1. Add one line to the Monday scorecard. The mastermind's weekly scorecard (made "weekly, permanently" on Call #41) has three input rows: designs, ads, emails. Add a fourth, derived, above them: "selling weeks in the trailing 8 with zero new designs published." Target 0–1. You are currently at 4–5.
2. Turn Seedwork from a publisher into a release queue. Today a batch finishes and the whole batch goes live, then nothing goes live for a month. Instead: a finished batch lands in a holding list, and one fixed Monday slot releases 8–12 designs into Shopify — never the whole batch, never zero. Worked example from your own history: site week 41 published 48 designs and week 42 published 72. Released at 12 a week, those two batches alone cover ten consecutive weeks instead of two.
3. Stop repeating the "24 designs/week" number, and fix the file it came from. Your member profile says "Target: 24 designs/week (above WeScale median of ~12–15/week)". Measured, you are at 12.9 against a cohort median of 12.4 — rank 31 of 65, the exact middle. The number to correct first is in wescale-kb/igl-action-plan.md (see "Where the belief came from" below); the same belief is repeated in six other places, listed at the bottom of this card.
IGL is not a high-volume design store with a conversion problem; it is a dead-median-volume store with the 5th-most intermittent publishing rhythm in the cohort — and the famous "+160% creative volume" is ads, not designs.
Franklin's trade judged a printer by the job: how big a run did you land, how many sheets came off the press. Franklin kept a different scorecard. His thirteen-virtue chart did not count virtuous acts and take an average — it had a column per day, and he put a black mark on the days he failed, weeding one bed of the garden at a time. The line that won him the Philadelphia stationers' credit was Dr. Baird's: "I see him still at work when I go home from club, and he is at work again before his neighbors are out of bed." That is a cadence claim, not a volume claim. He also named the failure mode: the speckled ax — the man who wanted a bright blade, found the grinding tedious, and decided he liked it speckled.
The mastermind asks the count question. But Chris Heckman's Black Friday Eve cohort analysis (Call #08, 2025-11-26) already published the Franklin answer right next to it: top-20% stores averaged 18 designs/week and missed only 8% of weeks; the middle 60% missed 24%; the bottom 20% averaged 6 and missed 39%. In ten months nobody has scored IGL against that table.
Scored: over the clean import window of the cohort scorecard database (site weeks 1–42), across all 65 members with 12 or more selling weeks — IGL averages 12.9 designs/week against a 12.4 median, rank 31 of 65, and published zero designs in 15 of its 27 design-reported selling weeks — 55.6%, rank 5 of 65. That is worse than Chris's bottom-tier benchmark of 39%. Meanwhile IGL averages 29.0 new ad creatives/week against a 17.0 median, rank 16 of 65. The "IGL runs huge creative volume" belief is true — of ads. It is false of designs, and it has been masking a rhythm problem for a year.
No single shelf produces this. founders-kb supplies the reframe (mark the days you failed, not the output). wescale-kb's call archive supplies the benchmark nobody applied. The scorecard database supplies the number, which appears in no prose document in any of the nine knowledge bases.
founders-kb/episodes/115-benjamin-franklin-an-american-life — seed, primary. Key Lesson 4: daily tracking of adherence to the thirteen virtues, "deliberate, measurable self-improvement applied consistently." Establishes that Franklin's reputation rested on observed daily rhythm.mylibrary-kb/books/autobiography-of-benjamin-franklin — primary; same man, second document, counted as ONE strand, not independent confirmation. Supplies the mechanics: black marks for faults rather than credits for successes, one virtue per week in four 13-week courses, the weeding-one-bed analogy, the speckled-ax parable, and Dr. Baird's testimony.wescale-kb/calls/call-08 — primary. Dated 2025-11-26 — a quotation from a 10-month-old document, not a live measurement. Carries the design-volume tiers (18 / 11.6 / 6 per week) and the consistency slice (8% / 24% / 39% of weeks with zero designs). Verified verbatim, including in wescale-kb/transcripts/call-08-transcript.md:170.wescale-kb/scorecard/wescale-scorecard.db — primary, structured artifact. 51,411 observations, 92 members, 49 weeks. All cohort figures in this card were computed over the full 65-member population, not a sample.wescale-kb/members/tim-dobyns — primary. The belief being corrected: line 36 "Target: 24 designs/week", line 99 "Target: 24 designs/week (above WeScale median of ~12-15/week)", line 222 "15/week → 24/week sustained".wescale-kb/igl-action-plan.md — primary, operator-added. Created 2026-02-16, window "Weeks 13–19". The origin of the "+160% / input is fine" doctrine. Note it already carries a caveat line added by an earlier night saying "'Input is fine' is true against the median" — which this card shows is itself false for designs.wescale-kb/themes/consistency-over-bursts — the single theme-aggregate. Cited only to show the prior art exists and that this move is its un-prescribed instance: it prescribes consistency in general and quotes Chris's "I'd rather see 10 ads every week for 8 weeks than 40 ads one week and 0 the next three", but names no IGL number and proposes no zero-week scorecard line.wescale-kb/igl-action-plan.md (see "Where the belief came from" below); the same belief is repeated in six other places, listed at the bottom of this card. ## Why (the one-liner) IGL is not a high-volume design store with a conversion problem; it is a dead-median-volume store with the 5th-most intermittent publishing rhythm in the cohort — and the famous "+160% creative volume" is ads, not designs. ## The insight Franklin's trade judged a printer by the job: how big a run did you land, how many sheets came off the press. Franklin kept a different scorecard. His thirteen-virtue chart did not count virtuous acts and take an average — it had a column per day, and he put a black mark on the days he failed, weeding one bed of the garden at a time. The line that won him the Philadelphia stationers' credit was Dr. Baird's: "I see him still at work when I go home from club, and he is at work again before his neighbors are out of bed." That is a cadence claim, not a volume claim. He also named the failure mode: the speckled ax — the man who wanted a bright blade, found the grinding tedious, and decided he liked it speckled. The mastermind asks the count question. But Chris Heckman's Black Friday Eve cohort analysis (Call #08, 2025-11-26) already published the Franklin answer right next to it: top-20% stores averaged 18 designs/week and missed only 8% of weeks; the middle 60% missed 24%; the bottom 20% averaged 6 and missed 39%. In ten months nobody has scored IGL against that table. Scored: over the clean import window of the cohort scorecard database (site weeks 1–42), across all 65 members with 12 or more selling weeks — IGL averages 12.9 designs/week against a 12.4 median, rank 31 of 65, and published zero designs in 15 of its 27 design-reported selling weeks — 55.6%, rank 5 of 65. That is worse than Chris's bottom-tier benchmark of 39%. Meanwhile IGL averages 29.0 new ad creatives/week against a 17.0 median, rank 16 of 65. The "IGL runs huge creative volume" belief is true — of ads. It is false of designs, and it has been masking a rhythm problem for a year. No single shelf produces this. founders-kb supplies the reframe (mark the days you failed, not the output). wescale-kb's call archive supplies the benchmark nobody applied. The scorecard database supplies the number, which appears in no prose document in any of the nine knowledge bases. ## Where the belief came from (operator addition — the hunt did not find this) wescale-kb/igl-action-plan.md, created 2026-02-16, is the origin of IGL's standing "fix the funnel, not the factory" doctrine. Its gap table, stated as "Weeks 13–19 Scorecard Data", reads: | Metric | IGL | Median | Gap | Verdict in the file | |---|---|---|---|---| | Designs/wk | 22 | ~18 | +22% | Input is fine | | Creatives/wk | 39 | ~15 | +160% | Input is fine | Both numbers are averages that drop the zero weeks. Re-derived from the database over the file's own stated window: - New designs, weeks 13–19: 0, 0, 0, 0, 24, 12, 30. The three non-zero weeks average exactly 22.0 — the number in the table. All seven weeks average 9.4, which is below the ~18 cohort figure, not 22% above it. - New ad creatives, weeks 13–19: 0, 0, 46, 105, 3, 6, 36. The five non-zero weeks average 39.2 — the number in the table. All seven average 28.0. So the diagnosis that told IGL its factory was fine, and to spend seven months on the funnel, was produced by exactly the arithmetic this nugget names: excluding the empty weeks from the denominator. The card's finding and the defect that hid it are the same defect. ## Evidence trail - founders-kb/episodes/115-benjamin-franklin-an-american-life — seed, primary. Key Lesson 4: daily tracking of adherence to the thirteen virtues, "deliberate, measurable self-improvement applied consistently." Establishes that Franklin's reputation rested on observed daily rhythm. - mylibrary-kb/books/autobiography-of-benjamin-franklin — primary; same man, second document, counted as ONE strand, not independent confirmation. Supplies the mechanics: black marks for faults rather than credits for successes, one virtue per week in four 13-week courses, the weeding-one-bed analogy, the speckled-ax parable, and Dr. Baird's testimony. - wescale-kb/calls/call-08 — primary. Dated 2025-11-26 — a quotation from a 10-month-old document, not a live measurement. Carries the design-volume tiers (18 / 11.6 / 6 per week) and the consistency slice (8% / 24% / 39% of weeks with zero designs). Verified verbatim, including in wescale-kb/transcripts/call-08-transcript.md:170. - wescale-kb/scorecard/wescale-scorecard.db — primary, structured artifact. 51,411 observations, 92 members, 49 weeks. All cohort figures in this card were computed over the full 65-member population, not a sample. - wescale-kb/members/tim-dobyns — primary. The belief being corrected: line 36 "Target: 24 designs/week", line 99 "Target: 24 designs/week (above WeScale median of ~12-15/week)", line 222 "15/week → 24/week sustained". - wescale-kb/igl-action-plan.md — primary, operator-added. Created 2026-02-16, window "Weeks 13–19". The origin of the "+160% / input is fine" doctrine. Note it already carries a caveat line added by an earlier night saying "'Input is fine' is true against the median" — which this card shows is itself false for designs. - wescale-kb/themes/consistency-over-bursts — the single theme-aggregate. Cited only to show the prior art exists and that this move is its un-prescribed instance: it prescribes consistency in general and quotes Chris's "I'd rather see 10 ads every week for 8 weeks than 40 ads one week and 0 the next three", but names no IGL number and proposes no zero-week scorecard line. ## Worked examples (from the database, full result sets) - All 16 IGL selling weeks with no designs published, inside the clean window: site weeks 2, 3, 4, 5, 6, 8, 10, 11, 15, 16, 20, 21, 29, 32, 34, 38. Fifteen of these are explicit zeros; week 4 is a blank cell (a row exists, its value is null), so the honest count is 15 of 27 design-reported selling weeks. - Five consecutive empty weeks — site weeks 2–6 — immediately after publishing 21 in week 1. - Four more — weeks 10, 11, 15, 16 — immediately after publishing 36 in week 9. - The bursts: week 41 = 48 designs, week 42 = 72. Seventy-two is four times the top-20% weekly average of 18. - The nine members who never missed a week (0.0% zero-design weeks, ≥12 selling weeks): Liz U (51.2/wk), Lori Scott (42.7), Nicole Maxner (35.4), Dave Shaffer (31.1), Patrizia Heinzl (24.9), Kristin Hernandez (22.1), Joshua C.K. (20.9), Paula Jakobs (14.3), Kelli Sanders (10.5). High volume and no misses is a real quadrant — IGL is in neither half of it. ## Assay verdicts - novelty: PASS (4/5) — no catalogued nugget scores IGL against the cohort on rhythm; the cross-member percentile computation exists in no prose document in any base. It refutes a written belief rather than confirming one. - grounding: PASS (5/5) — the judge re-ran the computation independently and every figure matched. - actionability: PASS (5/5) — one new scorecard line with a stated target, plus a release queue at 8–12/week. Solo-executable, no capital. - Assay-office flag, carried forward honestly: this move directly contradicts another candidate in the same batch ("publish each finished design the day it clears quality control"). They cannot both ship. This card is the one that survived at 14/15; the other scored 10. ## Operator corrections (all re-derived from disk before publication) 1. 57.1% → 55.6%. Week 4 carries a new_designs row whose value is null — a blank cell, not a reported zero. Strict count: 15 zeros in 27 design-reported selling weeks. Under the lenient reading (blank counts as zero) it is 16 of 28 = 57.1%. Rank 5 of 65 holds under both. 2. 12.5 → 12.9 designs/week (strict; 12.46 lenient) against a 12.35 cohort median. Rank 31 of 65 holds under both. 3. Dave Shaffer: 51.9 → 31.1 designs/week. 51.9 is his weeks 1–49 figure — drawn from the dirty window the candidate itself excluded. Inside the stated window he is at 31.1. 4. The never-missed list is nine members, not seven. The candidate omitted Nicole Maxner and Kristin Hernandez. Eighth consecutive night the winner's only defect class is an incomplete enumeration. 5. Added: the origin of the belief (igl-action-plan.md, weeks 13–19), which the hunt did not find. See the section above. 6. Added: the belief's full footprint — seven files. wescale-kb/igl-action-plan.md (lines 9, 27, 28, 30) · wescale-kb/members/tim-dobyns.md · nuggets/catalog/nugget-001-...md (whose premise this corrects) · librarian/workflows/night-librarian.js:65, the Night Librarian's own standing context block · and three wiki-link asides in mylibrary-kb/books/my-life-and-work.md, recall-kb/Business/Marketing/The Tactics Trader Joe's Uses...md and recall-kb/Creativity/Content Creation/The 5 Levels of Creative Diversity...md. Git blame says all three asides were written by this librarian's own weavers (2026-08-06, 2026-08-12, 2026-09-05) — the night-#37 amplification pattern, repeating. Thread: founders-kb ep-115 Franklin's virtue chart (mark faults, not output) → mylibrary-kb autobiography for the chart's mechanics and Dr. Baird's cadence testimony → wescale-kb call-08's Six-Week Consistency Benchmark → the cohort scorecard database, full 65-member population → wescale-kb/members/tim-dobyns.md, the "24/week" belief it contradicts → (operator) wescale-kb/igl-action-plan.md, where that belief was manufactured.First, today, open Meta Ads Manager and pause these two ads by name:
| Ad name | Campaign | Live since | Spent | Returned |
|---|---|---|---|---|
Catalog_ShopAll_EssentialTees_Single_2026-CW36 | 30_CBO_CATALOG_SHOPALL_US_2026-CW36 | 2026-08-30 | $300.41 | $113.15 |
Catalog_NewArrivals_Home_Single_2026-CW36 | 30_CBO_CATALOG_NEWARRIVALS_US_2026-CW36 | 2026-08-30 | $300.66 | $108.17 |
Together: $601.07 spent, 6 purchases, $221.32 back — a 0.37 return. The Essential Tees one has made zero sales in its last seven days on $167.51. Your own break-even, written in your own plan, is a 1.87 return; these are running at one fifth of it. At 52.3% gross margin that is about $485 of gross profit destroyed in thirteen days, and they are still spending roughly $45–60 a day as of 2026-09-11.
