wear and wash both in her garden, and measure body length on each. Janet F. (5★, ~10 shirts) says IGL shirts run "about 5 inches too long" on her. Pick the shorter body and publish the measured length in the product description.
The mastermind's most-repeated action item — move best sellers to heavyweight Comfort Colors — pushes IGL in the exact direction its own customer reviews are pushing back against.
Kelly Johnson's refusals were mostly refusals of other people's confident advice. He told two lead designers their airplane was unstable on his first assignment, went over NACA's head for wind-tunnel time, returned government money on work he didn't believe in, and scrapped his own hydrogen-powered plane when it proved impractical. Rule 9 of his 14: the contractor must flight-test his own product, or he "rapidly loses his competency to design other vehicles." IGL has never flight-tested a blank. It has never refused a cohort prescription either.
The Comfort Colors migration is the most-repeated action item in wescale-kb — Call #32's action list, Pati Heinzl's write-up, Tim's own IGL-applications notes. All of it traces to one member's one week. And Call #32 states its own mechanism as a hypothesis, not a finding: the same designs "look like new products in a CC mockup (brighter, more vivid) — the previous mockup may have been holding them back, not the designs." The honest counter-signal, which belongs in the same breath: that call also says her shirts now sell "better than the Gildan versions." So the cohort evidence is one uncontrolled week with a stated alternative explanation — not nothing, but not a fit finding, and nobody in that chain checked it against IGL's customers.
IGL's customers have already voted, in writing, in the opposite direction. The Loox export (551 reviews, 537 active, 88.2% five-star, covering 2024-08-31 to 2025-12-09) contains four independent praises of thinness and breathability in heat, and four fit complaints that all point the same way: too long, too oversized, armholes too big. The file names it in its own pain-point list: "Shirt length (multiple 'too long' mentions)." Comfort Colors 1717 is the heavyweight, boxy, garment-dyed option — 6.1 oz, per IGL's own Sloth Hiking Club teardown. Jo Conlon classifies it that way herself and notes her warm-climate customers always need the lightweight option. IGL's buyers are 68–70% female with a core age of 55–75+, gardening outdoors through a Missouri summer.
No single shelf produces this. The founders shelf supplies the permission to refuse; the mastermind shelf supplies both the prescription and the garden-niche counter-example; the brand shelf supplies the customer sentences that decide it.
founders-kb/episodes/419-kelly-johnson-skunk-works — primary, seed. Founder episode, no status flag. Rule 9 (flight-test your own product or lose competency), the returned government money, the killed hydrogen plane.igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox review export, export and processing date 2025-12-11, no status header. 8½ months old at the time of writing — old enough that you should confirm nothing changed, but this is review text, not sales data. Source of all eight quotes above, the summary statistics, and the file's own "Shirt length (multiple 'too long' mentions)" pain point.wescale-kb/calls/call-32 — primary, 2026-06-04. The prescription itself at line 103. Also carries its own limits: line 12 states no transcript or chat log was published for Call #32, so it is reconstructed from the debrief, and Pati's $13–14K → $46K is a self-reported 7-day figure with a hypothesised mockup-brightness mechanism.wescale-kb/members/jo-conlon — primary. The garden-niche peer. Slack, 2026-02-12: "I differentiate Bella Canvas and Comfort Colors in terms of Lightweight and Heavyweight garments, both premium products… my core customers are in Florida, California & Texas and lightweight tees and tanks are always going to be good to have as an option."igl-brand-kb/competitive/happy-plant-people-email-analysis — primary, 2026-03-17. Added by the operator; the hunt never opened it. Its comparison table lists IGL's blanks as "Comfort Colors (via Printify)" — contradicting the Gildan review quote and making Step 0 necessary.igl-brand-kb/core/buyer-profile-deep-dive — theme-aggregate. Status: CURRENT SoR, last updated February 2026, data dated 2026-01-01. ~1,000-profile Klaviyo sample (~22% of the base): 68–70% female, core age 55–75+ inferred from name-cohort and email-domain patterns, not verified.wescale-kb/calls/call-32.md:103, restated in your own to-do list at wescale-kb/members/tim-dobyns.md:80) — off the list, and run the same premium-second-SKU experiment on a lightweight blank instead: Bella Canvas 3001 or Gildan 6400. That is the exact pair Jo Conlon (Happy Plant People, the garden-niche member closest to IGL) sells as "Lightweight" beside her Comfort Colors "Heavyweight," priced the same, customer chooses. You already have a Bella Canvas relationship — the 3500 V-neck is in your own email templates. Before you list anything: have Derylin order one current-blank tee and one candidate blank in her own size from Printify, wear and wash both in her garden, and measure body length on each. Janet F. (5★, ~10 shirts) says IGL shirts run "about 5 inches too long" on her. Pick the shorter body and publish the measured length in the product description. ## Why (the one-liner) The mastermind's most-repeated action item — move best sellers to heavyweight Comfort Colors — pushes IGL in the exact direction its own customer reviews are pushing back