Promotion Mode in Google Ads 2026: Scheduled Peak Windows for Search and PMax
On day one of a sale, demand triples — and your target ROAS campaign serves roughly the same impression volume it did last Tuesday. The algorithm is not being stubborn. It simply does not know today is different, so it keeps honouring the target you set. The usual workarounds (drop the target manually, raise the budget, remember to undo both afterwards) work, but they cost you: every target edit disturbs learning, and a forgotten rollback quietly eats margin for another week. That is exactly the scenario promotion mode in Google Ads was built for.
In June 2026 Google announced promotion mode — a scheduled window that lets you relax bidding constraints for a defined peak and then reverts on its own. This article covers what it actually changes, how it differs from seasonality adjustments, how to launch a window properly, and how to measure whether it did anything at all.
What promotion mode in Google Ads actually does
Promotion mode is a scheduled window of 3 to 14 days during which a campaign gets two temporary changes:
- a wider ROAS tolerance — the algorithm is allowed to buy traffic it would normally discard at your stated target;
- extra daily budget on top of the campaign’s current daily budget, for the duration of the window only.
The important design detail: this is not a toggle you have to remember to switch off. The first thing the setup flow asks for is the start and end date. When the window closes, the campaign returns to its original settings automatically.
At the time of writing it is in beta for Search and Performance Max campaigns. It works with daily budgets and can be combined with campaign total budgets.
Promotion mode does not make a promotion profitable. It removes an artificial constraint that causes campaigns to under-buy on exactly the days when conversion rate and order value are genuinely higher than usual.
How it differs from the tools you already use
The most common misreading is treating promotion mode as a new wrapper around seasonality adjustments. They pull different levers.
| Tool | What it changes | When it fits | Risk |
|---|---|---|---|
| Seasonality adjustments | Expected conversion rate — you tell the system “CR will be X% higher on these dates” | Short spikes (1–7 days) with a known CR lift | Overstate it and you overpay per click; understate it and you under-buy |
| Manually lowering tCPA/tROAS | The strategy target itself | A genuine, lasting change in economics | Sharp edits disturb learning; easy to forget to revert |
| Raising the daily budget | The spend ceiling | Campaign is budget-constrained, not target-constrained | If the campaign is target-constrained, extra budget simply will not spend |
| Promotion mode | ROAS tolerance and daily budget, inside a fixed window | A planned 3–14 day event: sale, season, launch | Too wide a tolerance buys irrelevant queries; measurement requires discipline |
Practical takeaway: seasonality adjustments and data exclusions remain the right tool for short conversion-rate spikes — the email-blast day when CR predictably jumps 30%. Promotion mode does something else entirely: it does not forecast conversion rate, it authorises the system to pay more across a wider set of queries and funds that decision.
Both can technically run over the same dates. Google does not document how they interact, so on a first run it is smarter not to stack them — measure promotion mode on its own first.
ROAS tolerance in plain terms
ROAS tolerance is the same mechanic that sits behind Smart Bidding Exploration: you define how far the system may drift from your target in pursuit of queries the campaign has not seen before. Google’s own figures for exploration show an average +18% in unique search query categories with conversions and +19% in conversions. Treat those as platform benchmarks, not a promise — the lift depends heavily on how narrowly you were targeting beforehand.
Three things worth internalising:
- It is not simply “a lower tROAS”. Your target still anchors the core of the traffic. Tolerance widens the perimeter, letting the system test adjacent intent with weaker expected returns.
- Average ROAS during the window will very likely fall. That is the expected shape, not a failure. Judge the window on absolute gross profit, not on a percentage.
- Exploration needs signal to work with. Thin conversion volume and missing value data mean nothing to learn from. Get conversion value and value rules right before you widen anything.
Extra budget and how pacing treats it
The second lever is additional daily budget for the window. Remember the underlying arithmetic: Google multiplies your daily budget by 30.4 to derive a monthly cap, and a single day can spend up to twice the daily budget. That mechanic is unpacked in our guide to Google Ads budget pacing and the 30.4× rule.
What follows for a promo window:
- Extra budget is not “permission to spend total X across the window”. It is a raised daily ceiling that still obeys the 2× daily overspend behaviour on peak days.
- If the campaign runs on a campaign total budget, promotion mode operates inside it — and that total remains a hard cap. Check it is not quietly neutralising the whole exercise.
- Schedule at least 3–5 post-window days at original settings. Without them you cannot separate the promotion’s effect from the algorithm’s momentum.
Launching a window, step by step
Step 1. Confirm the campaign is target-constrained, not something else
Open the campaign and read impression share. High lost IS (rank) with near-zero lost IS (budget) means the target is the brake and promotion mode is the right instrument. The reverse means you just need more money. Our breakdown of impression share and auction insights covers how to read those columns.
Step 2. Set dates with a buffer on both sides
Minimum 3 days, maximum 14. Start the window a day before the promotion actually goes live — landing pages, codes and feed need to be ready before traffic climbs. End it a day after, so delayed conversions land inside the measured period.
Step 3. Set tolerance conservatively on the first run
Do not max it out first time. You are buying information: a moderate tolerance on run one tells you which new query categories actually convert, and lets you widen deliberately at the next peak.
