← All posts

Performance Max and Shopping for Black Friday: What to Change

For Black Friday, your Performance Max and Shopping results are decided in the feed, not the campaign. Here's exactly what to change, and what to leave alone.

For ecommerce, Black Friday is won or lost in Performance Max and Shopping, and most of what decides the outcome happens in the feed and Merchant Center, not in the campaign settings everyone fiddles with. The advertisers who struggle at peak are usually the ones optimising bids while their promotions sit unapproved and their feed quietly disapproves half the catalogue.

Here's the practical list of what to actually change for Black Friday across PMax and Shopping, and, just as important, what to leave alone.

The TL;DR

Get the plumbing right first. Submit deals through the Merchant Center Promotions Editor and signal them with the sale_price and sale_price_effective_date feed attributes, so shoppers see real struck-through prices. Build holiday-specific asset groups inside your existing Performance Max campaigns rather than new campaigns, and upload them at least two weeks ahead for approval. Give budgets real headroom and move targets gradually, days in advance. Then stop. Resetting bidding or rebuilding structure the week before peak spends the event in the learning phase. The feed is the campaign; treat it that way.

Start with the feed, because it decides everything

In an automated world, Performance Max and Shopping serve your products based on the quality and completeness of your feed. During peak, that's doubly true, because more clean, in-stock, competitively-priced offers give the machine more to work with.

The feed is the campaign The feed is the campaign Product feed titles, GTINs, stock Merchant Center Promotions Editor sale_price attributes Performance Max matches + serves ShoppingSearchYouTubeDisplay · DiscoverGmail · Maps
In an automated world the feed, not the campaign settings, is the lever. Clean offers in; relevant placements out.

Jobs to do now:

  • Fix disapprovals and warnings before peak. A disapproved product is invisible on the biggest day of the year, and reinstatement isn't instant.
  • Tighten titles and attributes on your priority products. Clear titles, correct GTINs, accurate availability.
  • Confirm stock and price accuracy. Showing a sold-out or wrongly-priced item at peak wastes spend and erodes trust at the worst time.

If your bidding runs on revenue rather than profit, peak is when that gap bites hardest, because a surge of low-margin sales can look like success while quietly losing money. That's the entire case for a margin feed that lets tROAS optimise real profit.

Signal your deals the way Google reads them

A sale only works if Google knows it's a sale. Two mechanisms matter:

  • Merchant Center promotions. Submit every offer through the Promotions Editor. It's the route that keeps your approval rate healthy and surfaces the "special offer" messaging on your listings. Ad-hoc edits are how deals end up unapproved on Black Friday morning.
  • Sale price attributes. Use sale_price and sale_price_effective_date so the struck-through original price and the sale annotation appear. This is what makes a deal look like a deal in the results, which is half the battle on a day when every listing is shouting.
What the sale_price attributes produce What sale_price produces in the results product image BLACK FRIDAY Wireless Headphones £89 £129 −31%
Set correctly, the feed shows the old price struck through, the sale price, and a deal badge — which is what makes shoppers stop.

Submit and get these approved early. Approvals take longer during peak, and a promotion that clears review on 28 November is a promotion you wasted.

Structure: add holiday asset groups, don't rebuild

The instinct to spin up shiny new Black Friday campaigns is the instinct to resist. New campaigns start cold, with no learning, right when you need stability.

Add holiday asset groups inside the existing campaign Add holiday asset groups — keep the campaign Existing Performance Max campaign keeps all its accumulated learning Evergreen asset groups always on running Holiday asset groups scheduled: on for BFCM, off after 27 Nov → 2 Dec
Keep the campaign (and its learning); add seasonal asset groups and switch them on with scheduling. Don't launch cold new campaigns at peak.

The better pattern:

  • Create holiday-specific asset groups inside your existing Performance Max campaigns. You keep the campaign's accumulated learning and add seasonal creative on top.
  • Use asset-group scheduling to switch holiday assets on for the event and off afterwards, so you're never advertising an expired deal.
  • Keep your evergreen asset groups running. Don't cannibalise what already works.