against. ## The insight Kelly Johnson's refusals were mostly refusals of other people's confident advice. He told two lead designers their airplane was unstable on his first assignment, went over NACA's head for wind-tunnel time, returned government money on work he didn't believe in, and scrapped his own hydrogen-powered plane when it proved impractical. Rule 9 of his 14: the contractor must flight-test his own product, or he "rapidly loses his competency to design other vehicles." IGL has never flight-tested a blank. It has never refused a cohort prescription either. The Comfort Colors migration is the most-repeated action item in wescale-kb — Call #32's action list, Pati Heinzl's write-up, Tim's own IGL-applications notes. All of it traces to one member's one week. And Call #32 states its own mechanism as a hypothesis, not a finding: the same designs "look like new products in a CC mockup (brighter, more vivid) — the previous mockup may have been holding them back, not the designs." The honest counter-signal, which belongs in the same breath: that call also says her shirts now sell "better than the Gildan versions." So the cohort evidence is one uncontrolled week with a stated alternative explanation — not nothing, but not a fit finding, and nobody in that chain checked it against IGL's customers. IGL's customers have already voted, in writing, in the opposite direction. The Loox export (551 reviews, 537 active, 88.2% five-star, covering 2024-08-31 to 2025-12-09) contains four independent praises of thinness and breathability in heat, and four fit complaints that all point the same way: too long, too oversized, armholes too big. The file names it in its own pain-point list: "Shirt length (multiple 'too long' mentions)." Comfort Colors 1717 is the heavyweight, boxy, garment-dyed option — 6.1 oz, per IGL's own Sloth Hiking Club teardown. Jo Conlon classifies it that way herself and notes her warm-climate customers always need the lightweight option. IGL's buyers are 68–70% female with a core age of 55–75+, gardening outdoors through a Missouri summer. No single shelf produces this. The founders shelf supplies the permission to refuse; the mastermind shelf supplies both the prescription and the garden-niche counter-example; the brand shelf supplies the customer sentences that decide it. ## Worked examples — the eight review sentences that decide it Pushing shorter and less boxy (4): - Janet F. (5★, ~10 shirts): "I am short, though and most of the shirts are about 5 inches too long. I try to be creative when I wear them. Perhaps you could consider making them less long? I'm not sure, am I the only one with this problem?" - Wendy L. (4★, tank): "The only thing I would change about this shirt is the size of the arm holes. They are too big and really one needs to wear another tank too or sports bra underneath." - Unnamed (colorwashed comfort tank): "The tank is a bit oversized and I found that the super soft texture of the fabric tends to stretch out a bit as you wear it. People may want to consider getting one size smaller if they don't like their tank tops baggie." - Jennifer Fox (5★, long sleeve): "the fit is comfortable and generous — I probably could have even sized down one size." Pushing lighter (4): - Wendy L.: "I liked this one the best as it was thin and that is appreciated when it gets very hot out." - Lydia N. (5★, racerback): "The fabric is soft, comfortable and breathable which is a must for these humid summers." - becca c. (5★, Zone 9a): "I already put it to the test while shoveling and digging new garden beds… The fabric breathes really well, which is a huge plus here in Zone 9a." - Joseph T. (5★): "made out of excellent material in order to keep you super comfortable in warm weather." ## Evidence trail - founders-kb/episodes/419-kelly-johnson-skunk-works — primary, seed. Founder episode, no status flag. Rule 9 (flight-test your own product or lose competency), the returned government money, the killed hydrogen plane. - igl-brand-kb/community/customer-reviews-q4-2025 — primary. Loox review export, export and processing date 2025-12-11, no status header. 8½ months old at the time of writing — old enough that you should confirm nothing changed, but this is review text, not sales data. Source of all eight quotes above, the summary statistics, and the file's own "Shirt length (multiple 'too long' mentions)" pain point. - wescale-kb/calls/call-32 — primary, 2026-06-04. The prescription itself at line 103. Also carries its own limits: line 12 states no transcript or chat log was published for Call #32, so it is reconstructed from the debrief, and Pati's $13–14K → $46K is a self-reported 7-day figure with a hypothesised mockup-brightness mechanism. - wescale-kb/members/jo-conlon — primary. The garden-niche peer. Slack, 2026-02-12: "I differentiate Bella Canvas and Comfort Colors in terms of Lightweight and Heavyweight garments, both premium products… my core customers are in Florida, California & Texas and lightweight tees and tanks are always going to be good to have as an option." - igl-brand-kb/competitive/happy-plant-people-email-analysis — primary, 2026-03-17. Added by the operator; the hunt never opened it. Its comparison table lists IGL's blanks as "Comfort Colors (via Printify)" — contradicting the Gildan review quote and making Step 0 necessary. - igl-brand-kb/core/buyer-profile-deep-dive — theme-aggregate. Status: CURRENT SoR, last updated February 2026, data dated 2026-01-01. ~1,000-profile Klaviyo sample (~22% of the base): 68–70% female, core age 55–75+ inferred from name-cohort and email-domain patterns, not