Second, this weekend, do the read that was never done. Your ten-cell destination test (Catalog_A1..E2_..._CW33) ran 2026-08-12 to 2026-08-31, spent $2,881.00 and returned $2,431.36 on 47 purchases. It cleared both stopping conditions your own plan set — 19 elapsed days against a 14-day minimum, and 47 orders against a ~40-order floor. Apply the plan's own decision rule (contribution per purchase = average order value × 52.3% − cost per purchase, break-even ≈ $21) and the verdict is not "which destination won." It is all five lost money:
| Destination | Spend | Purchases | Revenue | Return | Cost per purchase | Contribution per purchase |
|---|---|---|---|---|---|---|
| Homepage | $605.84 | 16 | $805.07 | 1.33 | $37.87 | −$11.55 |
| Essential Tees collection | $763.18 | 14 | $732.95 | 0.96 | $54.51 | −$27.13 |
| Product pages (deep links) | $765.15 | 13 | $639.01 | 0.84 | $58.86 | −$33.15 |
| Top Picks collection | $373.99 | 2 | $155.31 | 0.42 | $187.00 | −$146.38 |
| Bundle & Save collection | $372.84 | 2 | $99.02 | 0.27 | $186.42 | −$160.53 |
Third, write two sentences into wescale-kb/members/tim-dobyns.md under WeScale Scorecard Health: "Catalog destination test CW33 completed 2026-08-31. No destination cleared break-even; the two collection-page destinations lost roughly six times as much per purchase as the homepage, so collection pages are out for catalog ads until something changes." Then amend section I of planning/igl-cro-audit-spec-for-iris-2026-06-18.md, line 116, which currently reads "Catalog ads must land on collections/all-products, NOT a thin 'best sellers' page" — that rule points at the worst-performing destination class you have measured.
You ran a complete, properly designed experiment, it finished twelve days ago, it says every destination you tested loses money — and in the three days after it ended you launched two more catalog ads that have since burned $601 without anyone reading the result.
Tobi Lutke's account of what nearly killed Shopify is not that people did bad work. It is that work happened where his attention was not. A Toronto team built supermarket features for months: "it just turned out that like, you know, in Toronto office, there was a fairly big project to build the features to add to Shopify so you could use it for running a supermarket." He explains why he never saw it: "People learned pretty quickly what of the projects that I'm interested in — talk me about both and nothing else. People learned how to direct my attention to the things that they wanted to have my attention." The fix was brute force: "The moment I went through absolutely every project I did myself, took like 60-hour days. I canceled probably 60% of the projects."
IGL's attention is directed at design count, because that is the number that goes into a cell every Tuesday. The destination test was not neglected out of laziness — the opposite. There is a careful reconciliation document dated 2026-08-27 that joined Meta to Shopify session by session, verified 20 completed checkouts against 20 Shopify orders with zero delta, and paused seven arms on the evidence. It is good work. But it read the test mid-flight, on two partial windows (2026-08-17 to 08-23, and 08-22 to 08-26), and it said so plainly: "This is an operational pruning decision, not a guaranteed statistical winner or scaling verdict." It kept two arms, D2 and SA1, as the survivors.
Then both survivors stopped spending on 2026-08-30 — and on that same day two brand-new catalog ads started. Nobody wrote the final read. The attention had moved to the next wave before the last one reported.
That is the non-obvious part, and it is worse than simple neglect. The mid-flight prune looked like the test had been handled. A test that gets partially read is more dangerous than one nobody touches, because the partial read consumes the feeling of having decided.
Two specifics make the full read worth the hour. First, the bottom of the table replicates and the top does not. Pool all four structured catalog waves (CW33, CW34, CW35, CW36 — $4,134 of spend) and the homepage advantage evaporates: Essential Tees 0.87, deep links 0.84, homepage 0.83, a three-way tie. What holds is the floor — Top Picks collection 0.29 and Bundle & Save 0.27 across every wave. The confident finding is which destinations to stop using, not which to scale. Second, that floor contradicts your own most recent audit. The 2026-08-09 ads and conversion audit called Bundle & Save "the best destination in the store" at a 6.09% conversion rate and recommended routing to it. /collections/bundle-save is still the single largest landing page in the store — 13,845 sessions in the last 90 days, nearly six times the biggest product page. For catalog ads specifically, it was the worst of the five.
No single knowledge base could produce this. The mastermind shelf holds the catalog doctrine the test was built from; the Skool classroom holds the argument for stopping; the founder shelf holds the mechanism that explains the failure. The result itself lives in an operations database outside all thirteen bases, which is exactly why it went twelve days unread.
Honest limits. These are Meta-attributed purchases and revenue, not the Shopify-verified, utm-matched orders your plan names as the judgment standard — Meta usually reports generously, so the real picture is likely worse, not better. Per-cell counts are small (0 to 13 purchases), so single-cell reads are directional. The two losing destinations also got about half the spend of the others and stopped earliest, so their ranking is partly a consequence of having been paused for losing. And the format half of the test is genuinely split: carousel beat single image in four of five destinations (0.99 versus 0.71 overall), but the single best cell in the grid was Home/Single at 1.46, and single image produced more purchases (25) on lower revenue.
All ten cells, ranked by return, 2026-08-12 to 2026-08-31:
| Cell | Destination | Format | Spend | Purchases | Revenue | Return |
|---|---|---|---|---|---|---|
| D2 | Homepage | single | $397.74 | 13 | $578.72 | 1.46 |
| E1 | Essential Tees | carousel | $381.75 | 9 | $490.22 | 1.28 |
| D1 | Homepage | carousel | $208.10 | 3 | $226.35 | 1.09 |
| A1 | Deep links | carousel | $399.00 | 7 | $421.22 | 1.06 |
| B1 | Top Picks | carousel | $188.38 | 1 | $119.80 | 0.64 |
| E2 | Essential Tees | single | $381.43 | 5 | $242.73 | 0.64 |
| A2 | Deep links | single | $366.15 | 6 | $217.79 | 0.59 |
| C1 | Bundle & Save | carousel | $188.57 | 2 | $99.02 | 0.53 |
| B2 | Top Picks | single | $185.61 | 1 | $35.51 | 0.19 |
| C2 | Bundle & Save | single | $184.27 | 0 | $0.00 | 0.00 |
The shape that actually works, for when you rebuild: Catalog_New Arrivals - Single Image - Top Picks Landing Page — New Arrivals product set, single image, pointed at the Top Picks landing page. Over the last 90 days (2026-06-14 to 2026-09-11) it spent $733.94 for 34 purchases and $1,498.46, a 2.04 return — the highest-revenue ad in the account over that window, and second by return only to Life's Too Short (2.67 on $507.45) among ads spending $300 or more. Account-wide over the same 90 days: $27,438.92 spent, 499 purchases, $25,265.60 returned, a 0.92 return across 395 ads. It stopped spending on 2026-08-31. That is the one catalog ad worth rebuilding, and note what it is: the Top Picks landing page, which is a different page from the /collections/top-picks-v2 destination that placed fourth in the grid.
Where to check any of this yourself: ```sql -- the completed test, one row per cell SELECT ad_name, MIN(date), MAX(date), ROUND(SUM(spend),2), SUM(purchases), ROUND(SUM(revenue),2) FROM fact_meta_daily WHERE level='ad' AND ad_name LIKE 'Catalog_%CW33%' GROUP BY ad_name ORDER BY SUM(revenue)/SUM(spend) DESC;
-- the two ads still running SELECT ad_name, date, spend, purchases, revenue FROM fact_meta_daily WHERE level='ad' AND ad_name LIKE 'Catalog_%CW36%' AND date >= '2026-09-05' ORDER BY date; `` Database: G:/iris/igl-ops-sage/data/igl.db`, synced through 2026-09-11.
