Preparing Google Ads for Peak Season: Budgets, Bidding, and Ramp-Up

Peak season punishes improvisation. By the time Black Friday week arrives, the auctions are at their most expensive, the platforms are at their busiest, and every structural weakness in an account (thin conversion data, capped budgets, a messy feed, an untested promotion plan) gets amplified by volume. The accounts that win November made their important decisions in September and October. The accounts that struggle made theirs on the Tuesday before the sale.

This is the preparation playbook we run for ecommerce and promotion-driven accounts in the roughly six weeks before peak, in sequence.

Six Weeks Out: Structure and Signal

Smart Bidding will carry your peak performance, and it performs exactly as well as the data reaching it. So the earliest work is signal work: verify conversion tracking end to end, fix any double counting or gaps (the same checks we run in a GA4 audit), and make sure conversion values are accurate, especially if you have moved to the value-based bidding setup that peak season rewards most, since order values swing hardest exactly when margin discipline matters most.

Structural changes belong now, not later. Campaign restructures, bidding strategy migrations and major feed title rewrites all trigger learning periods, and a learning period during peak week is money on fire. The rule we hold accounts to: no structural changes inside the final three weeks. Clean the feed now too, because Merchant Center disapprovals spike when promotional prices and availability start changing, and each disapproved product is an auction you silently left. Set a weekly diagnostics check from here through January.

Four Weeks Out: Budgets, Forecasts and the Promotion Calendar

Set peak budgets from evidence, not vibes: last year's daily spend curve, this year's growth, and auction inflation. Your own history is the best forecaster available, the approach we detailed in setting paid search budgets from historical data. The specific failure to engineer against is budget capping during the highest-intent hours of the year; a campaign that exhausts its budget at 2pm on Black Friday spent its money in the cheap morning auctions and sat out the expensive evening ones. Whatever your target efficiency, decide now what you will do if volume runs hot: cap and hold efficiency, or fund the upside. Peak week is a bad time to convene that meeting.

Lock the promotion calendar with commercial stakeholders and translate it into ad operations: promotion assets, sitelinks and countdowns built in advance, ad variants written, Merchant Center promotions submitted early (approval queues lengthen in November), and landing pages load-tested with the speed stakes in mind. Freshen negatives while you are there: promotional queries surface new irrelevant traffic, and negative keyword hygiene pays double at peak CPCs.

Two Weeks Out: Teach the Algorithm About the Cliff

Smart Bidding handles gradual seasonality on its own; what it cannot anticipate is a step change in conversion rate the moment your sale goes live. That is the specific and narrow job of seasonality adjustments: you tell the system "expect conversion rate up roughly X per cent for these dates", it bids accordingly from hour one instead of spending two days discovering the sale, and critically, it unwinds cleanly when the event ends, avoiding the post-sale hangover where inflated bids chase vanished conversion rates. Set them for the sale window only, one to seven days, based on last year's observed lift. Do not use them for the whole of November; they are a scalpel for cliffs, not a lever for seasons, and Google's own guidance says exactly that.

If you run brand-building alongside, this is also when upper-funnel and Demand Gen activity should already be building audiences for the sale to harvest; starting demand creation in sale week is paying peak CPMs to introduce yourself.

During and After

In the event itself, monitor budgets, impression share and feed diagnostics daily, but resist optimisation theatre: the data is violently non-stationary, and reactive bid changes mid-event mostly add noise. Intervene on operational faults (capped budgets, disapprovals, broken pages), not on efficiency wobbles.

Afterwards, two disciplines. Remove or let expire the seasonality adjustments and watch the come-down week, because the post-peak conversion rate trough is where inflated targets quietly burn budget. And run the retro while the data is fresh: what capped, what broke, what the actual lift was against forecast, written into the account's testing and events register. Next September, that document is the head start.

Peak preparation is mostly the unglamorous enforcement of sequence: signal first, structure early, budgets from evidence, adjustments only at the cliff. If you want experienced hands running that sequence for your account this quarter, our paid search team is doing it now. Get in touch.