verified. ## Assay verdicts - Novelty — PASS, 4. Matches the advice layer head-on (tim-dobyns.md:80 carries the assignment verbatim) and that is exactly the case the priority-check nuance protects: not a restatement, a documented refusal of the prescription with evidence the prescriber never had. No prior nugget argues against a cohort prescription on customer-fit grounds. - Grounding — PASS, 5. "The most carefully sourced candidate in the batch." Verified call-32.md:103, the self-reported limits at line 12, Pati's figures at :90 and :122, jo-conlon.md:143-146, and the review quotes. - Consequence — PASS, 5. Kills a queued action item and substitutes a named alternative. Every noun defined: the two candidate blanks, the person who wears and washes them, the measurement to take, the number to publish. Cost is two sample tees on an account he already has. Score: 14 of 15 — the highest-scoring survivor of five. ## Operator corrections applied before shipping 1. The rebate claim was reversed into a certainty it does not have. The candidate said flatly that Comfort Colors is excluded from the Printify 9% rebate. The record is contested: Call #33 (2026-06-11) says the 9% rebate covers "Gildan, Comfort Colors, Bella Canvas," confirmed in writing Feb 16 to renew in perpetuity — with the formalised proposal still not arrived. Call #38 (2026-07-16) says Comfort Colors "and other products outside the current rebate won't be added this round." The later record wins on recency, but the conflict is real and unresolved. Ask Chris before relying on it. The direction of the risk is unchanged either way: Gildan and Bella Canvas are named as covered in both readings, so the lightweight route keeps the rebate under either. 2. Pati's mechanism was stated as a finding; the source states it as a maybe. Corrected in the text above, and the omitted counter-signal ("better than the Gildan versions") restored. 3. Two IGL documents disagree about which blank IGL prints on today. The candidate assumed Gildan and built a migration argument on it. Step 0 added. 4. "The Springfield garden" is an undefined noun by the consequence judge's own test. Replaced with "her garden." 5. A strengthener the hunt missed: IGL already sells Bella Canvas — the 3500 V-neck, per igl-brand-kb/voice/email-templates.md:926. The lightweight test is not a new supplier relationship. Thread: ep-419 Kelly Johnson rule 9 (flight-test your own product) + killing his own hydrogen plane → call-32 action item "move best sellers to Comfort Colors" → jo-conlon profile: Comfort Colors = Heavyweight, warm-climate garden customers need Lightweight → igl customer-reviews-q4-2025: four heat praises, four fit complaints, all pointing lighter and shorter → operator: happy-plant-people-email-analysis contradicts the assumed starting blankOpen igl-brand-kb/methodology/design-flywheel-framework.md and write a new thread into the section headed "The Five Identity Threads" (line 69). That section already names five: two proven in prose — Bird Lover and Tomato Grower — plus three written out as ### Thread 1 (Exhausted but Alive Gardener), ### Thread 2 (Second Half of Life Gardener) and ### Thread 3 (Creature Welcomer). Yours becomes ### Thread 4 in the numbering and the sixth thread overall, so change the heading on line 69 to "The Six Identity Threads" in the same edit. Line 73 already authorises it: IGL needs "5–6 distinct identity threads." Five exist. This is the sixth.
Name it for the constraint, not the sorrow: the gardener whose reach got smaller. Not grief, not retirement — physical reach. Every existing thread assumes a body that still works; Thread 2 is about finding the garden later, not losing access to it.
Source the phrases from three places you already own, in this order:
1. igl-brand-kb/community/tlc-thematic-stories.md, lines 81–95 — the population, already written down: "Raised beds built at wheelchair height. Pulley systems for lowering hanging baskets. Greenhouse spaces created from recycled windows"; micro-dwarf tomatoes grown "in small pots, on windowsills, in office spaces"; members who arrived after "medical diagnoses that shattered future plans" or carrying "the invisible wounds of military service." 2. Two named interviews inside your own community, not one. Lauren Blanchard Zalewski (community/lauren-and-rowena-stories.md:26) founded the Tomato Lovers Collective, has lived with chronic illness 20+ years, and runs "Attitude of Gratitude with Chronic Pain" — 11,000 members across 82 countries. Amanda (community/tlc-featured-gardener-stories.md:357) breeds dwarf and micro-dwarf varieties herself and builds communities specifically designed to "help disabled people and urban dwellers garden." Amanda is the closer match to the thread and has never been asked about it. 3. The quarterly survey you already run. igl-brand-kb/methodology/observation-mining-methodology.md line 182 lists "Quarterly TLC surveys for specific collection feedback" as a standing input, and lauren-and-rowena-stories.md:184 shows one actually ran (TLC Tank Tester Program, 2025). Add one open question: "What did you change so you could keep growing?"
Write the thread at Tier 2 — the framework's own definition (line 113) is "Speaking to the gardener — more contemplative, theology lives in the image," no scripture cited. Hold the no-dark-humor line absolutely: this thread is tender, never a punchline.
IGL's best-selling shirt is documented being bought for a man in a nursing home who can no longer garden, and the brand has no thread, collection, or persona for the gardener whose reach got smaller.