founders-kb/interviews/009-tobi-lutke-transcript — primary. The seed, and the mechanism. The Toronto supermarket project he never knew about (line 34–36), the attention-routing admission and the fix (line 40). Raw transcript of the David Senra interview show, published 2026-01-18.wescale-kb/calls/call-28 — primary. Jakob & Petra's four-step catalog pipeline, the doctrine this test was built on: a low-budget tag-based catalog auto-runs on every launch, designs with 10+ clicks graduate out, and "always remove promoted designs from the testing catalog." It is the rotation discipline that the CW36 pair skipped. It also sets the depth gate at 10–15 purchases in 7 days — a bar no cell in the grid came close to.stress-free-scaling-kb/lessons/most-ad-accounts-need-two-things-not-twenty — primary. Skool classroom lesson, captured 2026-08-24: "unless you're consistently spending $100k+ per month toward a problem and persona, most of what you need to do in your ad account is remarkably simple." The argument for reading wave two before launching wave four.wescale-kb/themes/catalog-ads-playbook — theme-aggregate, the one permitted. Establishes that the cohort treats catalog ads as the discovery engine and statics as the scaler, and records the standing IGL prescription — which is what makes the unread result the gap, rather than the general advice.G:/iris/igl-ops-sage/data/igl.db — primary, outside all thirteen knowledge bases. fact_meta_daily: every figure in this card, re-derived by the operator. fact_landing_page_daily: /collections/bundle-save at 13,845 sessions over 90 days.planning/igl-cro-audit-spec-for-iris-2026-06-18.md — primary. Dated 2026-06-18, a spec, not a status record. Section I, line 116: catalog ads must land on collections/all-products. Line 119 hedges on format — "Catalog = carousel (more products visible), though test single-image for bestseller catalogs" — which the test vindicates.G:/iris-03a/deliverables/ads-test-implementation-plan-2026-08-10.md — primary. The test design and, crucially, its own read rule: collapse format twins into five destination buckets; judge on contribution per purchase (AOV × 52.3% − cost per purchase, break-even ≈ $21); minimum 14 elapsed days and a ~40-order floor before a verdict.G:/iris-03a/deliverables/catalog-14-arm-shopify-reconciliation-and-action-plan-2026-08-27.md — primary. The careful mid-flight prune: 14-arm registry, 20 checkouts reconciled to 20 Shopify orders with zero delta, seven arms paused, D2 and SA1 kept. Explicitly "not a guaranteed statistical winner or scaling verdict."G:/iris-03a/deliverables/sor-ads-cro-audit-2026-08-09.md — primary. Calls Bundle & Save "the best destination in the store" at 6.09% conversion (72 Shopify orders / 1,183 sessions, 30 days) — the claim the catalog test contradicts for catalog traffic specifically.Catalog_ShopAll_EssentialTees_Single_2026-CW36 | 30_CBO_CATALOG_SHOPALL_US_2026-CW36 | 2026-08-30 | $300.41 | $113.15 | | Catalog_NewArrivals_Home_Single_2026-CW36 | 30_CBO_CATALOG_NEWARRIVALS_US_2026-CW36 | 2026-08-30 | $300.66 | $108.17 | Together: $601.07 spent, 6 purchases, $221.32 back — a 0.37 return. The Essential Tees one has made zero sales in its last seven days on $167.51. Your own break-even, written in your own plan, is a 1.87 return; these are running at one fifth of it. At 52.3% gross margin that is about $485 of gross profit destroyed in thirteen days, and they are still spending roughly $45–60 a day as of 2026-09-11. Second, this weekend, do the read that was never done. Your ten-cell destination test (Catalog_A1..E2_..._CW33) ran 2026-08-12 to 2026-08-31, spent $2,881.00 and returned $2,431.36 on 47 purchases. It cleared both stopping conditions your own plan set — 19 elapsed days against a 14-day minimum, and 47 orders against a ~40-order floor. Apply the plan's own decision rule (contribution per purchase = average order value × 52.3% − cost per purchase, break-even ≈ $21) and the verdict is not "which destination won." It is all five lost money: | Destination | Spend | Purchases | Revenue | Return | Cost per purchase | Contribution per purchase | |---|---:|---:|---:|---:|---:|---:| | Homepage | $605.84 | 16 | $805.07 | 1.33 | $37.87 | −$11.55 | | Essential Tees collection | $763.18 | 14 | $732.95 | 0.96 | $54.51 | −$27.13 | | Product pages (deep links) | $765.15 | 13 | $639.01 | 0.84 | $58.86 | −$33.15 | | Top Picks collection | $373.99 | 2 | $155.31 | 0.42 | $187.00 | −$146.38 | | Bundle & Save collection | $372.84 | 2 | $99.02 | 0.27 | $186.42 | −$160.53 | Third, write two sentences into wescale-kb/members/tim-dobyns.md under WeScale Scorecard Health: "Catalog destination test CW33 completed 2026-08-31. No destination cleared break-even; the two collection-page destinations lost roughly six times as much per purchase as the homepage, so collection pages are out for catalog ads until something changes." Then amend section I of planning/igl-cro-audit-spec-for-iris-2026-06-18.md, line 116, which currently reads "Catalog ads must land on collections/all-products, NOT a thin 'best sellers' page" — that rule points at the worst-performing destination class you have measured. ## Why (the one-liner) You ran a complete, properly designed experiment, it finished twelve days ago, it says every destination you tested loses money — and in the three days after it ended you launched two more catalog ads that have since burned $601 without anyone reading the result. ## The insight Tobi Lutke's account of what nearly killed Shopify is not that people did bad work. It is that work happened where his attention was not. A Toronto team built supermarket features for months: "it just turned out that like, you know, in Toronto office, there was a fairly big project to build the features to add to Shopify so you could use it for running a supermarket." He explains why he never saw it: "People learned pretty quickly what of the projects that I'm interested in — talk me about both and nothing else. People learned how to direct my attention to the things that they wanted to have my attention." The fix was brute force: "The moment I went through absolutely every project I did myself, took like 60-hour days. I canceled probably 60% of the projects." IGL's attention is directed at design count, because that is the number that goes into a cell every Tuesday. The destination test was not neglected out of laziness — the opposite. There is a careful reconciliation document dated 2026-08-27 that joined Meta to Shopify session by session, verified 20 completed checkouts against 20 Shopify orders with zero delta, and paused seven arms on the evidence. It is good work. But it read the test mid-flight, on two partial windows (2026-08-17 to 08-23, and 08-22 to 08-26), and it said so plainly: "This is an operational pruning decision, not a guaranteed statistical winner or scaling verdict." It kept two arms, D2 and SA1, as the survivors. Then both survivors stopped spending on 2026-08-30 — and on that same day two brand-new catalog ads started. Nobody wrote the final read. The attention had moved to the next wave before the last one reported. That is the non-obvious part, and it is worse than simple neglect. The mid-flight prune looked like the test had been handled. A test that gets partially read is more dangerous than one nobody touches, because the partial read consumes the feeling of having decided. Two specifics make the full read worth the hour. First, the bottom of the table replicates and the top does not. Pool all four structured catalog waves (CW33, CW34, CW35, CW36 — $4,134 of spend) and the homepage advantage evaporates: Essential Tees 0.87, deep links 0.84, homepage 0.83, a three-way tie. What holds is the floor — Top Picks collection 0.29 and Bundle & Save 0.27 across every wave. The confident finding is which destinations to stop using, not which to scale. Second, that floor contradicts your own most recent audit. The 2026-08-09 ads and conversion audit called Bundle & Save "the best destination in the store" at a 6.09% conversion rate and recommended routing to it. /collections/bundle-save is still the single largest landing page in the store — 13,845 sessions in the last 90 days, nearly six times the biggest product page. For catalog ads specifically, it was the worst of the five. No single knowledge base could produce this. The mastermind shelf holds the catalog doctrine the test was built from; the Skool classroom holds the argument for stopping; the founder shelf holds the mechanism that explains the failure. The result itself lives in an operations database outside all thirteen bases, which is exactly why it went twelve days unread. Honest limits. These are Meta-attributed purchases and revenue, not the Shopify-verified, utm-matched orders your plan names as the judgment standard — Meta usually reports generously, so the real picture is likely worse, not better. Per-cell counts are small (0 to 13 purchases), so single-cell reads are directional. The two losing destinations also got about half the spend of the others and stopped earliest, so their ranking is partly a consequence of having been paused for losing. And the format half of the test is genuinely split: carousel beat single image in four of five destinations (0.99 versus 0.71 overall), but the single best cell in the grid was Home/Single at 1.46, and single image produced more purchases (25) on lower revenue. ## Worked examples All ten cells, ranked by return, 2026-08-12 to 2026-08-31: | Cell | Destination | Format | Spend | Purchases | Revenue | Return | |---|---|---|---:|---:|---:|---:| | D2 | Homepage | single | $397.74 | 13 | $578.72 | 1.46 | | E1 | Essential Tees | carousel | $381.75 | 9 | $490.22 | 1.28 | | D1 | Homepage | carousel | $208.10 | 3 | $226.35 | 1.09 | | A1 | Deep links | carousel | $399.00 | 7 | $421.22 | 1.06 | | B1 | Top Picks | carousel | $188.38 | 1 | $119.80 | 0.64 | | E2 | Essential Tees | single | $381.43 | 5 | $242.73 | 0.64 | | A2 | Deep links | single | $366.15 | 6 | $217.79 | 0.59 | | C1 | Bundle & Save | carousel | $188.57 | 2 | $99.02 | 0.53 | | B2 | Top Picks | single | $185.61 | 1 | $35.51 | 0.19 | | C2 | Bundle & Save | single | $184.27 | 0 | $0.00 | 0.00 | The shape that actually works, for when you rebuild: Catalog_New Arrivals - Single Image - Top Picks Landing Page — New Arrivals product set, single image, pointed at the Top Picks landing page. Over the last 90 days (2026-06-14 to 2026-09-11) it spent $733.94 for 34 purchases and $1,498.46, a 2.04 return — the highest-revenue ad in the account over that window, and second by return only to Life's Too Short (2.67 on $507.45) among ads spending $300 or more. Account-wide over the same 90 days: $27,438.92 spent, 499 purchases, $25,265.60 returned, a 0.92 return across 395 ads. It stopped spending on 2026-08-31. That is the one catalog ad worth rebuilding, and note what it is: the Top Picks landing page, which is a different page from the /collections/top-picks-v2 destination that placed fourth in the grid. Where to check any of this yourself: ``sql -- the completed test, one row per cell SELECT ad_name, MIN(date), MAX(date), ROUND(SUM(spend),2), SUM(purchases), ROUND(SUM(revenue),2) FROM fact_meta_daily WHERE level='ad' AND ad_name LIKE 'Catalog_%CW33%' GROUP BY ad_name ORDER BY SUM(revenue)/SUM(spend) DESC; -- the two ads still running SELECT ad_name, date, spend, purchases, revenue FROM fact_meta_daily WHERE level='ad' AND ad_name LIKE 'Catalog_%CW36%' AND date >= '2026-09-05' ORDER BY date; ` Database: G:/iris/igl-ops-sage/data/igl.db, synced through 2026-09-11. ## Evidence trail - founders-kb/interviews/009-tobi-lutke-transcript — primary. The seed, and the mechanism. The Toronto supermarket project he never knew about (line 34–36), the attention-routing admission and the fix (line 40). Raw transcript of the David Senra interview show, published 2026-01-18. - wescale-kb/calls/call-28 — primary. Jakob & Petra's four-step catalog pipeline, the doctrine this test was built on: a low-budget tag-based catalog auto-runs on every launch, designs with 10+ clicks graduate out, and "always remove promoted designs from the testing catalog." It is the rotation discipline that the CW36 pair skipped. It also sets the depth gate at 10–15 purchases in 7 days — a bar no cell in the grid came close to. - stress-free-scaling-kb/lessons/most-ad-accounts-need-two-things-not-twenty — primary. Skool classroom lesson, captured 2026-08-24: "unless you're consistently spending $100k+ per month toward a problem and persona, most of what you need to do in your ad account is remarkably simple." The argument for reading wave two before launching wave four. - wescale-kb/themes/catalog-ads-playbook — theme-aggregate, the one permitted. Establishes that the cohort treats catalog ads as the discovery engine and statics as the scaler, and records the standing IGL prescription — which is what makes the unread result the gap, rather than the general advice. - G:/iris/igl-ops-sage/data/igl.db — primary, outside all thirteen knowledge bases. fact_meta_daily: every figure in this card, re-derived by the operator. fact_landing_page_daily: /collections/bundle-save at 13,845 sessions over 90 days. - planning/igl-cro-audit-spec-for-iris-2026-06-18.md — primary. Dated 2026-06-18, a spec, not a status record. Section I, line 116: catalog ads must land on collections/all-products. Line 119 hedges on format — "Catalog = carousel (more products visible), though test single-image for bestseller catalogs" — which the test vindicates. - G:/iris-03a/deliverables/ads-test-implementation-plan-2026-08-10.md — primary. The test design and, crucially, its own read rule: collapse format twins into five destination buckets; judge on contribution per purchase (AOV × 52.3% − cost per purchase, break-even ≈ $21); minimum 14 elapsed days and a ~40-order floor before a verdict. - G:/iris-03a/deliverables/catalog-14-arm-shopify-reconciliation-and-action-plan-2026-08-27.md — primary. The careful mid-flight prune: 14-arm registry, 20 checkouts reconciled to 20 Shopify orders with zero delta, seven arms paused, D2 and SA1 kept. Explicitly "not a guaranteed statistical winner or scaling verdict." - G:/iris-03a/deliverables/sor-ads-cro-audit-2026-08-09.md — primary. Calls Bundle & Save "the best destination in the store" at 6.09% conversion (72 Shopify orders / 1,183 sessions, 30 days) — the claim the catalog test contradicts for catalog traffic specifically. ## Assay verdicts - Novelty — PASS (4/5). No prior art in any of the thirteen knowledge bases: a grep for CW33, TopPicks_Home and "destination verdict" across every .md file returns zero hits, and none of the 37 prior nuggets reads this test. Operator amendment: the judge's phrasing "nothing has read it" was too strong and is corrected in this card — the 2026-08-27 reconciliation read it mid-flight and pruned on it. What is new is the completed-test verdict, the contribution-per-purchase read, the pooled four-wave floor, and the two ads still running. - Grounding — PASS (5/5). Every figure re-derived independently by the operator from igl.db; all document quotations checked verbatim. Four corrections applied before publication (below). - Actionability — PASS (5/5). Two ads named exactly, a rule amended at a named file and line, a decision recorded in a named block. One sitting, no new spend, no staff. ## Operator corrections applied before publication 1. "The homepage won" — removed. True within CW33 alone (1.33). Pooling all four structured waves ($4,134) drops the homepage to 0.83, third of five and tied with the other two. What replicates is the floor, not the peak. 2. "All ten cells… the homepage won and the collection pages lost" reframed to "all five lost money." By the plan's own decision metric every destination came back negative; the best was −$11.55 per purchase. 3. "Nothing in the knowledge bases has read it" narrowed. True of the knowledge bases, false of the workspace: the 2026-08-27 reconciliation document pruned seven arms on partial windows. 4. "The single best-performing ad in your account" corrected to "the highest-revenue ad." Life's Too Short returned 2.67 against its 2.04. 5. "Instead of running a third wave blind" corrected. Three further waves (CW34, CW35, CW36) launched after CW33 began; the one still running is the fourth. 6. "More than six times any product page" corrected to "nearly six times" — 13,845 against 2,325 is 5.96×. 7. Move's format instruction de-contradicted. The hunt prescribed carousel while naming a single-image ad as the shape to copy. The card now reports the split honestly rather than resolving it by assertion. Thread: Lutke on work happening where attention is not → the mastermind's catalog pipeline doctrine the test was built from → the catalog-ads playbook's standing IGL prescription → the ten CW33 cells in fact_meta_daily`, finished and unread → the June landing-page rule the result contradicts → the Skool lesson arguing against launching the next wave first.Delete the sentence "explicit religious messaging tested at 0% sales" everywhere it appears and replace it with the measured line: "Faith-angle ads, 2024-11-05 to 2026-08-27: 39 purchase-optimised ads, $7,350.70 spend, 215 purchases, $9,685.35 revenue, 1.32 return on ad spend; best single ad cost $22.74 per purchase against the account's $29.03 (igl.db, synced 2026-09-10)." It is in 15 files, not the 4 you would guess — five are the doctrine bullets themselves, and ten are wiki-link asides the Night Librarian itself wrote into other knowledge bases between 2026-08-08 and 2026-09-09, so deleting the five originals leaves ten live copies behind. The full list with line numbers is below; start with igl-brand-kb/voice/divine-language-tiers.md:23, because that file is also mirrored as a live skill at C:/Users/Tim Dobyns/.claude/skills/igl-brand-guidelines/divine-language-tiers.md. Leave the 60/30/10 tier split alone for now and mark it untested rather than validated — the cap rests entirely on the line you are deleting.
The number that caps faith at 10% of IGL's content comes from a December 2025 document whose very next line claims humor designs sold 0 units, when the bird-feeder family had already sold 4,176.
