Black Friday Seasonality Adjustments Now Cost More Than They Return

Optmyzr compared 5,886 Google Ads accounts over Black Friday-Cyber Monday 2024. Advertisers who used seasonality adjustments saw CPCs rise 34% and ROAS fall 15.7%. Accounts that left Smart Bidding alone saw CPCs rise 17% and ROAS rise 5.7%. The adjusters got only about 6 extra points of revenue growth for that cost, and on 2024's numbers those extra dollars probably lost money.

The new write-up on MarTech is sponsored by Optmyzr, so keep that in mind. The underlying data comes from Optmyzr's three-year BFCM study, which covers 4,292 accounts in 2022, 5,114 in 2023 and 5,886 in 2024. Its recommendation is short: "Our data says to leave it alone."

I mostly agree with that. I also think the headline ROAS number understates the problem, and the study has a hole worth being honest about before you rebuild your Q4 plan around it.

The setting was built for surprises, and Black Friday is not one

A seasonality adjustment tells Smart Bidding to expect a temporary change in conversion rate during a set window. Google's own help page on seasonality adjustments says to use them "only if you expect major changes to conversion rates." It also says "Smart Bidding already manages seasonal events." Google recommends windows of one to seven days and warns that windows over 14 days tend to work less well.

Read those two sentences together and the guidance is fairly clear. The tool is meant for things the algorithm cannot see coming: your flash sale, a friends-and-family code, a big email push. Black Friday is the single most predictable shopping event of the year. Smart Bidding has years of auction history on it.

Plenty of advertisers use it anyway. A LinkedIn poll covered by PPC Land found that 65% of 169 respondents apply seasonality adjustments to Black Friday campaigns. That poll is self-selected, so I wouldn't treat it as an industry figure. It still suggests the habit is common rather than fringe.

Optmyzr's explanation for the backfire is double-counting. The bidding model already expects the conversion rate spike. When you add an adjustment, you stack a second, manual forecast of the same spike on top of the first one.

The lift you're forecasting keeps shrinking

This chart from the study is the one I'd show anyone who still adjusts by habit. Conversion rate lift during BFCM for advertisers who did not adjust, measured against a pre-holiday baseline:

• 2022: +17.5%
• 2023: +11.9%
• 2024: +7.5%

In two years the peak lost more than half its height. Optmyzr credits the change to shoppers spreading purchases across a longer promotional season. That fits what most of us have watched happen: "Black Friday" deals now start in early November and sometimes in late October.

So if you still enter +20% or +30% because that worked in 2021, you're bidding for a spike that probably isn't coming. The MarTech piece uses that exact case: guess 20% when the real lift is 7.5%, and you overbid across the entire window.

Think of it like adding salt to a dish the chef already salted. The kitchen didn't need your help, and the bill arrives anyway.

The math on those 6 points of revenue

Defenders of adjustments have a fair point here. Adjusters did grow revenue faster. In 2024, conversion value grew 39.9% for accounts with adjustments and 33.8% for accounts without them. If you have a pure volume mandate from the CFO, that 6-point gap is real.

Where it gets ugly is the cost of that gap. You can back out spend from the published percentages, because spend equals revenue divided by ROAS. (I did this rough math myself. It isn't in the study, so treat it as an estimate.)

• Non-adjusters: revenue up 33.8% at ROAS up 5.7% works out to spend up roughly 27%.
• Adjusters: revenue up 39.9% at ROAS down 15.7% works out to spend up roughly 66%.

So the adjusters spent about 39 extra points of baseline spend to get about 6 extra points of baseline revenue. That puts the marginal ROAS on the extra money at roughly 15% of the account's normal ROAS. An account that normally runs at 400% ROAS would get back about 62 cents per incremental dollar, and that's revenue, not margin.

Every extra point of revenue cost roughly six and a half points of spend.

In 2022 the same calculation comes out closer to 47% of normal ROAS. Back then the conversion rate spike was big enough that overbidding into it was at least partly defensible. With a 7.5% spike, it mostly isn't.

The hole in the study (and why I still buy the conclusion)

This is the part sponsored research tends to skip, although to its credit the MarTech piece does mention it. The study isn't an experiment. Advertisers picked their own group. The people who apply a +30% adjustment on Black Friday are often the same people who raise budgets, loosen tROAS and turn on every broad match keyword they own. The authors acknowledge that adjusters "may also be more aggressive about volume in other ways."

So some of that 15.7% ROAS drop is probably caused by general aggression, not by the adjustment itself. How much? Nobody knows from this data. Optmyzr also doesn't publish significance tests or the split between adjusters and non-adjusters in each year's sample.

Honestly though, the self-selection problem points the same direction as the conclusion. If the adjusters were also the most aggressive bidders, then the "extra revenue" credited to adjustments is inflated too. The mechanism (a model that already forecasts BFCM, plus a manual forecast stacked on top) holds up regardless of how clean the cohorts are. And it matches Google's own documentation, which isn't nothing coming from the company that sells the clicks.

One more caveat: the most recent seasonality data is from 2024. Optmyzr's separate Black Friday vs. Cyber Monday 2025 analysis looks at timing, not adjustments. Per MarTech's summary, Black Friday beat Cyber Monday on conversion value by 34% ($144.8M vs. $107.9M across 5,134 ecommerce accounts) and on ROAS (835% vs. 674%). That's useful for budget weighting. It doesn't tell us whether 2025 adjusters did better or worse.

If you need volume, move the target instead

The study's alternative is simple: if you want to trade efficiency for volume, lower the tROAS target instead of telling Google the conversion rate will jump. I think that's correct, and the reason is that you know exactly what you're asking for. A tROAS change says "I'll accept a 650% return instead of 750% this week." A seasonality adjustment says "I predict conversion rate goes up 25%." That's a forecast you're probably not qualified to make, and if you get it wrong, it shows up as overbidding in every auction for the whole window.

Black Friday 2026 is November 27 and Cyber Monday is November 30. Here's the plan I'd put in place in the next couple of weeks:

1. Pull your own 2024 and 2025 BFCM numbers. If you ran adjustments, compare CPC and ROAS change against your pre-holiday baseline, using the study's window (the Wednesday before Black Friday through the Wednesday after Cyber Monday). If your CPC inflation was well above the ~17% non-adjusters saw, the adjustment was probably paying for it.
2. Delete any scheduled BFCM adjustment for Black Friday, Cyber Monday and Christmas. Keep adjustments for events only you know about, such as a 24-hour flash sale or a big SMS send.
3. Decide your volume trade in tROAS terms. If leadership wants more revenue and will accept lower efficiency, write that down as a number ("drop tROAS 10% for 7 days") and apply it directly.
4. Set alerts before the week starts. MarTech's guardrail thresholds are reasonable defaults: flag spend running more than 20% ahead of pace, or CPC up more than 30% versus the prior seven days.
5. Check the conversion signal first. A double-counted purchase action will wreck BFCM bidding faster than any adjustment. We covered this in how Smart Bidding ends up optimizing for the wrong conversions.

Budget allocation matters too, especially now that Google lets you steer spend between channels. Our breakdown of Google's new channel prioritization controls is worth a read before you lock Q4 budgets.

A forecast you probably didn't need to make

My prediction: if Optmyzr reruns the analysis on 2025 data, conversion rate lift for non-adjusters lands somewhere around 5%, and the ROAS gap between the groups stays above 15 points. The peak keeps flattening, and the manual forecast keeps landing on top of it.

I don't think adjustments are useless. They're good for the events your account history can't explain. Black Friday stopped being one of those a while ago, and the extra revenue from adjusting now seems to cost more than it brings in.