# Turkey 5 Playbook: Running Ads Live, Day by Day

URL: https://aiadking.com/blog/turkey-5-amazon-ads-playbook
Updated: 2026-09-18

The operational playbook for Nov 26-30: what each Turkey 5 day rewards, the 10-minute pacing check that is the event's real work, what you may touch live, an…

## Operating mode: what changes when the event starts
Normal-week PPC management is analytical: settled windows, weekly reviews, deliberate steps. Event-week management inverts one variable — budgets become an hourly concern — while everything else becomes more conservative, not less. The paradox trips sellers every year: surrounded by five days of maximum noise, they make maximum changes, then spend December unable to tell which change did what.
The operating rules, stated once:
Bids were decided before the event (at deal-price break-evens, in the lead-in). Live bid-chasing buys the same impressions at panic prices.
Budgets are the live control surface — headroom raises for capping winners, caps for bingeing losers. That's the whole whitelist, plus listing-side firefighting.
No structural changes, no new campaigns, no negation sprees mid-event: today's "waster" may simply have unlanded attribution (event shoppers convert late), and the thin-data rule applies double under surge noise.
Judgment waits: the event's real scorecard arrives in the two-pass harvest at T+2 and T+10.
## The five days, characterised
Each day's one-line strategy:
Thanksgiving: moderate presence; expect soft same-day conversion (carts build now, close later). Don't read the day's ACoS at all.
Black Friday: full headroom on Tier-A campaigns from the morning; the day's only unforgivable error is a winner dark at 7pm on a cap you could have raised at 1pm.
The Saturday-Sunday trough: the sleeper. Volume dips, but CPCs dip harder as less-prepared competitors' budgets exhaust — the cheapest event-condition clicks of the five days go to whoever kept meaningful presence. Cutting the weekend "because Friday is over" is the classic overreaction (the pacing guide's most-repeated warning).
Cyber Monday: the double pattern — desk-hours browsing, then the biggest late-evening close of the week. Category tilt toward electronics/gadgets, but the late peak is broad. Keep budget alive to the end; the last three hours matter.
## The 10-minute pacing check
The check runs at each day's scheduled times (more on peak days, fewer on the trough). Its design matters: four looks, two permitted actions. Look 1 (pacing) produces most raises; look 2 (binge detection) produces the rare cap; looks 3-4 route problems away from ads settings toward their real fixes — a vanished deal badge or a lost Buy Box is a listing emergency that no bid can compensate, and a deal burning stock at triple forecast needs its ads capped, counterintuitively, so the final units serve organic buyers at full margin instead of paid clicks at deal margin.
What the check deliberately excludes: ACoS reading (unsettled), search-term judgments (unsettled), bid changes (decided pre-event), placement tinkering (no settled data), and dashboard-refreshing between scheduled checks (anxiety, not management). Ten minutes, on schedule, then close the console — the event rewards operators, not spectators.
## The live-touch whitelist, precisely
Everything you may change while the event runs:
Raise a budget cap on a Tier-A campaign pacing to exhaust before the evening peak. The default event action; take it early — a raise at 1pm buys the whole evening, a raise at 8pm buys regret.
Cap or pause a proven binge: meaningful spend, zero orders, and a confirmed cause (dead badge, lost Buy Box, broken variation). Without the confirmed cause, prefer capping to pausing — pausing mid-event resets delivery pacing you can't rebuild until it's over.
Fix listing-side emergencies: repricer floors breached, suppressed variations, deal-status surprises — the quadrant-four rebase if a deal silently died.
Negate a genuinely alien term — the rare event-traffic surprise that's categorically irrelevant (your bottle ad serving on a costume query). Relevance judgments need no sample; performance judgments wait.
Everything else is off-list. Write the whitelist somewhere visible before Thursday: the mid-event urge to "do something" is real, and a pre-committed list is cheaper than discipline.
## Reading the event without fooling yourself
You'll watch numbers anyway, so calibrate what they can and can't say live:
Orders and revenue: directionally real. Units sold today mostly happened today. Fine for stock-burn and gross-sanity checks.
ACoS: systematically overstated all five days — spend registers immediately, attribution lands over the following days, and event shoppers convert later than normal (Thursday's carts close Monday). Friday-morning ACoS panic is an annual tradition with no survivors' benefits; the number that matters arrives with the settled window in December.
Traffic quality shifts: event pages are badge-saturated, so full-price products' CTRs sag against their normal baselines — expected, not a creative emergency.
Comparisons: event-to-event only. This Black Friday vs last Black Friday (or vs October's rehearsal), never vs last Tuesday. Keep last year's numbers open in a tab; they're the only honest yardstick in the room.
## The handoff: Monday midnight to December
The event's end is an operational moment of its own:
Don't cut budgets at midnight. Tuesday-Wednesday carry the immediate halo — shoppers returning to carts, late-attribution conversions still landing on your recent velocity. The December glide starts as a step-down over days, not a cliff at 12:01am.
T+2: the first harvest pass — obvious new winners graduated, categorically-alien terms negated, and the war-room log (every cap raised, every binge, every listing incident) written up while memory is fresh. That log is next year's prep document.
T+10: the second pass, once attribution has landed — the pass that catches Thursday's carts closing Monday, and un-convicts "wasters" that were actually slow closers.
Re-derive bids on post-event CVR where the event durably shifted conversion (it often does, in both directions), and restore deal-product bids to list-price break-evens as discounts expire.
Then the account returns to its normal cadence — reviews weekly, judgments settled, the report routine resuming — with December's cheaper, longer gift-tail to harvest.
## The war-room log: the event's most underrated artifact
Every pacing check produces observations; the sellers who improve year over year write them down. The log format that works is deliberately trivial — timestamp, observation, action (or "none"), one line each:
Fri 1:05pm — Hero SKC at 82% budget, day 54% elapsed, ACoS-live fine → cap +40%. · Fri 7:20pm — Accessory campaign $61 spend, 0 orders; checked: coupon badge missing since ~4pm → capped ads, ticketed the coupon. · Sat 10am — CPCs on category terms visibly down vs Friday → no action, noted for next year's trough budgeting. · Mon 10:15pm — Late peak real: 31% of day's orders after 8pm → keep next year's Monday budget alive to midnight.
Three reasons this beats memory:
The T+2 write-up writes itself — and the T+10 second pass can check whether Friday's "binge" was actually a slow closer, because you know exactly when and why you capped it.
Next year's prep document is this year's log. The October fix-window works through a breakage list; the log is that list, pre-written under battle conditions.
It converts anxiety into procedure. The urge to "do something" between checks gets a legitimate outlet: write the observation down, take no action, close the console. Half of event discipline is having somewhere to put the nervousness.
## Running the Turkey 5 without living in the console
Read back through the playbook and notice what the human is actually doing: threshold checks on pacing, binge detection, stock-burn math, and a short whitelist of mechanical responses — on a schedule, for five days, during the same week the rest of the business also peaks. It's the most automatable week of the year, and the week automation earns the most: AIAdKing runs the pacing checks continuously instead of thrice daily, raises headroom before winners cap, flags binges with their probable cause (badge, Buy Box, stock) instead of just their symptom, holds the whitelist discipline structurally — and logs every event-week action with its reason for the T+2 write-up you'd otherwise reconstruct from memory. Shadow mode lets you keep final approval during the highest-stakes week; the flat fee doesn't care that it's Black Friday.