The seed asked whether faith designs can work in cold paid social. Before answering, the thread chased where IGL's existing answer came from. It traces to one bullet in igl-brand-kb/evolution/strategic-synthesis-v1-1.md — Status: HISTORICAL, Last Updated December 2025 — reading "Explicit religious messaging = 0% sales." That bullet was copied verbatim into four files that declare Status: CURRENT SoR, including content/three-tier-content-system.md, where it sits under the heading "The tier system is not theoretical — it is validated by sales data."
Three lines below it, the same December 2025 document says: "Cardinal Birds (24 units) vs. humor designs (0 units) validates thesis approach." As of 2025-12-31 the "Excuse Me, Bird Feeder's Empty" family had sold 4,176 units and $132,486 — 53% of IGL's entire 7,859 units and $248,429 to that date. The Cardinal Birds Graphic T-Shirt has sold 70 units lifetime, not 24. If the humor half of the comparison is wrong by 4,176 units, the faith half carries no authority either.
And it is wrong. In the live ad account, 39 purchase-optimised ads classified copy_angle = FAITH spent $7,350.70 and produced 215 purchases and $9,685.35 — a 1.32 return on ad spend. The best of them, "Catalog_New Arrivals - Single Image - Top Picks Landing Page," returned $1,810.28 on $955.09 (1.90, 42 purchases, $22.74 per purchase) — better than the account-wide $29.03 across all 2,178 ads. That is not zero. Honest caveat: the strongest faith-angle performers are catalog ads, so their revenue spans the whole catalog rather than faith designs alone. The claim this kills is "0%", not "faith designs are proven winners in cold social."
No single knowledge base could produce this. The brand shelf holds the claim and repeats it as doctrine; the sales and ad data that falsify it live in an operations database outside all six knowledge bases, which is exactly why the line went unchecked for nine months. Claude Hopkins is why it matters: his rule in Founders #207 is that mail-order is the honest standard because "Every claim is tested by sales." A claim that was never tested by sales has been governing IGL's content mix — and, through this librarian's own link-weaving, spreading.
igl-brand-kb/evolution/strategic-synthesis-v1-1 — theme-aggregate. Status HISTORICAL, December 2025. Line 80 carries the claim; line 83 carries the "humor designs (0 units)" comparison that falsifies its own method.G:/iris/igl-ops-sage/data/igl.db — primary, outside the six knowledge bases. fact_product_sales_daily: bird-feeder family 4,176 units / $132,486.09 through 2025-12-31 against 7,859 total units / $248,429.14; Cardinal Birds Graphic T-Shirt 70 units lifetime. fact_meta_daily joined to dim_ad_classifications: 39 faith-angle purchase-optimised ads, $7,350.70, 215 purchases, $9,685.35.founders-kb/episodes/207-claude-hopkins-scientific-advertising — primary. Line 40: "Every claim is tested by sales." Line 30: "The weight of an argument may often be multiplied by making it specific" — why an undated, unsourced "0%" carried authority for nine months.igl-brand-kb/audience/profiles/03-ruth-slow-living-prophet — primary. Status CURRENT SoR, created 2026-08-29. "The overlooked thing is the holy thing" — faith without faith vocabulary, the lever a blanket ban on explicit language obscured.mylibrary-kb/books/braiding-sweetgrass — primary. The source of the implicit sacred register ("seeing the world as a gift rather than a commodity") the tier system calls Tier 1 and treats as the safe alternative.igl-brand-kb/voice/divine-language-tiers.md:23, because that file is also mirrored as a live skill at C:/Users/Tim Dobyns/.claude/skills/igl-brand-guidelines/divine-language-tiers.md. Leave the 60/30/10 tier split alone for now and mark it untested rather than validated — the cap rests entirely on the line you are deleting. ## Why (the one-liner) The number that caps faith at 10% of IGL's content comes from a December 2025 document whose very next line claims humor designs sold 0 units, when the bird-feeder family had already sold 4,176. ## The insight The seed asked whether faith designs can work in cold paid social. Before answering, the thread chased where IGL's existing answer came from. It traces to one bullet in igl-brand-kb/evolution/strategic-synthesis-v1-1.md — Status: HISTORICAL, Last Updated December 2025 — reading "Explicit religious messaging = 0% sales." That bullet was copied verbatim into four files that declare Status: CURRENT SoR, including content/three-tier-content-system.md, where it sits under the heading "The tier system is not theoretical — it is validated by sales data." Three lines below it, the same December 2025 document says: "Cardinal Birds (24 units) vs. humor designs (0 units) validates thesis approach." As of 2025-12-31 the "Excuse Me, Bird Feeder's Empty" family had sold 4,176 units and $132,486 — 53% of IGL's entire 7,859 units and $248,429 to that date. The Cardinal Birds Graphic T-Shirt has sold 70 units lifetime, not 24. If the humor half of the comparison is wrong by 4,176 units, the faith half carries no authority either. And it is wrong. In the live ad account, 39 purchase-optimised ads classified copy_angle = FAITH spent $7,350.70 and produced 215 purchases and $9,685.35 — a 1.32 return on ad spend. The best of them, "Catalog_New Arrivals - Single Image - Top Picks Landing Page," returned $1,810.28 on $955.09 (1.90, 42 purchases, $22.74 per purchase) — better than the account-wide $29.03 across all 2,178 ads. That is not zero. Honest caveat: the strongest faith-angle performers are catalog ads, so their revenue spans the whole catalog rather than faith designs alone. The claim this kills is "0%", not "faith designs are proven winners in cold social." No single knowledge base could produce this. The brand shelf holds the claim and repeats it as doctrine; the sales and ad data that falsify it live in an operations database outside all six knowledge bases, which is exactly why the line went unchecked for nine months. Claude Hopkins is why it matters: his rule in Founders #207 is that mail-order is the honest standard because "Every claim is tested by sales." A claim that was never tested by sales has been governing IGL's content mix — and, through this librarian's own link-weaving, spreading. ## The full population — all 15 files (operator-enumerated, not sampled) The five doctrine bullets (the originals): | File | Line | Status | |---|---|---| | igl-brand-kb/evolution/strategic-synthesis-v1-1.md | 80 | HISTORICAL, Dec 2025 — the origin | | igl-brand-kb/voice/divine-language-tiers.md | 23 | CURRENT SoR, 2026-02-10 | | igl-brand-kb/collections/collection-strategy.md | 27 | CURRENT SoR, 2026-02-10 | | igl-brand-kb/core/strategic-synthesis-v4.md | 571 | CURRENT SoR, 2026-02-11 | | igl-brand-kb/content/three-tier-content-system.md | 187 | CURRENT SoR, March 2026 | The ten propagated copies — every one a Night Librarian wiki-link aside: | File | Line | Written by night of | |---|---|---| | competitors-kb/trend-reports/shopify-gardening-apparel-trend-report.md | 912 | 2026-08-08 | | competitors-kb/hardcore-catholics/README.md | 76 | 2026-08-11 | | wescale-kb/members/beth-waller.md | 106 | 2026-08-12 | | recall-kb/Creativity/Content Creation/ChatGPT 03_IGL_MKTG_Andrew-Tim_GPT_Flip script examples.md | 36 | 2026-08-12 | | mylibrary-kb/books/we-make-the-road-by-walking.md | 254 | 2026-08-18 | | competitors-kb/miners/faith-gardening/module-b-wescale-transplant.md | 32 | 2026-08-22 | | mylibrary-kb/books/the-social-principles-of-jesus.md | 124 | 2026-08-22 | | igl-brand-kb/competitive/faith-gardening-pod-brands.md | 7 | 2026-08-25 | | igl-brand-kb/collections/faith-gardening-collection-strategy-v2.md | 25 | 2026-08-26 | | recall-kb/Philosophy/Spirituality/Brian McLaren Faith after Fundamentalism.md | 69 | 2026-09-09 | One further file, igl-brand-kb/competitive/top-20-premium-brands.md:250, matches the search string but is a false positive — "25–50% sales" is a discount, not the claim. It needs no edit. ## Evidence trail - igl-brand-kb/evolution/strategic-synthesis-v1-1 — theme-aggregate. Status HISTORICAL, December 2025. Line 80 carries the claim; line 83 carries the "humor designs (0 units)" comparison that falsifies its own method. - G:/iris/igl-ops-sage/data/igl.db — primary, outside the six knowledge bases. fact_product_sales_daily: bird-feeder family 4,176 units / $132,486.09 through 2025-12-31 against 7,859 total units / $248,429.14; Cardinal Birds Graphic T-Shirt 70 units lifetime. fact_meta_daily joined to dim_ad_classifications: 39 faith-angle purchase-optimised ads, $7,350.70, 215 purchases, $9,685.35. - founders-kb/episodes/207-claude-hopkins-scientific-advertising — primary. Line 40: "Every claim is tested by sales." Line 30: "The weight of an argument may often be multiplied by making it specific" — why an undated, unsourced "0%" carried authority for nine months. - igl-brand-kb/audience/profiles/03-ruth-slow-living-prophet — primary. Status CURRENT SoR, created 2026-08-29. "The overlooked thing is the holy thing" — faith without faith vocabulary, the lever a blanket ban on explicit language obscured. - mylibrary-kb/books/braiding-sweetgrass — primary. The source of the implicit sacred register ("seeing the world as a gift rather than a commodity") the tier system calls Tier 1 and treats as the safe alternative. ## Assay verdicts - Novelty — PASS (4/5). Nuggets #012 and #008 both flagged this line as unverified ("a quotation, not a measurement"); neither falsified it. This traces it to its origin document and kills it with its own companion line. - Grounding — PASS (5/5). Document chain and database both verified end to end, then independently re-derived by the operator from disk. - Consequence — PASS (5/5). A false line propagating into current source-of-record files, one of them under a heading claiming it is "validated by sales data." ## Operator corrections applied to the hunt's draft 1. Population: 5 files → 15. The hunt named five and did not run the search across the whole workspace. Seventh consecutive night the winner's only defect is an incomplete enumeration, and the first since the full-result-set rule shipped on night #34. 2. Ten of the fifteen are this librarian's own wiki-link asides, dated 2026-08-08 through 2026-09-09 — git-blamed, all ten. The hunt did not detect that the claim had been spreading, or by whose hand. 3. Faith-angle revenue $9,703 → $9,685.35 (re-derived; the 1.32 return on ad spend is unchanged). 4. Date range 2026-08-31 → 2026-08-27, the true last day of faith-angle spend. 5. Caveat added that the top faith-angle performers are catalog ads. Thread: the faith transition (live problem) → igl.db dim_ad_classifications copy_angle='FAITH' (39 ads, 1.32) → igl-brand-kb/voice/divine-language-tiers.md "0% sales" → traced back to evolution/strategic-synthesis-v1-1.md (Dec 2025, HISTORICAL) → its companion line "humor designs (0 units)" → fact_product_sales_daily: 4,176 bird-feeder units by 2025-12-31 → Founders #207 Hopkins "every claim is tested by sales" → Ruth dossier → Braiding SweetgrassOpen the ad rule contract in G:\iris\igl-ops-sage\data\igl.db before you let it drive anything: select contract_json from ops_ad_rule_contracts where rule_contract_id = 'ads-rule-contract-step6-fixture-v1'. It was approved 2026-08-02 and still carries "status": "FIXTURE_ONLY_CONFLICT_REMAINS_UNRESOLVED". It sets run_roas_floor to 1.45 at spend_usd 250 — meaning: when an ad has spent $250, kill it if it has not returned $1.45 for every $1. Backtested against every ad in fact_meta_daily that ever reached $250 of lifetime spend (111 ads, 2024-08-01 to 2026-09-09), that floor cuts 57 of them, avoids $46,832 of later spend and forfeits $66,857 of later revenue: net minus $20,025. Make two edits. (1) Change the $250 checkpoint from a cut gate into a raise-budget-only gate — keep the 1.87 scale_roas_floor as the trigger to increase budget, and delete 1.45 as a trigger to stop anything. (2) Promote protected_winner (currently "bounded_scale_status": "DESIGN_APPROVED_BUT_NOT_YET_AN_EXECUTABLE_RULE") to executable before you switch anything else on, since it is the only clause in the contract that protects a winner rather than killing one. Keep the zero_purchase hard stop at $54 exactly as written — it is the one rule that costs nothing.
Cutting ads at a return floor has never once paid off in IGL's own ad history, and the floor written into the draft contract is the most expensive one tested.
Schwartz's awareness scale — tonight's seed section — is a diagnosis whose clock runs in years. He puts a "psychological wall" between each stage, and says a wrongly-aimed headline meets indifference, not a weak signal. His Completely Unaware strategy explicitly forbids naming the product in the headline, and his rule for what a headline owes you is brutally modest: "Your headline has only one job — to stop your prospect and compel him to read the second sentence of your ad." That is a deliberate sacrifice of the early number. The ad is built to earn a second sentence, not to convert on day two.
The mastermind decides on the opposite clock. Call #32 (2026-06-04, Chris Heckman) formalised the Bump and Cut tree: spend to 2× cost-per-purchase — about 48 hours at $25/day — then "under 2.0 = cut." igl-action-plan Action 3.3 hardens it further, sourcing Jakob's "Kill anything under 1.8 ROAS after 2-3 days" and instructing IGL to "Set hard kill rule. No exceptions." One member dissents, and only softly: Eric Corry's profile says "Don't kill ads after 24 hours for bad ROAS... give the campaign time." Nobody has ever said how much time.