On control, remember that Performance Max is less of a black box than it used to be. Account-level and campaign-level negatives, brand controls and channel reporting all still apply during peak, and they matter more when spend is high. If you're not using them, the search categories report and account-level negatives are where to start. And if you lean on audience signals to guide the machine, peak is a good time to make sure they're genuinely helping and not just reassuring you, a point I've made about PMax audience signals.

Budgets and targets: headroom, then patience

  • Budget headroom. Google suggests planning for up to roughly 3× normal spend on retail peaks (around 2× if you have no prior benchmark), and campaigns can spend up to twice their daily budget on high-demand days. Set the headroom before the event so you're not capped mid-surge.
  • Targets, moved gently. If you want more volume, ease your Target ROAS down in steps a few days ahead, letting each change settle. If the goal is maximum sales within a fixed budget, Maximise Conversion Value is often cleaner than a strict tROAS for the peak itself.
  • New-customer acquisition. If repeat buyers would convert anyway, Performance Max's new-customer goals let you value first-time buyers more highly, so your peak spend buys growth rather than discounts for people who'd have bought regardless.

During the event: leave it alone

Once the feed is clean, the promotions are approved and the budgets have headroom, the highest-value thing you can do during Cyber Week is not much.

DoDon't
Monitor pacing and approvalsWatch hourly ROAS and panic
Lift a budget that's cappingReset the bid strategy mid-event
Use a data exclusion if tracking breaksRebuild asset groups during peak
Let learning hold steadyReach for a seasonality adjustment to "fix" bad data

A short composite

A composite from the ecommerce accounts I've run, not one store. A mid-size retailer used to launch three brand-new "Black Friday" PMax campaigns each November. Every year they underperformed the evergreen campaigns, because they spent the first few days learning from scratch. We switched to holiday asset groups inside the existing campaigns, got the Merchant Center promotions approved a fortnight early, and gave budgets proper headroom. The peak ran calmer and the spend worked harder, not because of a new tactic, but because nothing was starting cold on the day it mattered.

Common mistakes

  • Optimising bids while the feed is broken. Fix disapprovals and prices first.
  • Submitting promotions late or outside the Promotions Editor. They won't clear review in time.
  • Launching brand-new campaigns for peak. They start cold. Use holiday asset groups instead.
  • Over-managing during the event. Resets cost you the learning you spent weeks building.
  • Ignoring PMax's controls. Negatives, brand controls and new-customer goals matter most when spend is highest.

Bottom line

  • Your feed and Merchant Center promotions decide Black Friday more than your bids do. Get them clean and approved early.
  • Signal deals with the Promotions Editor and sale-price attributes so they look like deals.
  • Add holiday asset groups inside existing campaigns; don't rebuild from scratch.
  • Give budgets headroom, move targets gradually, and value new customers.
  • During the event, monitor and leave the bidding alone.

Peak rewards preparation and punishes panic. The retailers who win Black Friday on Performance Max aren't reacting faster than everyone else on the day; they got the feed, the promotions and the structure right two weeks earlier and then had the nerve to sit still.

The feed is the campaign. Get it clean two weeks early, then have the nerve to sit still.

If you want your feed and PMax setup checked before the rush, the free audit still has just enough runway before the 27th. To have it set up and managed for you through peak, see my Google Ads management service.

Sources and further reading

Use Smart Bidding with Shopping and Performance Max — Google Ads Help
Create and manage promotions — Google Merchant Center Help
About promotion assets — Google Ads Help
About seasonality adjustments — Google Ads Help

Related Reading · Performance Max & Campaign Types
Google Ads Management →

Demand Gen vs Performance Max: 90-Day Test With Real Data

Demand Gen vs Performance Max (PMax) compared in a 90-day side-by-side test. When to use each, when to run both, and how to stop them cannibalising.

PMax Audience Signals: Do They Actually Matter? (Test Data)

Performance Max audience signals tested across six campaigns in three accounts. What they do, when they matter, and how to use them on new vs mature accounts.

Performance Max for Lead Gen: Offline Conversion Setup

How to set up Performance Max for lead gen: offline conversions, GCLID capture and pipeline tracking, with a B2B SaaS account that cut cost per opportunity 58%.

Found this useful?

That's the kind of thinking I bring to every account I work on. If your Google Ads need a pair of senior eyes, the call is free and I respond personally within 24 hours.

Book a free 30-min call →Or get a free account audit