## FAQ

Q: What is the Turkey 5 in Amazon advertising?
A: The five-day stretch from Thanksgiving through Cyber Monday — November 26-30 in 2026 — Amazon's highest-traffic week. Each day behaves differently: Thanksgiving builds carts with an evening browse surge, Black Friday runs twin peaks all day, Saturday-Sunday form a cheaper-CPC value trough as competitors' budgets exhaust, and Cyber Monday doubles work-hours browsing with the week's biggest late-evening close.

Q: What should I change on my Amazon ads during Black Friday itself?
A: Almost nothing except budgets: raise caps on winning campaigns before they go dark (the day's only unforgivable error is a winner capped at 1pm missing the evening peak), cap a campaign bingeing spend with a confirmed cause like a dead deal badge or lost Buy Box, and fix listing-side emergencies. Bids stay where the pre-event calibration put them — live bid-chasing buys the same impressions at panic prices.

Q: Should I lower my Amazon ads on the Saturday after Black Friday?
A: Usually the opposite. Volume dips but CPCs dip harder as under-prepared competitors exhaust their budgets, making the Saturday-Sunday trough the cheapest event-condition clicks of the five days for sellers who keep meaningful presence. Cutting the weekend because "Black Friday is over" is the most common Turkey 5 overreaction.

Q: Why does my ACoS look terrible during the Turkey 5?
A: Because spend registers immediately while attribution lands over the following days — and event shoppers convert later than normal, with Thursday's carts closing as late as Cyber Monday. Event-week ACoS is systematically overstated on every live reading; the real number arrives with settled data in early December. Judge the event against last year's event, never against a normal week.

Q: What should I do right after Cyber Monday ends?
A: Step budgets down gradually rather than cutting at midnight (the immediate halo converts through midweek), run a first search-term harvest at T+2 and a second at T+10 once attribution lands, write up the event log while memory is fresh, re-derive bids where the event durably shifted conversion rates, and restore deal products to list-price break-evens as discounts expire.