Run that clock backwards over IGL's own history and it eats the business. Of the 43 ads that ever produced $1,000 or more in Meta-reported revenue ($236,960 combined), 26 were returning under 1.8 at their first $50 of spend, and those 26 went on to produce $169,763. The median winner takes six days just to reach $50. The largest ad in the account's history — "Excuse Me The Bird Feeder's Empty," $16,316 spent, $38,221 returned, 2.34 lifetime over 221 active days — had earned $34.78 when it crossed $50 on day eight. A 0.69. Any 1.8-after-three-days rule deletes it. "Life's Too Short," IGL's best ad of 2026, was at 0.88 when it crossed $50 on day six, then ran 137 days to 209 purchases at 2.22 — and the seven words on that shirt were written by a Tomato Lovers Collective member named Jennifer Fox, not by a designer.
The closing argument is the one that usually kills a finding like this: freed money is not destroyed, it buys more of something else. Except the something else is worse. The 57 ads the 1.45 floor cuts returned 1.43 on their post-checkpoint spend, while money spent on any ad after it had already crossed $250 returned 1.73 ($224,259 on $129,429) — against 1.50 for the account all-time and 1.24 in 2026 so far. Old ads are IGL's best dollars, and the floor spends them on new ones. Ogilvy's shelf already said so, via Albert Lasker: "No idea is big unless it will work for 30 years," with the episode's own gloss — "Retiring a winning ad is a decision made for the agency's boredom, not the client's results."
No single shelf reaches this. The book explains why early returns are structurally slow, the mastermind supplies the numeric rule, the founders shelf supplies the counter-principle, and only the operations database can say what the rule would actually have cost.
mylibrary-kb/books/breakthrough-advertising — tonight's seed, section "Framework: The 5 Levels of Market Awareness" (lines 288–303) plus the Chapter 2 notes at lines 46, 57 and 408. Supplies the "psychological wall"/indifference framing and the headline's "one job… the second sentence" rule — why a correctly aimed ad reads as a zero early. Enriched 2026-02-26; a book note, no business figures taken from it.founders-kb/episodes/082-david-ogilvy-ogilvy-advertising — lines 80–84. Lasker via Ogilvy, "No idea is big unless it will work for 30 years," and "Retiring a winning ad is a decision made for the agency's boredom, not the client's results." The long-horizon counterweight.wescale-kb/calls/call-32 — 2026-06-04, Chris Heckman. The Bump and Cut Decision Tree, verbatim at lines 31 and 42: "spend to 2× CPA (~48 hours at $25/day)… under 2.0 = cut."wescale-kb/members/eric-corry — line 164, the cohort's lone dissent on horizon: "Don't kill ads after 24 hours for bad ROAS." Never quantified — the gap this nugget fills.wescale-kb/igl-action-plan — theme-aggregate, and dated: created 2026-02-16, seven months old, and a plan rather than a result. Action 3.3, lines 127–130: Jakob's "Kill anything under 1.8 ROAS after 2-3 days," with the IGL instruction "Set hard kill rule. No exceptions." This is the prescription the backtest contradicts.igl-brand-kb/community/lauren-and-rowena-stories — lines 56–58 and 174: Tomato Lovers Collective member Jennifer Fox submitted "Life's too short to grow boring tomatoes," which the file calls one of IGL's best-selling designs.G:\iris\igl-ops-sage\data\igl.db — tables fact_meta_daily (ad level, 2024-08-01 to 2026-09-09) and ops_ad_rule_contracts. Every figure here that is not a document quotation comes from this database, re-derived independently by the operator on 2026-09-10. All returns are Meta-reported, not Shopify-verified. The knowledge bases hold no ad-level or product-level sales data at all, so none of these numbers could have come from them.G:\iris\igl-ops-sage\data\igl.db before you let it drive anything: select contract_json from ops_ad_rule_contracts where rule_contract_id = 'ads-rule-contract-step6-fixture-v1'. It was approved 2026-08-02 and still carries "status": "FIXTURE_ONLY_CONFLICT_REMAINS_UNRESOLVED". It sets run_roas_floor to 1.45 at spend_usd 250 — meaning: when an ad has spent $250, kill it if it has not returned $1.45 for every $1. Backtested against every ad in fact_meta_daily that ever reached $250 of lifetime spend (111 ads, 2024-08-01 to 2026-09-09), that floor cuts 57 of them, avoids $46,832 of later spend and forfeits $66,857 of later revenue: net minus $20,025. Make two edits. (1) Change the $250 checkpoint from a cut gate into a raise-budget-only gate — keep the 1.87 scale_roas_floor as the trigger to increase budget, and delete 1.45 as a trigger to stop anything. (2) Promote protected_winner (currently "bounded_scale_status": "DESIGN_APPROVED_BUT_NOT_YET_AN_EXECUTABLE_RULE") to executable before you switch anything else on, since it is the only clause in the contract that protects a winner rather than killing one. Keep the zero_purchase hard stop at $54 exactly as written — it is the one rule that costs nothing. ## Why (the one-liner) Cutting ads at a return floor has never once paid off in IGL's own ad history, and the floor written into the draft contract is the most expensive one tested. ## The insight Schwartz's awareness scale — tonight's seed section — is a diagnosis whose clock runs in years. He puts a "psychological wall" between each stage, and says a wrongly-aimed headline meets indifference, not a weak signal. His Completely Unaware strategy explicitly forbids naming the product in the headline, and his rule for what a headline owes you is brutally modest: "Your headline has only one job — to stop your prospect and compel him to read the second sentence of your ad." That is a deliberate sacrifice of the early number. The ad is built to earn a second sentence, not to convert on day two. The mastermind decides on the opposite clock. Call #32 (2026-06-04, Chris Heckman) formalised the Bump and Cut tree: spend to 2× cost-per-purchase — about 48 hours at $25/day — then "under 2.0 = cut." igl-action-plan Action 3.3 hardens it further, sourcing Jakob's "Kill anything under 1.8 ROAS after 2-3 days" and instructing IGL to "Set hard kill rule. No exceptions." One member dissents, and only softly: Eric Corry's profile says "Don't kill ads after 24 hours for bad ROAS... give the campaign time." Nobody has ever said how much time. Run that clock backwards over IGL's own history and it eats the business. Of the 43 ads that ever produced $1,000 or more in Meta-reported revenue ($236,960 combined), 26 were returning under 1.8 at their first $50 of spend, and those 26 went on to produce $169,763. The median winner takes six days just to reach $50. The largest ad in the account's history — "Excuse Me The Bird Feeder's Empty," $16,316 spent, $38,221 returned, 2.34 lifetime over 221 active days — had earned $34.78 when it crossed $50 on day eight. A 0.69. Any 1.8-after-three-days rule deletes it. "Life's Too Short," IGL's best ad of 2026, was at 0.88 when it crossed $50 on day six, then ran 137 days to 209 purchases at 2.22 — and the seven words on that shirt were written by a Tomato Lovers Collective member named Jennifer Fox, not by a designer. The closing argument is the one that usually kills a finding like this: freed money is not destroyed, it buys more of something else. Except the something else is worse. The 57 ads the 1.45 floor cuts returned 1.43 on their post-checkpoint spend, while money spent on any ad after it had already crossed $250 returned 1.73 ($224,259 on $129,429) — against 1.50 for the account all-time and 1.24 in 2026 so far. Old ads are IGL's best dollars, and the floor spends them on new ones. Ogilvy's shelf already said so, via Albert Lasker: "No idea is big unless it will work for 30 years," with the episode's own gloss — "Retiring a winning ad is a decision made for the agency's boredom, not the client's results." No single shelf reaches this. The book explains why early returns are structurally slow, the mastermind supplies the numeric rule, the founders shelf supplies the counter-principle, and only the operations database can say what the rule would actually have cost. ## Worked examples — the full result set, not a sample The floor ladder at the contract's own $250 checkpoint (population: all 111 ads that ever reached $250 lifetime spend). Net = later spend avoided minus later revenue forfeited: | Floor | Ads cut | Spend avoided | Revenue forfeited | Net | |---|---|---|---|---| | 0.70 | 12 | $7,352 | $8,920 | −$1,567 | | 1.00 | 26 | $15,153 | $19,853 | −$4,701 | | 1.20 | 37 | $21,827 | $27,939 | −$6,112 | | 1.45 (the contract's) | 57 | $46,832 | $66,857 | −$20,025 | | 1.87 (the contract's scale floor) | 82 | $68,514 | $104,133 | −$35,619 | Tested at four checkpoints — $100, $250, $500 and $1,000 — and no floor at any checkpoint ever made money. At $100 the same ladder runs from −$3,585 to −$45,365; at $500 from −$57 to −$38,266; at $1,000 from $0 to −$32,781. (Precision: at the $1,000 checkpoint the 0.70 and 1.00 floors cut nothing at all, so they lose nothing. They are neutral, not profitable — the honest statement is that no floor ever gains, not that every floor loses.) Named ads, all figures Meta-reported, from fact_meta_daily: - Excuse Me The Bird Feeder's Empty (ad_id 120212557648610726) — $16,316 spend, $38,221 revenue, 2.34, 1,004 purchases, 221 active days. At its first $50, on day eight: 0.69. At the $250 checkpoint: 3.35. It survives the 1.45 floor, but not the cohort's 1.8-after-three-days rule. - Life's Too Short (ad_id 120239217019100726) — 2026-02-06 to 2026-07-18, 137 days, $5,533 spend, $12,258 revenue, 209 purchases, 2.22. At its first $50, day six: 0.88. At the $250 checkpoint: 4.26. Has not delivered since 2026-07-18. - Advantage+ New Release – Bundle & Save Landing Page (ad_id 120212117651740726) — $16,364 spend, $25,903 revenue, 1.58, 583 purchases, 278 active days. At the contract's exact $250 checkpoint it sat at 1.43 — two hundredths under the floor. This one the 1.45 rule does delete. The contract's other two clauses, over the same 43 revenue-producing ads: protected_winner (≥2.0 return at $100 spend) would ever have qualified 17 of 43 — so it protects under 40% of the ads worth protecting, which is why it needs promoting before anything else is switched on. zero_purchase (0 purchases by $54 spend) touches 2 of 43 — cheap and safe. Keep it. ## Evidence trail - mylibrary-kb/books/breakthrough-advertising — tonight's seed, section "Framework: The 5 Levels of Market Awareness" (lines 288–303) plus the Chapter 2 notes at lines 46, 57 and 408. Supplies the "psychological wall"/indifference framing and the headline's "one job… the second sentence" rule — why a correctly aimed ad reads as a zero early. Enriched 2026-02-26; a book note, no business figures taken from it. - founders-kb/episodes/082-david-ogilvy-ogilvy-advertising — lines 80–84. Lasker via Ogilvy, "No idea is big unless it will work for 30 years," and "Retiring a winning ad is a decision made for the agency's boredom, not the client's results." The long-horizon counterweight. - wescale-kb/calls/call-32 — 2026-06-04, Chris Heckman. The Bump and Cut Decision Tree, verbatim at lines 31 and 42: "spend to 2× CPA (~48 hours at $25/day)… under 2.0 = cut." - wescale-kb/members/eric-corry — line 164, the cohort's lone dissent on horizon: "Don't kill ads after 24 hours for bad ROAS." Never quantified — the gap this nugget fills. - wescale-kb/igl-action-plan — theme-aggregate, and dated: created 2026-02-16, seven months old, and a plan rather than a result. Action 3.3, lines 127–130: Jakob's "Kill anything under 1.8 ROAS after 2-3 days," with the IGL instruction "Set hard kill rule. No exceptions." This is the prescription the backtest contradicts. - igl-brand-kb/community/lauren-and-rowena-stories — lines 56–58 and 174: Tomato Lovers Collective member Jennifer Fox submitted "Life's too short to grow boring tomatoes," which the file calls one of IGL's best-selling designs. - (outside the knowledge bases) G:\iris\igl-ops-sage\data\igl.db — tables fact_meta_daily (ad level, 2024-08-01 to 2026-09-09) and ops_ad_rule_contracts. Every figure here that is not a document quotation comes from this database, re-derived independently by the operator on 2026-09-10. All returns are Meta-reported, not Shopify-verified. The knowledge bases hold no ad-level or product-level sales data at all, so none of these numbers could have come from them. ## Assay verdicts - novelty — PASS (5/5). Grepped ops_ad_rule_contracts, run_roas_floor, protected_winner, "rule contract" and "backtest" across all thirteen bases and the 35-nugget catalog. No prior art; the only backtest hits are unrelated. - grounding — PASS (5/5). The judge re-ran the backtest independently and it reproduced, including flagging the revenue figure the operator also corrected. - actionability — PASS (5/5). Stops a rule being switched on that would have deleted the ads the business runs on, and names the row, the two edits, and what to keep. Score: 15/15 — the first perfect score in the vault. ## Operator corrections applied to the hunt's card Every number was re-derived from the database. The finding held and got stronger; six figures did not. | Claim | Hunt said | Verified | Effect | |---|---|---|---| | Net cost of the 1.45 floor | −$19,768 | −$20,025 | headline corrected, loss is larger | | Revenue forfeited | $66,600 | $66,857 | corrected | | Winners under 1.8 at first $50 | 27 of 43, $171,610 | 26 of 43, $169,763 | corrected | | 43 winners' combined revenue | $236,488 | $236,960 | corrected | | protected_winner qualifiers | 16 of 43 | 17 of 43 | corrected | | The row to open | igl_ads_rule_contract_v1 | ads-rule-contract-step6-fixture-v1 | that string is the schema_version column, not the row id — as written the query returns nothing | | "Every floor loses at every checkpoint" | as stated | at $1,000 the 0.70 and 1.00 floors cut nothing, so they are neutral | narrowed to "no floor ever gains" | Exactly reproduced: 111 ads reaching $250, 57 cut, $46,832 avoided to the dollar, the 1.00/1.20/1.87 cut counts, median six days to $50, both named winners' lifetime figures and first-$50 returns, the Advantage+ 1.43, post-$250 return 1.73, account 1.50 all-time and 1.24 in 2026, zero_purchase touching 2 of 43, and every one of the six document quotations verbatim at the cited line numbers. Thread: breakthrough-advertising, 5 Levels of Market Awareness (Level 5 never names the product; the "psychological wall" means indifference, not a weak signal) → so a correctly aimed ad reads as a zero early → wescale call-32 Bump and Cut, "spend to 2× CPA, ~48h, under 2.0 = cut" → igl-action-plan Action 3.3, "kill under 1.8 after 2-3 days, no exceptions" → eric-corry's lone dissent, "don't kill after 24 hours" → founders ep-082, Lasker's "30 years" → igl-brand-kb lauren-and-rowena-stories, Jennifer Fox wrote "Life's too short to grow boring tomatoes" → operator: backtest every floor against all 111 ads that reached $250 in igl.db; no floor ever gains, and 1.45 loses $20,025.Re-render two design files this week and republish them onto the same Shopify listings — same handle, same reviews, no new product — using the image-seed re-render prompt IGL already wrote for exactly this ("Use the attached image (our own existing design) as the reference… Modernize, don't reinvent", competitors-kb/miners/tomato/render-prompts-top24-v2-image-seed-2026-07-09.md:249). File 1: "Excuse Me, Bird Feeder's Empty" — the squirrel's front paw is drawn as a human hand with fingernails; eight separate customers said so in the ad comments between 2024-11-15 and 2025-07-31, two of them refusing to buy over it. File 2: "Listen to Birds Not News" — the jay's back is brown instead of blue and the treble clefs and note stems are mirrored; four customers said so between 2025-10-27 and 2026-02-20, three of them phrased as blocked purchases.
