Performance Max at Black Friday 2026: what to touch

Black Friday 2026 falls on Friday 27 November, with Cyber Monday on the 30th. The date almost nobody circles is Tuesday 1 December: Performance Max walks into December spending at peak pace against a conversion rate that has already collapsed. I’ve seen accounts hand back in one week a good chunk of the margin they earned over the previous weekend.

This post is about one thing: how PMax behaves when demand spikes and, above all, when it stops. It isn’t the build guide or the Shopping comparison. It’s what I do with campaigns that already exist.

In 30 seconds:

• A brand new PMax won’t make it in time: Google asks for six weeks of runway before you evaluate a campaign.
• Switching bid strategy type counts as a new strategy. Raising the daily budget isn’t on that list.
• Listing groups, not asset groups, are the real structural lever at peak.
• Brand traffic spikes in November and flatters your PMax ROAS.
• When the seasonality adjustment ends, the campaign reverts on its own. The targets you set don’t.

Can I launch a new Performance Max campaign for Black Friday?

If you’re reading this in November, no. Google is explicit: “Run new campaigns for at least 6 weeks. This allows the machine learning algorithm to ramp up and gather sufficient data to optimize performance” (Google Ads). Six weeks before 27 November lands you in mid-October.

On top of that sits the bid strategy’s own learning, which can take up to 3 weeks or 1-2 conversion cycles to calibrate. If your conversion cycle runs ten days, learning isn’t measured in clicks. It’s measured in calendar days, and the calendar doesn’t negotiate.

So what do I do when someone turns up on 10 November asking for “a PMax for Black Friday”? I work inside the campaign that’s already running: a seasonal asset group, a listing group that isolates the discounted range, and extra budget. Switching something on from scratch usually means paying for the learning at the most expensive auction prices of the year. If you’re still building PMax, start with the guide to Performance Max for ecommerce.

How do I raise budget without resetting learning?

In steps, and without touching the strategy type. The confusion I see most in audits is treating “raise the budget” and “change the bidding” as the same move.

Google lists the reasons a strategy enters the “Learning” status: a new or recently reactivated strategy, a settings change, and a composition change when campaigns or ad groups are added or removed. Daily budget isn’t on that list. That doesn’t make it a free-for-all either, because the PMax documentation says to “avoid making frequent changes to budget, bid strategy, or campaign status within this initial period”.

Google goes further than most people expect: for the highest-volume promotional days it recommends you “set a budget of 3 times what you want or expect to spend to ensure ample headroom across retail holidays”, and notes that campaigns may spend up to 2 times your average daily budget if demand increases. Budget stops being the ceiling. If your campaign is capped by 11am on the 27th, someone else wins the auction.

On target ROAS, the same page advises increments of 30% while waiting one or two conversion cycles between changes. That usually forces you to close the final correction in early November. The mechanics are in my guide to Smart Bidding in Google Ads.

What do I prepare in listing groups and asset groups?

Listing groups. It’s the part of PMax that gets ignored most in peak-season prep, and the one that gives you the most control over where the money goes.

Google says plainly that listing groups organize your listings by the attributes assigned to them in Merchant Center. Translated into November: isolate the range that’s genuinely discounted, rather than spreading budget across 4,000 SKUs of which only 300 carry a markdown. In October I create a custom label in the feed and build the listing group on top of it.

That unlocks the second half, because Google confirms “you may create multiple asset groups per campaign to group assets by theme or to group assets with different targeting”. An asset group with Black Friday creative pointing at genuinely discounted products is coherent. The same group advertising “up to 40% off” across the whole catalogue isn’t.

Two warnings, because I’ve watched both fail:

• Don’t fragment in November. Going from two asset groups to seven splits your data into small buckets exactly when you need dense signals.
• Watch exclusivity. If a product sits in listing groups across two campaigns, you’re bidding against yourself. That split is covered in Shopping versus Performance Max and in the Shopping campaign structure guide.

How do I refresh creative without breaking the campaign?

By adding, not swapping everything out, and around 10 November. It’s one of the few changes worth making close to the peak, as long as it happens inside the structure you already have.

Seasonal assets go in as extra material in the group that already exists. Google recommends that “if you make significant changes to an existing campaign, allow a similar learning period of 1-2 weeks (or at least one full conversion cycle)”, and emptying a whole asset group belongs in that category. Adding six images to a group of fifteen doesn’t.

The detail people forget is the expiry date. A headline saying “Black Friday” still serving on 4 December is a broken promise you pay for in CTR. Google’s post-holiday list says to “remove any assets or Merchant Center-based promotions that are no longer relevant”. Promotions and their review timelines I covered in Google Shopping before Black Friday.

What happens with brand and search themes in November?

Brand traffic spikes, and if you don’t separate it, your PMax looks better than it is. People who already know you search your name to see what you’ve put on offer. Those conversions were coming anyway.

Brand exclusions are the tool built for this. Google defines them as follows: “brand exclusions keep your campaigns from serving for branded queries you want to avoid”, and specifies that for Performance Max, brand exclusions apply to only Search, Shopping, and YouTube search inventory.

