Account memory & who's driving

Lessons are mined from this account's own concluded experiments in the knowledge corpus — they override generic best practice and are enforced on every recommendation above (price steps cite the proven-safe finding; URL/handle changes are hard-blocked).

⛔ GUARDRAIL NEVER change URL handles or consolidate listings on revenue products -- Merchant Center treats the result as a NEW product and resets its history (proven twice on this account: Leghorn 4,000+ clicks/28d -> 159; Orpington same pattern). [corpus.db concluded experiment]

⛔ GUARDRAIL NOTHING is pushed to ad platforms automatically. Every recommendation, creative batch, and audience test produces exact manual steps; a human executes and confirms with the done-tap. [Chris directive 2026-07-13]

⛔ GUARDRAIL NEVER generate product imagery from imagination -- each coop model looks a specific way. Ad images originate from real CCC media (Shopify product photos, Drive library, customer photos); AI may only enhance with overlays, text, graphics, backgrounds. [Chris directive 2026-07-13; fabricated Recraft coops quarantined]

✅ PROVEN SAFE Price edits on revenue products are proven SAFE on this account -- no Merchant Center history reset observed when Rhode Island / Bundle prices were raised. [corpus.db concluded experiment]

ℹ CONTEXT Variant additions to revenue products create new MC entries (per shopify_mc.md Rule 3) -- plan feed changes accordingly. [corpus.db concluded experiment]

ℹ CONTEXT Cascade effect. 20+ replacement parts entered MC as new individual products, diluting the feed and creating disapproval risk for each. Several of the current 246 MC disapprovals trace to this event. [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): no clear change -- store contribution +72/day before vs +128/day after (season-adjusted x1.21); change +56/day (95% range -152 to +264; 80% range -81 to +194) [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): too little data -- store contribution -25/day before vs +30/day after (season-adjusted x1.31); change +55/day (95% range -30 to +158; 80% range -6 to +122), a [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): no clear change -- store contribution +84/day before vs +86/day after (season-adjusted x1.36); change +1/day (95% range -166 to +185; 80% range -112 to +121), [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +323/day before vs +767/day after (season-adjusted x0.71); change +444/day (95% range -248 to +1,208; 80% range -9 t [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +438/day before vs +305/day after (season-adjusted x0.79); change -132/day (95% range -679 to +437; 80% range -510 t [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): no clear change -- store contribution +54/day before vs +49/day after (season-adjusted x0.66); change -5/day (95% range -190 to +227; 80% range -135 to +140), [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): no clear change -- store contribution +200/day before vs +280/day after (season-adjusted x0.63); change +81/day (95% range -288 to +512; 80% range -177 to +35 [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): no clear change -- store contribution +54/day before vs +55/day after (season-adjusted x0.61); change +1/day (95% range -193 to +248; 80% range -134 to +157), [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +278/day before vs +277/day after (season-adjusted x0.70); change -2/day (95% range -509 to +572; 80% range -349 to [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): LIKELY improved -- store contribution -17/day before vs +78/day after (season-adjusted x1.03); change +94/day (95% range -38 to +255; 80% range +1 to +192), a [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): no clear change -- store contribution -9/day before vs +49/day after (season-adjusted x1.21); change +58/day (95% range -62 to +200; 80% range -28 to +148), a [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): too little data -- store contribution -44/day before vs -2/day after (season-adjusted x0.72); change +41/day (95% range -8 to +115; 80% range +2 to +89), abou [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +416/day before vs +171/day after (season-adjusted x0.88); change -245/day (95% range -843 to +335; 80% range -632 t [corpus.db concluded experiment]

ℹ CONTEXT platform read (all bing campaigns, bing): no clear change -- store contribution +121/day before vs +58/day after (season-adjusted x0.78); change -63/day (95% range -326 to +238; 80% range -231 to +120 [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): no clear change -- store contribution +54/day before vs +126/day after (season-adjusted x0.72); change +71/day (95% range -132 to +302; 80% range -75 to +212) [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +447/day before vs +168/day after (season-adjusted x0.89); change -278/day (95% range -860 to +289; 80% range -668 t [corpus.db concluded experiment]

ℹ CONTEXT platform read (all meta campaigns, meta): no clear change -- store contribution +53/day before vs +153/day after (season-adjusted x0.81); change +100/day (95% range -103 to +322; 80% range -39 to +248 [corpus.db concluded experiment]

ℹ CONTEXT platform read (all google campaigns, google): no clear change -- store contribution +398/day before vs +207/day after (season-adjusted x0.97); change -191/day (95% range -821 to +427; 80% range -625 t [corpus.db concluded experiment]

Who is changing the accounts (10,335 recorded changes)

⚠ Unidentified actors made recent changes: Administrator, Analytics Brain, admin@chickencoopcompany.com, chris@chrisbowman.us — Chris/Danny: who are these? Access audit if unrecognized.

Notable recent changes (bid strategy / budgets)

• 2026-10-05 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(1) PMax - Shopping Only - Coops & Runs'

• 2026-10-05 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(1) PMax - Shopping Only - Low Priced Coops'

• 2026-09-22 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(1) PMax - Shopping Only - Low Priced Coops'

• 2026-09-15 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(NB) Chicken Coops'

• 2026-09-11 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(1) PMax - Shopping Only - Coops & Runs'

• 2026-09-11 — Other/unknown (chris@chrisbowman.us) changed maximize_conversion_value.target_roas on google '(B) Branded'

Change visibility by platform

facebook: full history via Meta change log (accumulated)

google: API trailing-30d only, accumulating since 2024-07-16 (1357 records); Jan-May 2026 needs the Google UI export (open item)

bing: not yet -- unlocks when the Bing API browser steps (docs/ops/bing_api_setup.md) are done

Takeover readiness — the work vs what Marcus covers

Every recorded change from the last year, classified into the kinds of work running these accounts actually consists of. This is the gap map: green is handled, amber is partial with a plan, red needs a decision before full takeover.

Change calendar — day by day, last 12 months

Every recorded change, grouped: the same actor doing the same thing on the same campaign in one day collapses to one line with a count. Months and days expand.