The most moving passage in the Goodyear episode is also the most specific. At 52 — joints wrecked by gout and lead poisoning, face yellow, needing a cane — he filled a journal with sketches of "rubber invalid cushions, baby jumpers and swings, ear trumpets for the hearing-impaired, abdominal supporters for hernia sufferers, life preservers." His stated fear was that if he died first, "hospital patients and invalids might never lie down to rest in a rubber air or water bed." He was designing for bodies that had stopped cooperating, at exactly the moment his own did.
Now read IGL's own best-selling product through that lens. Kimberly J., five stars, 2025-07-16, on the bird-feeder shirt: "My father-in-law is in a nursing home and we have put a bird feeder outside his room. Of course the squirrels enjoy the bird seed more than the birds... he absolutely loved it when he saw it. He laughed so hard." That is the megalodon working on a man who cannot garden. The garden came to the window. It is the only such review in the whole brand knowledge base.
The population is not hypothetical and IGL has already written it down twice — as community stories, never as a market. Wheelchair-height beds. Pulley systems. Windowsill micro-dwarfs. The founder of IGL's own community is a chronic-pain advocate with 11,000 members. Another member breeds micro-dwarfs specifically to help disabled people garden. And the cohort supplies the confirming case from the faith side: Kelli Sanders' customer wrote from hospice with ALS, "There are things from my life and this ALS journey that are reflected in your shirts," and placed two large orders.
Why no single knowledge base could produce this: founders-kb has the practice but no gardeners. igl-brand-kb has the customer, the community stories and the framework with an empty slot, but nothing tells it the empty slot is the same shape as the customer. wescale-kb has the cohort proof that buyers in physical decline attach real meaning to apparel, about a different brand. Goodyear, who could barely walk, wrote this thread first and called it a journal.
founders-kb/episodes/069-charles-goodyear-inventors-obsession — the practice: at 52, crippled, sketching invalid cushions, ear trumpets and abdominal supporters so that "hospital patients and invalids" would not go without (lines 120–124, 238).igl-brand-kb/community/customer-reviews-q4-2025 — Kimberly J., 5★, 2025-07-16: the nursing-home bird feeder outside the father-in-law's window (Story #8, line 98; the same review is quoted again in the summary list at line 202).igl-brand-kb/community/tlc-thematic-stories — the adapted-gardener population in IGL's own words: wheelchair-height beds, pulley systems, recycled-window greenhouses, windowsill micro-dwarfs, "medical diagnoses that shattered future plans" (lines 9, 81–95).igl-brand-kb/community/tlc-featured-gardener-stories — Amanda, who breeds dwarf and micro-dwarf varieties and builds communities to "help disabled people and urban dwellers garden" (line 357). Added by the operator; the hunt missed her.igl-brand-kb/community/lauren-and-rowena-stories — Lauren Blanchard Zalewski: TLC founder, 20+ years chronic illness, 11,000-member chronic-pain community across 82 countries, and gardening as the flow state where she stops "obsessing about your pain and your troubles" (lines 26, 28).wescale-kb/members/kelli-sanders — FaithWise + Co: the hospice/ALS customer, two large orders, quoted verbatim (line 79); replacement fee waived (line 35).igl-brand-kb/methodology/design-flywheel-framework — theme-aggregate. Filed 2026-03-03. Names five threads and states IGL needs "5–6 distinct identity threads" (line 73). All five assume a body that still works. This is the file the move edits.igl-brand-kb/methodology/design-flywheel-framework.md and write a new thread into the section headed "The Five Identity Threads" (line 69). That section already names five: two proven in prose — Bird Lover and Tomato Grower — plus three written out as ### Thread 1 (Exhausted but Alive Gardener), ### Thread 2 (Second Half of Life Gardener) and ### Thread 3 (Creature Welcomer). Yours becomes ### Thread 4 in the numbering and the sixth thread overall, so change the heading on line 69 to "The Six Identity Threads" in the same edit. Line 73 already authorises it: IGL needs "5–6 distinct identity threads." Five exist. This is the sixth. Name it for the constraint, not the sorrow: the gardener whose reach got smaller. Not grief, not retirement — physical reach. Every existing thread assumes a body that still works; Thread 2 is about finding the garden later, not losing access to it. Source the phrases from three places you already own, in this order: 1. igl-brand-kb/community/tlc-thematic-stories.md, lines 81–95 — the population, already written down: "Raised beds built at wheelchair height. Pulley systems for lowering hanging baskets. Greenhouse spaces created from recycled windows"; micro-dwarf tomatoes grown "in small pots, on windowsills, in office spaces"; members who arrived after "medical diagnoses that shattered future plans" or carrying "the invisible wounds of military service." 2. Two named interviews inside your own community, not one. Lauren Blanchard Zalewski (community/lauren-and-rowena-stories.md:26) founded the Tomato Lovers Collective, has lived with chronic illness 20+ years, and runs "Attitude of Gratitude with Chronic Pain" — 11,000 members across 82 countries. Amanda (community/tlc-featured-gardener-stories.md:357) breeds dwarf and micro-dwarf varieties herself and builds communities specifically designed to "help disabled people and urban dwellers garden." Amanda is the closer match to the thread and has never been asked about it. 3. The quarterly survey you already run. igl-brand-kb/methodology/observation-mining-methodology.md line 182 lists "Quarterly TLC surveys for specific collection feedback" as a standing input, and lauren-and-rowena-stories.md:184 shows one actually ran (TLC Tank Tester Program, 2025). Add one open question: "What did you change so you could keep growing?" Write the thread at Tier 2 — the framework's own definition (line 113) is "Speaking to the