Then add a sixth question to the Quality Check that already sits in Seedwork's Phase 4 (Prompts -> KITTL/NanoBanana Pro -> Quality Check -> MyDesigns -> Shopify, igl-brand-kb/core/strategic-synthesis-v4.md:369). Do not build a new step and do not fold this into phrase scoring — Seedwork's auto-scorer grades the words, and all five Quality Check questions that exist today grade meaning: voice, thesis, wearer, collection, community (igl-brand-kb/evolution/brand-thesis-v3-synthesis.md lines 477, 487, 497, 507, 517). Question six is the first one about the picture: "Open a reference photo. Is every depicted fact right — species coloration, animal anatomy, plant identity, musical notation?" Answer it against an actual reference image, not from memory.
Last, change the ad-comment reply rule: when a comment reports a drawing error, the reply names the fix and a date. Today every one of the thirteen got a joke — IGL answered "Paint that jay's back blue instead of brown" with "If you ever want to moonlight as our Head of Bird Accuracy, the position may have just opened," and the same ad kept running with the brown jay.
IGL's own Meta ad comments hold thirteen specific, unfixed factual drawing errors across three designs — including the shirt that is 36% of all company units — and every one of them was answered with wit instead of a correction.
The seed was a recall-kb capture whose whole argument is "stop selling the tool, solve the customer's problem," and whose most concrete tactic is a critique agent: "I have separate agents that all they do is check the work of my other agents… it writes a list of things to improve, sends it back to the coder." It also warns that "amateurs with a hammer make more problems, create more holes, waste more resources than anything else." IGL runs the hammer at roughly 24 designs a week through Seedwork, and nothing in that pipeline ever asks whether the picture is factually right.
What makes it non-obvious is where the evidence was hiding. The 3,210-comment ad-comment export has been indexed for a year and never read as a defect log. Read as one, the complaints are not scattered — they cluster on the designs that carry the business. The flagship squirrel has been publicly described as having "a human hand with curled fingernails" since November 2024, and that single anatomy error is also the only reason IGL was ever accused of using AI: "Hahaha! Your squirrel has a human hand! AI generated? Lol!"
No single shelf could produce this. The comment file is a raw artifact outside the knowledge bases. The persona shelf added "Would Ruth fact-check it and would it SURVIVE?" as a kill filter (igl-brand-kb/audience/buyer-portraits.md:429) on 2026-08-29 — a forward-looking rule for designs not yet made, written nine months after these two had already failed it in public, and never run backwards against the shipped catalog. And the mastermind supplied the stakes in one line from Chris Hall on call #16: the fear of a design error found "after you've sold 100 and someone points it out." IGL's flagship is 36% of all company units and has been pointed at thirteen times.
recall-kb/Business/Entrepreneurship/If you're trying to get rich with AI, you need to hear this….md — the seed. Supplies the critique-agent mechanism and the "amateurs with a hammer" warning. No IGL numbers taken from it.G:/iris/igl-ads/data/ad-comments-2026-08-04.json — the ad-comment export, pulled 2026-08-04, 3,210 comments across 1,722 ad stories, 2,224 unique by message-plus-timestamp. Outside the knowledge bases; indexed at igl-brand-kb/audience/who-data-map. Every quote above re-read from this file by the operator.igl-brand-kb/audience/profiles/05-carol-garden-witness — Status: CURRENT SoR, created 2026-08-29. Source of the $176k / 49.5% / 36%-of-company-units figures, quoted as document claims.igl-brand-kb/core/strategic-synthesis-v4 :369 — Seedwork Phase 4 pipeline; the Quality Check step the move attaches to.igl-brand-kb/evolution/brand-thesis-v3-synthesis — all five existing Quality Check questions, none about depicted fact.igl-brand-kb/audience/buyer-portraits :429 — the "Would Ruth fact-check it and would it SURVIVE?" filter, written 2026-08-29, forward-looking only.wescale-kb/raw-data/chats/call-16.md:1079-1097 — Shona Todd: "Just about sent out a design with a typo the other day"; Chris Hall: "Or after you've sold 100 and someone points it out"; bethwaller: "i did the typo."wescale-kb/raw-data/chats/call-20.md:225 — Alex Chao catches Jo Conlon's live typo after publish: "Don't know if you're aware, but I think there's a typo. It would be ニャンポップ."igl-brand-kb/community/customer-reviews-q4-2025 — Loox export 2025-12-11, 551 reviews; the two designs rank 1 (215 reviews) and 10 (6 reviews).competitors-kb/miners/tomato/render-prompts-top24-v2-image-seed-2026-07-09.md:249 — the existing re-render-from-our-own-image prompt the move reuses.competitors-kb/miners/tomato/render-prompts-top24-v2-image-seed-2026-07-09.md:249). File 1: "Excuse Me, Bird Feeder's Empty" — the squirrel's front paw is drawn as a human hand with fingernails; eight separate customers said so in the ad comments between 2024-11-15 and 2025-07-31, two of them refusing to buy over it. File 2: "Listen to Birds Not News" — the jay's back is brown instead of blue and the treble clefs and note stems are mirrored; four customers said so between 2025-10-27 and 2026-02-20, three of them phrased as blocked purchases. Then add a sixth question to the Quality Check that already sits in Seedwork's Phase 4 (Prompts -> KITTL/NanoBanana Pro -> Quality Check -> MyDesigns -> Shopify, igl-brand-kb/core/strategic-synthesis-v4.md:369). Do not build a new step and do not fold this into phrase scoring — Seedwork's auto-scorer grades the words, and all five Quality Check questions that exist today grade meaning: voice, thesis, wearer, collection, community (igl-brand-kb/evolution/brand-thesis-v3-synthesis.md lines 477, 487, 497, 507, 517). Question six is the first one about the picture: "Open a reference photo. Is every depicted fact right — species coloration, animal anatomy, plant identity, musical notation?" Answer it against an actual reference image, not from memory. Last, change the ad-comment reply rule: when a comment reports a drawing error, the reply names the fix and a date. Today every one of the thirteen got a joke — IGL answered "Paint that jay's back blue instead of brown" with "If you ever want to moonlight as our Head of Bird Accuracy, the position may have just opened," and the same ad kept running with the brown jay. ## Why (the one-liner) IGL's own Meta ad comments hold thirteen specific, unfixed factual drawing errors across three designs — including the shirt that is 36% of all company units — and every one of them was answered with wit instead of a correction. ## The insight The seed was a recall-kb capture whose whole argument is "stop selling the tool, solve the customer's problem," and whose most concrete tactic is a critique agent: "I have separate agents that all they do is check the work of my other agents… it writes a list of things to improve, sends it back to the coder." It also warns that "amateurs with a hammer make more problems, create more holes, waste more resources than anything else." IGL runs the hammer at roughly 24 designs a week through Seedwork, and nothing in that pipeline ever asks whether the picture is factually right. What makes it non-obvious is where the evidence was hiding. The 3,210-comment ad-comment export has been indexed for a year and never read as a defect log. Read as one, the complaints are not scattered — they cluster on the designs that carry the business. The flagship squirrel has been publicly described as having "a human hand with curled fingernails" since November 2024, and that single anatomy error is also the only reason IGL was ever accused of using AI: "Hahaha! Your squirrel has a human hand! AI generated? Lol!" No single shelf could produce this. The comment file is a raw artifact outside the knowledge bases. The persona shelf added "Would Ruth fact-check it and would it SURVIVE?" as a kill filter (igl-brand-kb/audience/buyer-portraits.md:429) on 2026-08-29 — a forward-looking rule for designs not yet made, written nine months after these two had already failed it in public, and never run backwards against the shipped catalog. And the mastermind supplied the stakes in one line from Chris Hall on call #16: the fear of a design error found "after you've sold 100 and someone points it out." IGL's flagship is 36% of all company units and has been pointed at thirteen times. ## Operator corrections to the hunt The hunt's population claim was wrong in the direction that strengthens the finding, and correcting it changed the move. 1. Not 7 errors on two designs — 13 on three. Re-enumerating all 2,224 unique comments (deduplicated by message plus timestamp; 3,210 raw) finds eight reports of the squirrel's human hand, not four. The hunt missed 2025-01-06 "Squirrel boy has man hands…….", 2025-02-25 "Creepy hand…lol", 2025-04-15 "That hand is so creepy", 2025-06-12 "Do squirrels 'hands' look like that?", and misread 2025-12-01 "but that's not a bird doesn't make sense" as an art defect when it is a reader missing the joke. Two of the missed four are explicit refusals to buy. It also missed a third affected design: Native Flowers Your Resistance, 2026-05-22, "Not native to Europe. Asclepias is a genus…" — a botanical accuracy complaint. Excluded on merit as preferences rather than errors: "It should be a rat", "the crop should be peppers", "instead of tomatoes it should be zucchini", "Aesthetically it should be centre justified". 2. The move was a new agent; it should be an existing step. Seedwork's Phase 4 already contains a Quality Check between rendering and publishing. Defining what that step must check is cheaper than building a parallel critique agent and hangs the change on a mechanism Tim already runs, per the standing rule. 3. Scoping the worst quote honestly. The 2025-07-31 "Do not buy this T-shirt" comment continues into an unrelated allegation about a different shirt entirely; only its first two sentences are the art complaint, and only those are cited here. A separate 2025-08-16 comment, "SCAM. Never received the merchandise," is a fulfilment complaint and is not counted among the thirteen. 4. Date and status. The ad-comment export was pulled 2026-08-04 and cannot show anything after that date. The $176k / 49.5% / "one SKU = 36% of company units" figures are quotations from 05-carol-garden-witness.md (Status: CURRENT SoR, created 2026-08-29, 11 days old), not measurements — the thirteen knowledge bases hold no product-level sales data. ## Worked examples — all thirteen, enumerated "Excuse Me, Bird Feeder's Empty" — the paw drawn as a human hand (8) | Date | Comment | |---|---| | 2024-11-15 | "Hahaha! Your squirrel has a human hand! AI generated? Lol!" (1 like) | | 2025-01-05 | "That critter has a human hand !!!Darwin would be amazed…." | | 2025-01-06 | "Squirrel boy has man hands……." | | 2025-02-25 | "Creepy hand…lol" | | 2025-03-26 | "I LOVE squirrels But…that 'paw' is creepy AF. That's a no from me" | | 2025-04-15 | "That hand is so creepy" | | 2025-06-12 | "Do squirrels 'hands' look like that?" | | 2025-07-31 | "Do not buy this T-shirt. Look closely at the 'Paw' and see that it is a human hand with curled fingernails." | "Listen to Birds Not News" — the jay's colour and the mirrored notation (4) | Date | Comment | IGL's reply | |---|---|---| | 2025-10-27 | "Get a real birder to design the shirt." (1 like) | "A 'real birder'? If they bring snacks and can identify a wren mid-flight, we're hiring." | | 2025-11-14 | "Paint that jay's back blue instead of brown, and flocks of Canadians would buy it!" | "One color correction and suddenly we've unlocked the Canadian market. If you ever want to moonlight as our Head of Bird Accuracy, the position may have just opened." | | 2026-02-20 | "If the treble clefs were the right way I would buy it!" | joke reply, 2026-03-13 | | 2026-02-20 | "And the notes too!" | — | "Native Flowers Your Resistance" — botanical accuracy (1) | Date | Comment | |---|---| | 2026-05-22 | "Not native to Europe. Asclepias is a genus of herbaceous, perennial, flowering plants known as milkweeds…" | ## Evidence trail - recall-kb/Business/Entrepreneurship/If you're trying to get rich with AI, you need to hear this….md — the seed. Supplies the critique-agent mechanism and the "amateurs with a hammer" warning. No IGL numbers taken from it. - G:/iris/igl-ads/data/ad-comments-2026-08-04.json — the ad-comment export, pulled 2026-08-04, 3,210 comments across 1,722 ad stories, 2,224 unique by message-plus-timestamp. Outside the knowledge bases; indexed at igl-brand-kb/audience/who-data-map. Every quote above re-read from this file by the operator. - igl-brand-kb/audience/profiles/05-carol-garden-witness — Status: CURRENT SoR, created 2026-08-29. Source of the $176k / 49.5% / 36%-of-company-units figures, quoted as document claims. - igl-brand-kb/core/strategic-synthesis-v4 :369 — Seedwork Phase 4 pipeline; the Quality Check step the move attaches to. - igl-brand-kb/evolution/brand-thesis-v3-synthesis — all five existing Quality Check questions, none about depicted fact. - igl-brand-kb/audience/buyer-portraits :429 — the "Would Ruth fact-check it and would it SURVIVE?" filter, written 2026-08-29, forward-looking only. - wescale-kb/raw-data/chats/call-16.md:1079-1097 — Shona Todd: "Just about sent out a design with a typo the other day"; Chris Hall: "Or after you've sold 100 and someone points it out"; bethwaller: "i did the typo." - wescale-kb/raw-data/chats/call-20.md:225 — Alex Chao catches Jo Conlon's live typo after publish: "Don't know if you're aware, but I think there's a typo. It would be ニャンポップ." - igl-brand-kb/community/customer-reviews-q4-2025 — Loox export 2025-12-11, 551 reviews; the two designs rank 1 (215 reviews) and 10 (6 reviews). - competitors-kb/miners/tomato/render-prompts-top24-v2-image-seed-2026-07-09.md:249 — the existing re-render-from-our-own-image prompt the move reuses. ## Assay verdicts - Novelty — PASS, 4/5. No prior art in any knowledge base or in the vault for the squirrel's hand, the jay's colour, or mirrored notation. The 3,210-comment file has never been read as a defect log in any catalogued nugget. Held at 4 because the flagship squirrel is the most-mined object in this vault (#010, #012, #014, #027, #029). - Grounding — PASS, 5/5. Operator re-opened the comment file and reproduced the denominator exactly (2,224 unique by message-plus-timestamp) and every quote verbatim with its date, then found six more the hunt had not. - Consequence — PASS, 5/5. Two named files, an existing pipeline step, and a reply rule; startable this week with no staff, inventory, or capital. Thread: recall-kb seed (critique agent, amateurs with a hammer) -> the ad-comment export read as a defect log -> 13 factual errors across 3 designs, two of them the review leaders -> wescale-kb call-16 and call-20 (the cohort's named failure mode: the error found after publish) -> igl-brand-kb Seedwork Phase 4, where the Quality Check step exists and checks only meaningBefore you contact Magicianly about flow setup (Action 2.1 in wescale-kb/igl-action-plan.md, created 2026-02-16), add two rows to the Tuesday 3 PM CT scorecard note, copied by hand out of Klaviyo: campaign revenue and flow revenue for that week. The WeScale portal stopped carrying both at scorecard week 42 (2026-W29, mid-July) — 9 of its 21 metrics now have zero rows for every one of the 60 members still reporting, including flow_revenue and campaign_revenue — so if you buy flows now, the one number that could prove they worked is no longer collected anywhere. In the 28 selling weeks before that cutoff your flows earned $8,749.77 and your 63 campaigns earned $4,537.67, about $72 a send, with flows ahead in 17 of those 28 weeks; so spend the Wednesday PM email block already in your member profile on flow architecture first and campaign design second.