It isn’t automatic. If your brand Search campaigns cover those queries well, excluding brand in PMax finally gives you a clean read on the channel. If you don’t have them, you leave a gap competitors walk straight into. It depends on the account. For finer exclusions, I have a guide to negative keywords in Google Ads.

On search themes, Google lets you add up to 50 search themes per asset group and clarifies they “will have the same prioritization as phrase match and broad match keywords from your search campaigns”. That changes the game in November: a seasonal search theme isn’t a harmless suggestion, it competes with your own Search for the same query.

What won’t you be able to see in reporting at peak?

Less than you think, and precisely when you need it most. The PMax search terms report exists, with limits: Google states that its data is available starting from March 2023 and that store visits and store sales conversions aren’t available there. For a retailer with a physical shop, that’s a real slice of the result.

The underlying problem isn’t one missing report, it’s the lag. Volume multiplies, conversion delay stays the same, and on Saturday morning you’re looking at a Friday that’s still missing conversions. Every year someone pauses something on the Saturday over a number that turned out fine by the 3rd.

Across those four days I watch two things: spend pace and feed status. Everything else I read with 72 hours of delay. The full phase plan sits in Q4 2026 advertising strategy.

What do I do in the days after the peak?

Lower expectations before the auction lowers them for you. This is the stretch the accounts I audit handle worst, and the mechanics are simple: from the 27th to the 30th you train the system on an exceptional conversion rate, then on 1 December that rate vanishes while budget and targets are still calibrated for the previous weekend.

Two things are worth separating. If you used a seasonality adjustment, Google is clear: “your campaigns will optimize their bids during the events and return to their pre-adjust performance after the event is finished”, and “no negative adjustment is needed when the promotion is over” (Google Ads). That part sorts itself out. The rest doesn’t: the tripled budget, the target you relaxed, the assets with the words “Black Friday” on them, and the listing group of the discounted range.

Here’s the order I work in from 1 to 8 December:

1. Day 1, no bid changes. I take budget back to a point between October and the peak. In most cases, December demand isn’t October demand.
2. Day 1, creative clean-up. Out go the Black Friday assets and the expired Merchant Center promotions.
3. Day 2 or 3, listing groups. The discounted range isn’t discounted any more. It’s usually better to adjust its weight than to leave it bidding as if the markdown were still live.
4. Day 4 or 5, target ROAS. This is where Google’s advice to “reassess ROAS targets to ensure they’re aligned with post-holiday expectations” comes in. In increments of 30%.
5. Day 6, real reading. Only here is the peak data settled.
6. Following week, remarketing. The cold traffic you paid for in November is now sitting in your lists.

Point six leaves the most margin and is the one fewest people execute. November is expensive acquisition: you pay for clicks from people who didn’t know you, to sell to them at a discount. That investment doesn’t pay off on the 27th. It pays off if you reach those people again in December and January, when a click costs half as much. The guide to remarketing in Google Ads covers how to structure it.

One last warning. When day 2 ROAS looks bad, the typical reaction is to switch bid strategy: the worst possible moment, because you push the campaign into learning with less volume than October and with peak noise still in the history. Whatever genuinely needs changing gets changed in January.

Frequently asked questions

What’s the last date to launch a new PMax for Black Friday 2026?

Mid-October, counting back from 27 November. Google recommends running new campaigns for at least six weeks before evaluating them, and on top of that sits the bid strategy learning: up to three weeks. Past that date, I’d rather work inside a campaign that already has history behind it.

Does raising Performance Max budget reset learning?

Daily budget isn’t among the reasons Google lists for the “Learning” status: a new or reactivated strategy, a settings change, and a composition change. Even so, the PMax documentation recommends avoiding frequent budget changes during the initial period. I tend to raise it in steps.

Should I exclude brand traffic in PMax during Black Friday?

It depends on whether you have brand Search campaigns covering those queries. If you do, excluding brand in PMax gives you a clean read on the channel in the month that traffic distorts it most. If you don’t, you leave the door open to competitors. Decide it in October.

How much budget should I set for the peak weekend?

For the highest-volume promotional days Google recommends a budget three times what you expect to spend, and notes a campaign may spend up to twice your average daily budget when demand rises. The point isn’t to spend triple. It’s that budget shouldn’t be what knocks you out of the auction.

When do I go back to normal targets after Black Friday?

In stages across the first week of December, not on the 1st. If you used a seasonality adjustment, the campaign returns to its pre-adjustment performance on its own once the event ends, and no negative adjustment is needed. What you fix by hand is budget, expired assets and target ROAS.

December decides what November was worth

The expensive Performance Max mistake at Black Friday almost never happens on the 27th. It happens in October, when someone switches on a new campaign thinking there’s still time, and on 1 December, when nobody brings budget and targets back to reality.

What surprises me most after several Q4 cycles is how attention gets shared out. Weeks go into preparing four days. Then nobody looks at the account until mid-December, exactly when the campaign is worst calibrated and an hour of work pays better than any of the hours you put in back in October.

If you want me to review your PMax structure before November, or help you design the exit from the peak, book 30 minutes of consulting.