gardener — more contemplative, theology lives in the image," no scripture cited. Hold the no-dark-humor line absolutely: this thread is tender, never a punchline. ## Why (the one-liner) IGL's best-selling shirt is documented being bought for a man in a nursing home who can no longer garden, and the brand has no thread, collection, or persona for the gardener whose reach got smaller. ## The insight The most moving passage in the Goodyear episode is also the most specific. At 52 — joints wrecked by gout and lead poisoning, face yellow, needing a cane — he filled a journal with sketches of "rubber invalid cushions, baby jumpers and swings, ear trumpets for the hearing-impaired, abdominal supporters for hernia sufferers, life preservers." His stated fear was that if he died first, "hospital patients and invalids might never lie down to rest in a rubber air or water bed." He was designing for bodies that had stopped cooperating, at exactly the moment his own did. Now read IGL's own best-selling product through that lens. Kimberly J., five stars, 2025-07-16, on the bird-feeder shirt: "My father-in-law is in a nursing home and we have put a bird feeder outside his room. Of course the squirrels enjoy the bird seed more than the birds... he absolutely loved it when he saw it. He laughed so hard." That is the megalodon working on a man who cannot garden. The garden came to the window. It is the only such review in the whole brand knowledge base. The population is not hypothetical and IGL has already written it down twice — as community stories, never as a market. Wheelchair-height beds. Pulley systems. Windowsill micro-dwarfs. The founder of IGL's own community is a chronic-pain advocate with 11,000 members. Another member breeds micro-dwarfs specifically to help disabled people garden. And the cohort supplies the confirming case from the faith side: Kelli Sanders' customer wrote from hospice with ALS, "There are things from my life and this ALS journey that are reflected in your shirts," and placed two large orders. Why no single knowledge base could produce this: founders-kb has the practice but no gardeners. igl-brand-kb has the customer, the community stories and the framework with an empty slot, but nothing tells it the empty slot is the same shape as the customer. wescale-kb has the cohort proof that buyers in physical decline attach real meaning to apparel, about a different brand. Goodyear, who could barely walk, wrote this thread first and called it a journal. ## Evidence trail - founders-kb/episodes/069-charles-goodyear-inventors-obsession — the practice: at 52, crippled, sketching invalid cushions, ear trumpets and abdominal supporters so that "hospital patients and invalids" would not go without (lines 120–124, 238). - igl-brand-kb/community/customer-reviews-q4-2025 — Kimberly J., 5★, 2025-07-16: the nursing-home bird feeder outside the father-in-law's window (Story #8, line 98; the same review is quoted again in the summary list at line 202). - igl-brand-kb/community/tlc-thematic-stories — the adapted-gardener population in IGL's own words: wheelchair-height beds, pulley systems, recycled-window greenhouses, windowsill micro-dwarfs, "medical diagnoses that shattered future plans" (lines 9, 81–95). - igl-brand-kb/community/tlc-featured-gardener-stories — Amanda, who breeds dwarf and micro-dwarf varieties and builds communities to "help disabled people and urban dwellers garden" (line 357). Added by the operator; the hunt missed her. - igl-brand-kb/community/lauren-and-rowena-stories — Lauren Blanchard Zalewski: TLC founder, 20+ years chronic illness, 11,000-member chronic-pain community across 82 countries, and gardening as the flow state where she stops "obsessing about your pain and your troubles" (lines 26, 28). - wescale-kb/members/kelli-sanders — FaithWise + Co: the hospice/ALS customer, two large orders, quoted verbatim (line 79); replacement fee waived (line 35). - igl-brand-kb/methodology/design-flywheel-framework — theme-aggregate. Filed 2026-03-03. Names five threads and states IGL needs "5–6 distinct identity threads" (line 73). All five assume a body that still works. This is the file the move edits. ## Assay verdicts - novelty — PASS (4). No prior art in any base or in the nugget catalog. Grepped wheelchair, chronic pain, chronic illness, disability, adaptive gardening, limited mobility, nursing home, hospice. Independently re-checked by the operator: every "accessibility" hit in igl-brand-kb is either a community story describing the population or the voice library using the word to mean theological approachability. Nearest existing constructs are genuinely different — Thread 2 is finding the garden later, and observation-mining-methodology.md:449 scopes Healing to depression/anxiety/grief/addiction, i.e. mental rather than physical constraint. - grounding — PASS (5). Every citation verified verbatim by the judge and re-verified by the operator against the files. - actionability — PASS (3). Docked for depending on another person (Lauren) and for an unverified survey. The operator overturned half of that deduction: the quarterly Tomato Lovers Collective survey is documented (observation-mining-methodology.md:182) and one demonstrably ran in 2025. The person-dependency is real and is why a second named interviewee (Amanda) was added. Thread: ep-069 Goodyear at 52 sketching ear trumpets → Kimberly J.'s nursing-home bird feeder, the only one in the brand knowledge base → tlc-thematic-stories' wheelchair-height beds and windowsill micro-dwarfs → Amanda breeding micro-dwarfs for disabled gardeners → Lauren's 11,000-member chronic-pain community → kelli-sanders' hospice/ALS customer → the Design Flywheel's five threads, every one written for a body that still worksMessage Lauren Blanchard Zalewski — founder of the Tomato Lovers Collective, the person IGL already works with, profiled in igl-brand-kb/community/lauren-and-rowena-stories.md — and offer Inspired Garden Life as the official shirt supplier for the 2027 growout teams. The product is one template reskinned per team: the team's category name on the front, the grower's last name on the back, jersey style.