IGL's automated email flows quietly earned twice what its hand-built campaigns did, and the scorecard metric that proves it went dark five weeks before any flow investment would be made.
Sam Altman's answer to "do you have to build your own products" is a platform rule: "Actually, I think we should be more of a platform company than a product company." One interface, one programming interface, best cost-performance, and let everything else compound on top. The email equivalent is exact — a flow is built once and runs on every visitor forever; a campaign is a product shipped on Wednesday and gone by Friday.
IGL's own recorded rows say the platform half is already winning without help. Across the 28 selling weeks from 2025-W43 to 2026-W29: $8,749.77 of flow revenue against $4,537.67 of campaign revenue from 63 sends, with flows ahead in 17 of the 28. The comparison store in the same niche, Jo Conlon of Happy Plant People, sends roughly one email per week, consistently on weekends, while IGL sends two or three. More sends is not the lever.
The half no plan in the workspace carries is the instrument. Altman's safety doctrine is the aviation one: "extremely robust accident reporting… They never try to, like, you know, hand wave over something. They want to extract as much information as possible." You cannot run that loop on a number nobody writes down. The portal importer that took over at week 43 carries 12 metrics; the nine it dropped include exactly the two that split email revenue into flows and campaigns. The advice layer says hire someone to build flows and expects email to move from 11.8% to 18–20% of revenue — but the instrument that would read that movement was switched off in mid-July, cohort-wide, for all 60 reporting members.
Honest limits. Flow and campaign revenue are not cleanly separable levers — flows ride on traffic, campaigns on list size — so the 2-to-1 ratio argues about where an hour of Tim's time goes, not about which channel is better. Every figure here is a quotation from self-reported scorecard rows, not from a sales system.
- The three selling weeks where IGL sold with zero campaigns sent: week 16 ($687.98), week 21 ($2,019.41), week 34 ($4,302.55). Week 34 is the week Chris reviewed at call #31. - The spend comparison over the same 28 weeks: $54,698.80 on Meta ads against $94,789.84 in sales (57.7%), versus $0 marginal cost for the flows that produced $8,749.77. - The cutoff is a schema change, not a gap: week 42 = 2026-W29 (V3B); week 43 = 2026-W30 (portal-v1).
founders-kb/interviews/035-sam-altman-transcript — raw verbatim transcript. Supplies the platform-over-product rule (line 92) and the aviation accident-reporting doctrine (line 72). Correction to the hunt's claim: these were not dropped from the enriched analysis half — both appear there (035-sam-altman.md lines 83 and 131). The founder material is the framing, not the finding; the finding is the dark metrics, and that is new.wescale-kb/scorecard/wescale-scorecard.db — structured artifact, self-reported rows, live through week 48. Every figure in this card was reproduced from it by the operator.igl-brand-kb/competitive/happy-plant-people-email-analysis — dated 2026-03-17, built from 25 emails captured 2025-10-27 to 2026-03-16. Jo Conlon's ~1-per-week weekend cadence.wescale-kb/members/tim-dobyns — member profile. Defines the "Email Optimization (Wednesday PM Block)" the move hangs on and the Tuesday 3 PM CT submission routine.wescale-kb/igl-action-plan (theme-aggregate) — plan created 2026-02-16. Action 2.1 "Hire Magicianly for Flow Setup," expected to move email from 11.8% to 18–20% — the prescription whose success metric is no longer recorded. Date-and-status note: the two comparison figures ($8,749.77 / $4,537.67) cover 2025-W43 through 2026-W29 and are self-reported scorecard rows, not measured sales-system data. The action plan is a proposal from 2026-02-16, not a live commitment. The dark-metric finding is current: verified against database weeks 43–48.wescale-kb/igl-action-plan.md, created 2026-02-16), add two rows to the Tuesday 3 PM CT scorecard note, copied by hand out of Klaviyo: campaign revenue and flow revenue for that week. The WeScale portal stopped carrying both at scorecard week 42 (2026-W29, mid-July) — 9 of its 21 metrics now have zero rows for every one of the 60 members still reporting, including flow_revenue and campaign_revenue — so if you buy flows now, the one number that could prove they worked is no longer collected anywhere. In the 28 selling weeks before that cutoff your flows earned $8,749.77 and your 63 campaigns earned $4,537.67, about $72 a send, with flows ahead in 17 of those 28 weeks; so spend the Wednesday PM email block already in your member profile on flow architecture first and campaign design second. ## Why (the one-liner) IGL's automated email flows quietly earned twice what its hand-built campaigns did, and the scorecard metric that proves it went dark five weeks before any flow investment would be made. ## The insight Sam Altman's answer to "do you have to build your own products" is a platform rule: "Actually, I think we should be more of a platform company than a product company." One interface, one programming interface, best cost-performance, and let everything else compound on top. The email equivalent is exact — a flow is built once and runs on every visitor forever; a campaign is a product shipped on Wednesday and gone by Friday. IGL's own recorded rows say the platform half is already winning without help. Across the 28 selling weeks from 2025-W43 to 2026-W29: $8,749.77 of flow revenue against $4,537.67 of campaign revenue from 63 sends, with flows ahead in 17 of the 28. The comparison store in the same niche, Jo Conlon of Happy Plant People, sends roughly one email per week, consistently on weekends, while IGL sends two or three. More sends is not the lever. The half no plan in the workspace carries is the instrument. Altman's safety doctrine is the aviation one: "extremely robust accident reporting… They never try to, like, you know, hand wave over something. They want to extract as much information as possible." You cannot run that loop on a number nobody writes down. The portal importer that took over at week 43 carries 12 metrics; the nine it dropped include exactly the two that split email revenue into flows and campaigns. The advice layer says hire someone to build flows and expects email to move from 11.8% to 18–20% of revenue — but the instrument that would read that movement was switched off in mid-July, cohort-wide, for all 60 reporting members. Honest limits. Flow and campaign revenue are not cleanly separable levers — flows ride on traffic, campaigns on list size — so the 2-to-1 ratio argues about where an hour of Tim's time goes, not about which channel is better. Every figure here is a quotation from self-reported scorecard rows, not from a sales system. ## The full result set — all 21 scorecard metrics, checked Verified by the operator directly against wescale-kb/scorecard/wescale-scorecard.db, not sampled. 9 of 21 metrics are dark (zero observations at week 43 or later, across all 60 reporting members); 12 survive. | Dark after week 42 (9) | Still carried (12) | |---|---| | campaign_revenue, flow_revenue, email_total, popup_rate, cogs_pct, overhead_pct, google_ad_spend, google_cpc, google_roas | total_sales, fb_ad_spend, fb_roas, fb_cpc, aov, conversion_rate, net_profit_margin, campaigns_sent, new_designs, new_ad_creatives, email_pct_revenue, mer | mer is technically alive but nearly so: it has post-cutoff rows for 3 members out of 60. ## Worked examples - The three selling weeks where IGL sold with zero campaigns sent: week 16 ($687.98), week 21 ($2,019.41), week 34 ($4,302.55). Week 34 is the week Chris reviewed at call #31. - The spend comparison over the same 28 weeks: $54,698.80 on Meta ads against $94,789.84 in sales (57.7%), versus $0 marginal cost for the flows that produced $8,749.77. - The cutoff is a schema change, not a gap: week 42 = 2026-W29 (V3B); week 43 = 2026-W30 (portal-v1). ## Evidence trail - founders-kb/interviews/035-sam-altman-transcript — raw verbatim transcript. Supplies the platform-over-product rule (line 92) and the aviation accident-reporting doctrine (line 72). Correction to the hunt's claim: these were not dropped from the enriched analysis half — both appear there (035-sam-altman.md lines 83 and 131). The founder material is the framing, not the finding; the finding is the dark metrics, and that is new. - wescale-kb/scorecard/wescale-scorecard.db — structured artifact, self-reported rows, live through week 48. Every figure in this card was reproduced from it by the operator. - igl-brand-kb/competitive/happy-plant-people-email-analysis — dated 2026-03-17, built from 25 emails captured 2025-10-27 to 2026-03-16. Jo Conlon's ~1-per-week weekend cadence. - wescale-kb/members/tim-dobyns — member profile. Defines the "Email Optimization (Wednesday PM Block)" the move hangs on and the Tuesday 3 PM CT submission routine. - wescale-kb/igl-action-plan (theme-aggregate) — plan created 2026-02-16. Action 2.1 "Hire Magicianly for Flow Setup," expected to move email from 11.8% to 18–20% — the prescription whose success metric is no longer recorded. Date-and-status note: the two comparison figures ($8,749.77 / $4,537.67) cover 2025-W43 through 2026-W29 and are self-reported scorecard rows, not measured sales-system data. The action plan is a proposal from 2026-02-16, not a live commitment. The dark-metric finding is current: verified against database weeks 43–48. ## Assay verdicts - Novelty — PASS (3/5). The flows-first half is already prescribed (igl-action-plan.md Action 2.1, sal-presti.md, kevin-godinez.md, frameworks/weekly-review-methodology.md line 321). What survives is the un-prescribed constraint on that assigned build: the two metrics that would prove the spend worked have had zero observations cohort-wide since mid-July. Nugget #032 touched the same importer but only the percent-unit defect in metrics it does carry; the nine it silently drops appear nowhere in the catalog. Docked to 3 because the headline leads with the known half. - Grounding — PASS (5/5). Every figure reproduces to the cent. Independently re-verified by the operator: flow $8,749.77, campaign $4,537.67, 63 campaigns, 28 selling weeks, flows ahead in 17, 9 of 21 metrics dark, week 42 = 2026-W29, Meta spend $54,698.80, sales $94,789.84. - Consequence — PASS (4/5). Changes a specific decision: what Tim records this Tuesday, and what he buys before he can measure it. Two rows in a note he already writes; no staff, capital, or inventory required. Thread: 035-sam-altman-transcript (platform over product; accident-reporting doctrine) → wescale-scorecard.db (flows $8,750 vs campaigns $4,538 over 28 weeks; 9 of 21 metrics dark after week 42) → igl-action-plan Action 2.1 (hire Magicianly for flows) → happy-plant-people-email-analysis (Jo Conlon: one consistent send a week) → tim-dobyns.md Wednesday PM email blockAdd one guard clause, in Chris Heckman's own words, to the Discover step of competitors-kb/miner-methodology.md and to any Claude run that asks "what do my winners have in common": "Do not give me feedback until you have verified that you are actually able to see the ad creative. Inferring from the naming convention alone is not enough." Operationally: the run must load the actual image files from competitors-kb/proven-winners/<theme>/cropped/ (they already exist, cropped to just the printed artwork) and print three visual attributes per design — line weight, palette, text-vs-graphic ratio — before it is allowed to state a pattern. Never let it reason from the title field in manifest.json, from Shopify handles, or from your Klaviyo naming convention. Run it first on the 57 files in proven-winners/garden/plant-humor/cropped/, not on faith-gardening — that is where the title pass scored zero, and it is the lane your own best-sellers live in.