Ask before the end of September 2026, not in December. Signups open the first week of January and close about a week later (Wednesday January 14 in 2026), teams are picked immediately after, and the shirt has to exist while people are being placed on teams rather than after. That is a four-month lead time on a decision Lauren makes once a year.
Get written permission before any surname is printed. igl-brand-kb/core/strategic-synthesis-v4.md carries a standing rule that the Collective is a partner community and a member's name may not be published without permission. This move puts individual growers' surnames on merchandise, so it needs Lauren's agreement plus each grower's. The precedent is in IGL's own files: the community-stories document records "Story permissions granted by both Lauren Blanchard Zalewski and Rowena Kotelniski" (line 187), so a permission step already exists and just has to be reused.
Size it honestly. The January 8, 2026 show states 14 tomato growout categories, 15 people per team plus 3 backups, and five parallel pepper categories on the same signup cycle. That is 342 team slots at the ceiling (19 categories × 18). Two cautions on that number, both from the source: participants may sign up for up to two teams, so the count of distinct people is lower than 342 and the knowledge bases do not say by how much; and the 15-plus-3 rule is stated for the growouts generally, so the pepper team sizes should be confirmed with Lauren rather than assumed. Whatever the true figure, every one of these people already holds the TLC20 code and already knows the product.
The one product IGL's weekly ad-testing system can never discover was pitched on IGL's own sponsored show fourteen months ago by somebody else, and nothing in the knowledge bases shows IGL ever answered.
This is not a design idea, it is a sizing problem. A fifteen-person team shirt cannot survive the cohort's testing machine. Call #32's Bump and Cut Decision Tree says let a new ad set spend to roughly 2× cost per acquisition — "~48h at $25/day" — then cut anything under 2.0 ROAS. A product whose entire addressable audience is fifteen people will never clear that bar, so a system built to find designs that win a 48-hour auction is structurally incapable of proposing one. It is below the threshold, permanently, and no amount of creative volume changes that.
Walton's whole company came out of exactly that gap. Helen's rule — "Ten thousand people is enough for me" — put every town worth fighting for off limits and forced him into places the major chains had already priced as too small to bother with. His conclusion: "there was much, much more business out there in small town America than anybody, including me, had ever dreamed of." He also named the failure mode of the competitors who half-committed: "They held on to their old variety store concepts way too long." Fifteen people is IGL's small town.
The time horizon is the other half. A growout is a ten-month commitment with a hard November 1 deadline, taken on in January. The buying decision, the identity, and the deadline are all fixed months before the season ends — the opposite of every input IGL's design pipeline currently consumes. There is no ad test, no ROAS threshold and no creative-fatigue curve involved. There is a named list and a date.
Why no single knowledge base gets here: founders-kb has the constraint-as-advantage argument, wescale-kb has the 48-hour rule that explains why the idea is invisible to IGL's system rather than merely unattempted, and the actual pitch is sitting in an AI summary of a gardening livestream in recall-kb that nobody has read as a business document.
recall-kb/Business/Agriculture/Lauren's TOMATO Awesomesauce!!-June 5, 2025 (Prizes from Inspired Garden Life!) — the pitch, verbatim (line 37): "Sherman has designed special shirts with the last name on the back, similar to a jersey style, for members of the growout teams, and the host is conducting an informal poll to see if this is something that would interest the team members." Same capture establishes IGL as that episode's sponsor (line 33, "Tim and Daryl from Inspired Garden Life"), the TLC20 forever code (line 22), and the group at "almost 7,800 members" (line 256). Record of a June 2025 broadcast, captured Dec 2025; used as a historical record of a pitch, not as current state. No evidence anywhere in the knowledge bases that IGL responded.recall-kb/Business/Agriculture/Lauren's TOMATO Awesomesauce!!-January 8, 2026 (Prizes from Inspired Garden Life!) — sizes the opportunity and sets the calendar: "a total of 14 categories" (line 43), "15 people per team, with three backups" (line 40), signups "open until Wednesday, January 14th" (line 73), pepper categories "sweet new Mexican varieties, jalapeno and jalapeno crosses, hots, super hots, and variegated peppers" (line 36), "send their plants to the host by November 1st" (line 48), and the two-team cap (line 50). Record of a January 2026 broadcast — seven months old, describing an annual recurring cycle, so used for structure and not as a live figure.wescale-kb/calls/call-32 — dated 2026-06-04. Supplies the rule that makes a 15-person product invisible: "New ad set → spend to 2× CPA (~48h at $25/day) before deciding… ROAS: <2.0 cut, 2.0–2.2 leave, 2.2–2.5 your call, 2.5+ bump." The file states its own limitation in its frontmatter — "Built from the Weekly Debrief alone… no transcript (.vtt) or chat log was made available for Call #32" — so its language is the debrief's, not verbatim speech.founders-kb/episodes/234-sam-walton-made-in-america — the seed. Helen Walton's constraint, verbatim: "I'll go with you any place you want to go, as long as you don't ask me to live in a big city. Ten thousand people is enough for me." Plus the small-town conclusion and "They held on to their old variety store concepts way too long." Historical narrative; no business figures borrowed from it.igl-brand-kb/community/lauren-and-rowena-stories — who to send the message to, and the permission precedent. IGL's own words: Lauren is "the visionary founder of the Tomato Lovers Collective itself — the community that would become the beating heart of Inspired Garden Life," "an obsessive planner." Line 187 records that story permissions were granted by both named members. No status header; brand-narrative material, so used only for identity and the permission precedent, not for any number.igl-brand-kb/community/lauren-and-rowena-stories.md — and offer Inspired Garden Life as the official shirt supplier for the 2027 growout teams. The product is one template reskinned per team: the team's category name on the front, the grower's last name on the back, jersey style. Ask before the end of September 2026, not in December. Signups open the first week of January and close about a week later (Wednesday January 14 in 2026), teams are picked immediately after, and the shirt has to exist while people are being placed on teams rather than after. That is a four-month lead time on a decision Lauren makes once a year. Get written permission before any surname is printed. igl-brand-kb/core/strategic-synthesis-v4.md carries a standing rule that the Collective is a partner community and a member's name may not be published without permission. This move puts individual growers' surnames on merchandise, so it needs Lauren's agreement plus each grower's. The precedent is in IGL's own files: the community-stories document records "Story permissions granted by both Lauren Blanchard Zalewski and Rowena Kotelniski" (line 187), so a permission step already exists and just has to be reused. Size it honestly. The January 8, 2026 show states 14 tomato growout categories, 15 people per team plus 3 backups, and five parallel pepper categories on the same signup cycle. That is 342 team slots at the ceiling (19 categories × 18). Two cautions on that number, both from the source: participants may sign up for up to two teams, so the count of distinct people is lower than 342 and the knowledge bases do not say by how much; and the 15-plus-3 rule is stated for the growouts generally, so the pepper team sizes should be confirmed with Lauren rather than assumed. Whatever the true figure, every one of these people already holds the TLC20 code and already knows the product. ## Why (the one-liner) The one product IGL's weekly ad-testing system can never discover was pitched on IGL's own sponsored show fourteen months ago by somebody else, and nothing in the knowledge bases shows IGL ever answered. ## The insight This is not a design idea, it is a sizing problem. A fifteen-person team shirt cannot survive the cohort's testing machine. Call #32's Bump and Cut Decision Tree says let a new ad set spend to roughly 2× cost per acquisition — "~48h at $25/day" — then cut anything under 2.0 ROAS. A product whose entire addressable audience is fifteen people will never clear that bar, so a system built to find designs that win a 48-hour auction is structurally incapable of proposing one. It is below the threshold, permanently, and no amount of creative volume changes that. Walton's whole company came out of exactly that gap. Helen's rule — "Ten thousand people is enough for me" — put every town worth fighting for off limits and forced him into places the major chains had already priced as too small to bother with. His conclusion: "there was much, much more business out there in small town America than anybody, including me, had ever dreamed of." He also named the failure mode of the competitors who half-committed: "They held on to their old variety store concepts way too long." Fifteen people is IGL's small town. The time horizon is the other half. A growout is a ten-month commitment with a hard November 1 deadline, taken on in January. The buying decision, the identity, and the deadline are all fixed months before the season ends — the opposite of every input IGL's design pipeline currently consumes. There is no ad test, no ROAS threshold and no creative-fatigue curve involved. There is a named list and a date. Why no single knowledge base gets here: founders-kb has the constraint-as-advantage argument, wescale-kb has the 48-hour rule that explains why the idea is invisible to IGL's system rather than merely unattempted, and the actual pitch is sitting in an AI summary of a gardening livestream in recall-kb that nobody has read as a business document. ## Evidence trail - recall-kb/Business/Agriculture/Lauren's TOMATO Awesomesauce!!