Your competitor corpus was labeled twice, once from product titles and once by looking at the pictures, and both labels were kept — so it contains a measured proof that titles cannot see humor, which is exactly the failure Chris wrote a guard clause against on call #32.
Call #32's transcript addendum contains a prompt Chris dictated live for analysing six months of ad creative. Buried in it is a rule that reads like paranoia: don't give me feedback until you have verified you can actually see the creative; inferring from the naming convention is not enough. It sounds like prompt-engineering hygiene.
It is not, and IGL has the receipt. The proven-winners corpus was built twice — once classifying each design from its title, once by a vision pass over the image — and both labels survive in manifest.json. Of 620 designs, 333 (54%) got no usable label from the title at all, and 89 of the 287 that did were overturned. Title-only was silent or wrong on 422 of 620, 68%.
But the aggregate hides the real finding, which is that the error is directional, not random. Sorted by theme, the failure is not spread evenly: it concentrates almost perfectly in humor. Of the 57 designs the corpus finally classifies as plant-humor, the title pass identified zero. Not a low number — zero. It filed 26 of them as "general-garden" and 13 as "veg-homestead," and the 4 designs it did call plant-humor were all something else. The reason is plain once seen: a product title says what a shirt is about (gardening, vegetables, plants), while the joke lives in the picture and the phrasing. "Maybe Plants Are Addicted to Me" reads as general-garden to a title classifier. So does "Most Likely To Talk To Plants."
That matters more for IGL than for any other shop in the corpus, because humor is IGL's proven winning mode. The four best-selling shirts of the last 90 days — Listen to the Birds, Quietly Cheering for Moss, Life's Too Short Tomatoes, Excuse Me Bird Feeder's Empty — are all wry garden humor. The single category the mining pipeline is blindest to when it reasons from titles is the exact category IGL wins in.
Why no single shelf could produce this: the mastermind call carries the rule but no proof; the corpus carries the proof but never states the rule; and the Wright Brothers episode carries the reason the rule matters at all. Lilienthal's flight tables were trusted by the entire field and were, in Orville's word, worthless, so the brothers built a wind tunnel and measured for themselves. A product title is Lilienthal's table: an authoritative-looking label somebody else wrote, standing in for a measurement nobody took. Dave Shaffer's Claude run that surfaced woodcut/linocut across his top 8 designs (call #27) is the counter-case — that answer is only reachable by looking at pixels, since no product title in his store contains the word "woodcut."
wescale-kb/calls/call-32 — SEED. Transcript addendum, Segment 3: Chris Heckman's live-dictated creative-analysis prompt, carrying the verbatim guard clause at line 229. Raw, unenriched call transcript.competitors-kb/proven-winners/manifest.json — the measured proof. 620 records, each carrying theme_title (title-derived), theme_vision (image-derived) and theme_final. Corpus built 2026-07-12. This is a classification artifact, not sales data — the knowledge bases hold no product-level sales figures.founders-kb/episodes/028-the-wright-brothers — the principle underneath the rule: "Clearly, those esteemed authorities had been guessing, groping in the dark. The accepted tables were, in a word, worthless. We had to go ahead and discover everything ourselves" (Orville Wright, line 93). They built a wind tunnel rather than trust Lilienthal's published tables.wescale-kb/calls/call-27 — the counter-case: Dave Shaffer had Claude analyse his top 8 designs and it surfaced woodcut/linocut engraving as the shared trait, a finding only reachable from the images.competitors-kb/proven-winners/README — theme aggregate. Documents the two-pass method that makes the disagreement measurable, and confirms every theme folder has a cropped/ subfolder of artwork-only images ready to feed the analysis.competitors-kb/miner-methodology.md and to any Claude run that asks "what do my winners have in common": "Do not give me feedback until you have verified that you are actually able to see the ad creative. Inferring from the naming convention alone is not enough." Operationally: the run must load the actual image files from competitors-kb/proven-winners/<theme>/cropped/ (they already exist, cropped to just the printed artwork) and print three visual attributes per design — line weight, palette, text-vs-graphic ratio — before it is allowed to state a pattern. Never let it reason from the title field in manifest.json, from Shopify handles, or from your Klaviyo naming convention. Run it first on the 57 files in proven-winners/garden/plant-humor/cropped/, not on faith-gardening — that is where the title pass scored zero, and it is the lane your own best-sellers live in. ## Why (the one-liner) Your competitor corpus was labeled twice, once from product titles and once by looking at the pictures, and both labels were kept — so it contains a measured proof that titles cannot see humor, which is exactly the failure Chris wrote a guard clause against on call #32. ## The insight Call #32's transcript addendum contains a prompt Chris dictated live for analysing six months of ad creative. Buried in it is a rule that reads like paranoia: don't give me feedback until you have verified you can actually see the creative; inferring from the naming convention is not enough. It sounds like prompt-engineering hygiene. It is not, and IGL has the receipt. The proven-winners corpus was built twice — once classifying each design from its title, once by a vision pass over the image — and both labels survive in manifest.json. Of 620 designs, 333 (54%) got no usable label from the title at all, and 89 of the 287 that did were overturned. Title-only was silent or wrong on 422 of 620, 68%. But the aggregate hides the real finding, which is that the error is directional, not random. Sorted by theme, the failure is not spread evenly: it concentrates almost perfectly in humor. Of the 57 designs the corpus finally classifies as plant-humor, the title pass identified zero. Not a low number — zero. It filed 26 of them as "general-garden" and 13 as "veg-homestead," and the 4 designs it did call plant-humor were all something else. The reason is plain once seen: a product title says what a shirt is about (gardening, vegetables, plants), while the joke lives in the picture and the phrasing. "Maybe Plants Are Addicted to Me" reads as general-garden to a title classifier. So does "Most Likely To Talk To Plants." That matters more for IGL than for any other shop in the corpus, because humor is IGL's proven winning mode. The four best-selling shirts of the last 90 days — Listen to the Birds, Quietly Cheering for Moss, Life's Too Short Tomatoes, Excuse Me Bird Feeder's Empty — are all wry garden humor. The single category the mining pipeline is blindest to when it reasons from titles is the exact category IGL wins in. Why no single shelf could produce this: the mastermind call carries the rule but no proof; the corpus carries the proof but never states the rule; and the Wright Brothers episode carries the reason the rule matters at all. Lilienthal's flight tables were trusted by the entire field and were, in Orville's word, worthless, so the brothers built a wind tunnel and measured for themselves. A product title is Lilienthal's table: an authoritative-looking label somebody else wrote, standing in for a measurement nobody took. Dave Shaffer's Claude run that surfaced woodcut/linocut across his top 8 designs (call #27) is the counter-case — that answer is only reachable by looking at pixels, since no product title in his store contains the word "woodcut." ## Evidence trail - wescale-kb/calls/call-32 — SEED. Transcript addendum, Segment 3: Chris Heckman's live-dictated creative-analysis prompt, carrying the verbatim guard clause at line 229. Raw, unenriched call transcript. - competitors-kb/proven-winners/manifest.json — the measured proof. 620 records, each carrying theme_title (title-derived), theme_vision (image-derived) and theme_final. Corpus built 2026-07-12. This is a classification artifact, not sales data — the knowledge bases hold no product-level sales figures. - founders-kb/episodes/028-the-wright-brothers — the principle underneath the rule: "Clearly, those esteemed authorities had been guessing, groping in the dark. The accepted tables were, in a word, worthless. We had to go ahead and discover everything ourselves" (Orville Wright, line 93). They built a wind tunnel rather than trust Lilienthal's published tables. - wescale-kb/calls/call-27 — the counter-case: Dave Shaffer had Claude analyse his top 8 designs and it surfaced woodcut/linocut engraving as the shared trait, a finding only reachable from the images. - competitors-kb/proven-winners/README — theme aggregate. Documents the two-pass method that makes the disagreement measurable, and confirms every theme folder has a cropped/ subfolder of artwork-only images ready to feed the analysis. ## Worked examples — the FULL result set, all 17 theme folders Rate = share of that folder's designs the title pass got wrong or could not label at all. | Rate | n | blank | overturned | theme (folder) | |---|---|---|---|---| | 100.0% | 57 | 12 | 45 | plant-humor — start here | | 100.0% | 28 | 27 | 1 | horse-heritage | | 100.0% | 23 | 21 | 2 | misc-nongarden | | 100.0% | 2 | 2 | 0 | slow-living | | 88.9% | 18 | 15 | 1 | spiritual-empower | | 88.2% | 17 | 15 | 0 | outdoor-adventure | | 86.2% | 29 | 23 | 2 | faith-gardening | | 82.4% | 17 | 10 | 4 | moss-mushroom | | 77.8% | 36 | 25 | 3 | food-italian | | 74.6% | 71 | 44 | 9 | garden-wildlife | | 67.2% | 58 | 36 | 3 | nature-outdoor | | 61.5% | 26 | 16 | 0 | auto-vintage | | 53.3% | 60 | 25 | 7 | veg-homestead | | 46.9% | 32 | 13 | 2 | general-garden | | 42.9% | 77 | 27 | 6 | wildflowers-floral | | 41.9% | 43 | 15 | 3 | pollinators | | 30.8% | 26 | 7 | 1 | ocean-coastal | Individual designs the title pass misfiled, all verified in manifest.json (brand-safe examples only): - "Maybe Plants Are Addicted to Me" — title said general-garden. Actually plant-humor. - "Most Likely To Talk To Plants" — title said general-garden. Actually plant-humor. - "Oops, I Bought More Plants" — title said general-garden. Actually plant-humor. - "Botanically Unhinged" — title said wildflowers-floral. Actually plant-humor. - "Certified Overplanter" — title said general-garden. Actually plant-humor. - "WTF: What's That Flower?" — title said wildflowers-floral. Actually plant-humor. - "Floral Heart T-Shirt" (faithwiseco.com) — title said wildflowers-floral. Actually faith-gardening, IGL's own core niche. - "In The Garden T-Shirt" (gardenistee.com) — title said general-garden. Actually faith-gardening. - "Ladybug Flowers T-shirt" — title said pollinators. Actually garden-wildlife. ## Assay verdicts - Novelty — PASS (4/5). The guard clause appears only in call-32.md:229 and its two transcript copies, purely as prompt hygiene for ad review; nothing ports it to mining, and nothing anywhere states the corpus's own title-vs-vision disagreement. Vault checked: #005, #010 and #031 all interrogate the same manifest and none touch classification provenance. Operator re-check: title-vs-vision returns zero hits across all 13 bases and all 32 prior nuggets. - Grounding — PASS (5/5). Every number reproduced independently from disk by both the judge and the operator: 620 records, 333 blank, 89 overturned, 422/620 = 68.06%. All worked examples verify verbatim. The guard clause is verbatim at line 229; the Orville Wright quote verbatim at line 93 of episode 028. - Actionability — PASS (4/5). Changes a named, repeated decision — how the Discover step classifies designs. Specific guard-clause insertion into a named file plus a hand-run over a named existing folder. No staff, capital, or inventory. Not a 5 because the payoff is analysis quality, one step removed from revenue. Thread: call-32 transcript addendum (Chris's "verify you can see the creative") → call-27 (Dave's top-8 woodcut finding, image-only) → founders ep-028 Wright Brothers (Lilienthal's worthless tables) → competitors-kb proven-winners/manifest.json (counted title-vs-vision labels: 68% wrong or blank, and zero of 57 for humor)