-June 5, 2025 (Prizes from Inspired Garden Life!) — the pitch, verbatim (line 37): "Sherman has designed special shirts with the last name on the back, similar to a jersey style, for members of the growout teams, and the host is conducting an informal poll to see if this is something that would interest the team members." Same capture establishes IGL as that episode's sponsor (line 33, "Tim and Daryl from Inspired Garden Life"), the TLC20 forever code (line 22), and the group at "almost 7,800 members" (line 256). Record of a June 2025 broadcast, captured Dec 2025; used as a historical record of a pitch, not as current state. No evidence anywhere in the knowledge bases that IGL responded. - recall-kb/Business/Agriculture/Lauren's TOMATO Awesomesauce!!-January 8, 2026 (Prizes from Inspired Garden Life!) — sizes the opportunity and sets the calendar: "a total of 14 categories" (line 43), "15 people per team, with three backups" (line 40), signups "open until Wednesday, January 14th" (line 73), pepper categories "sweet new Mexican varieties, jalapeno and jalapeno crosses, hots, super hots, and variegated peppers" (line 36), "send their plants to the host by November 1st" (line 48), and the two-team cap (line 50). Record of a January 2026 broadcast — seven months old, describing an annual recurring cycle, so used for structure and not as a live figure. - wescale-kb/calls/call-32 — dated 2026-06-04. Supplies the rule that makes a 15-person product invisible: "New ad set → spend to 2× CPA (~48h at $25/day) before deciding… ROAS: <2.0 cut, 2.0–2.2 leave, 2.2–2.5 your call, 2.5+ bump." The file states its own limitation in its frontmatter — "Built from the Weekly Debrief alone… no transcript (.vtt) or chat log was made available for Call #32" — so its language is the debrief's, not verbatim speech. - founders-kb/episodes/234-sam-walton-made-in-america — the seed. Helen Walton's constraint, verbatim: "I'll go with you any place you want to go, as long as you don't ask me to live in a big city. Ten thousand people is enough for me." Plus the small-town conclusion and "They held on to their old variety store concepts way too long." Historical narrative; no business figures borrowed from it. - igl-brand-kb/community/lauren-and-rowena-stories — who to send the message to, and the permission precedent. IGL's own words: Lauren is "the visionary founder of the Tomato Lovers Collective itself — the community that would become the beating heart of Inspired Garden Life," "an obsessive planner." Line 187 records that story permissions were granted by both named members. No status header; brand-narrative material, so used only for identity and the permission precedent, not for any number. ## Assay verdicts - Novelty — PASS (4). Grep for "sherman", "jersey", "team shirt", "growout" across all knowledge bases plus CATALOG.md and catalog/*.md returns only nugget #019, which mines the same show for seed want lists (design subjects), not for a team-shirt product. A genuine narrowing, not a restatement. - Grounding — PASS (5). Every quote verified verbatim by the judge and re-verified by the operator. Operator caveat: the judge passed a citation whose file path did not exist — see the corrections below. - Consequence — PASS (5). Names a specific decision, a specific person to contact, and a deadline that is four months out. ## Operator corrections applied Four, made after the 3-of-3 verdict and before this card shipped. 1. Broken evidence path corrected. The winner cited recall-kb/Community/Lauren's TOMATO Awesomesauce!!-June 5, 2025…. That path does not exist. The file is real but lives in recall-kb/Business/Agriculture/; only the July 10, 2025 capture sits under Community/, which is the likely source of the confusion. The content is genuine and verified verbatim — this was a wrong folder, not a fabrication. But the grounding judge's own instruction says "Missing file = FAIL", and it passed this citation with a score of 5. Logged to the ledger as a judge defect: the judge verified the quote by finding the file some other way and never checked that the path it would print for Tim resolved. 2. Arithmetic corrected — same defect class as night #19. The winner said "14 tomato categories at 15 growers plus 3 backups each, plus the five pepper categories, is roughly 250 pre-committed buyers." 14 × 18 = 252 before peppers are counted at all; including the five pepper categories the ceiling is 342 slots. The winner also called them "buyers" when they are growers who would still have to choose to buy, and it ignored the source's own two-team cap (line 50), which pushes the count of distinct people below 342. Rewritten as a ceiling with both cautions stated. Night #19 recorded that the grounding judge verifies quotes without checking the arithmetic built on top of them; that is now four nights running. 3. False date precision removed. The winner asserted "IGL's next sponsored first-Thursday episode (Thursday, September 3, 2026)." The show is monthly on a Thursday, but the record shows it is not reliably the first: of five captures on disk, July 10 2025 and January 8 2026 were both second Thursdays (January 1, 2026 was itself a Thursday). The newest capture of the show in any knowledge base is January 8, 2026 — over seven months old — so nothing on disk confirms the September 2026 schedule or that the monthly sponsorship still runs. Replaced with "before the end of September 2026" and an instruction to confirm the date with Lauren. The partnership itself is current: wescale-kb/reports/igl-cro-plan-2026-08-06.md (Status: CURRENT SoR, 19 days old) treats the Collective as a live partner community. 4. Permission constraint added. The move prints individual members' surnames on merchandise and said nothing about consent. strategic-synthesis-v4.md carries a standing rule that the Collective is a partner community and names may not be published without permission. Added as a required step, with the existing precedent named so it reads as reusing a process rather than inventing one. Thread: ep-234 Sam Walton (Helen's 10,000-person rule) → recall-kb June 5 2025 (Sherman's jersey-style growout shirts, polled on IGL's sponsored episode) → recall-kb Jan 8 2026 (14 categories × 15+3, Jan 14 close, Nov 1 deadline) → wescale call-32 (the 48-hour bump-and-cut rule that makes a 15-person product invisible) → igl-brand-kb lauren-and-rowena-stories (who to ask, and the permission precedent)