Step 4. Budget 30–50% above your projection
The classic mistake is sizing extra budget exactly to planned revenue. Wider tolerance buys more expensive traffic by definition; if the ceiling is hit on day two, the system throttles impressions precisely when demand peaks.
Step 5. Capture a baseline before you start
Export the 14 days before the window: spend, conversions, conversion value, AOV, CR, impression share. Without that snapshot, the post-mortem becomes an argument about impressions rather than numbers.
Step 6. Make sure the landing pages survive the surge
Promo traffic means a load spike and new URLs — sale pages, coupon landers. Two failures show up every peak season: the page starts returning 5xx under load, or the new URL is unreachable to the crawler and the ad gets disapproved. The first is covered in landing page speed and Core Web Vitals for paid traffic, the second in our guide to AdsBot and landing page crawlability. Both checks take under an hour and can save the entire window’s budget.
Promotion mode in Performance Max: what changes
In Search, wider tolerance is legible: new phrasings appear in the search terms report and you can trim what does not belong. In Performance Max the picture is murkier, because the expansion touches several channels and your whole catalogue at once.
- Expansion is uneven across channels. Extra budget rarely lands where you assume — the cheapest inventory usually absorbs the increment first. Look at how spend splits by channel before the window opens.
- The product feed becomes the bottleneck. If part of the catalogue is disapproved or thinly populated, extra budget concentrates on what remains, which is typically the cheapest and least profitable subset. Run a pass over Merchant Center product disapprovals and feed diagnostics first.
- Audience signals and search themes act as the frame. Too narrow and there is nowhere to expand; too wide and the mode accelerates drift into irrelevant traffic. See search themes in Performance Max.
- Cannibalisation intensifies. A PMax campaign with wider tolerance takes queries that previously went to Search. Running both on the same products makes the window’s effect hard to separate from traffic simply moving between campaigns — see campaign cannibalisation across PMax, Search and Shopping.
Recommendation: if both PMax and Search cover the same products, enable the mode on one campaign type only for the first promotion. You will get an interpretable result instead of a blend of two effects.
Measuring the effect honestly
A promo window overlaps, by definition, with a period when demand rises on its own. “Sale week versus last week” therefore proves nothing about promotion mode — it measures the sale.
- Year-over-year on the same event. If last year’s sale ran without the mode, compare spend and gross profit trajectories adjusted for overall account growth. Crude, but better than nothing.
- Split by campaign. Enable it on half of a set of structurally similar campaigns and leave the rest untouched. Not clean, but directional.
- A proper experiment. The most defensible route is running the comparison through Experiment Center, where traffic is split randomly rather than by your choice of campaigns.
| Metric | Expected | Warning sign |
|---|---|---|
| Spend | Rises within the extra budget you allowed | Hits the ceiling on day two |
| Conversions | Grow faster than spend in the first days | Spend +60%, conversions +5% |
| ROAS | Falls — that is the mode working as designed | Falls below your margin break-even |
| New query categories | Converting queries you had not seen before | A flood of irrelevant queries with no conversions |
| Impression share | Rises; lost IS (rank) falls | Flat — the tolerance never engaged |
One habit matters more during a window than in ordinary weeks: read the search terms report daily. Wider tolerance is precisely the condition under which new phrasings enter the campaign, and some of them will be waste.
Seven mistakes that eat the upside
- Starting the window the day of the peak. The system needs adaptation time; a window that opens on sale morning spends half its value ramping.
- Editing targets at the same time. Dropping tROAS manually on day one makes the result uninterpretable and hits learning twice. See the Smart Bidding learning period.
- Booking 14 days “just in case”. Long windows dilute the effect and inflate the bill. Window length should equal promotion length plus a day or two.
- Ignoring the campaign total budget. Extra daily budget is meaningless if a period total caps the campaign anyway.
- No baseline. Without a “before” snapshot, any outcome can be argued either way.
- Running a window on broken tracking. Wider tolerance amplifies noise: more options for the algorithm, but a distorted picture if conversions are not arriving. Run a 30-minute conversion tracking diagnosis before the peak.
- No plan for the aftermath. The campaign snaps back to its original target while remarketing lists stay inflated. Allow 3–5 days of normalisation and do not panic at a dip two days after the sale.
Two-week pre-peak checklist
- Baseline exported and stored (spend, conversions, value, CR, impression share).
- Verified what constrains the campaign: target or budget.
- Acceptable CPA/ROAS recalculated for promotional margin — discounted margin means a different threshold.
- Extra budget agreed with a 30–50% buffer.
- Promo URLs built, crawlable, speed-tested, mobile-checked.
- Prices and offers synchronised between feed and ads (for retail, also Merchant Center promotions and annotations).
- Measurement method chosen (experiment / campaign split / year-over-year).
- Someone named as owner of the daily search terms review during the window.
The wider preparation playbook — offer, warm-up, staffing the peak — is covered in the four-week promo peak playbook. Promotion mode belongs to the final week of that plan; it is not a substitute for it.
When not to use it
- The campaign is already budget-constrained. Widening tolerance changes nothing — fund it first.
- Fewer than 15–20 conversions a month. Too little signal; exploration becomes a lottery.
- No genuine peak. The mode assumes a demand spike. Without a sale, season or launch, it just raises spend.
- Near-zero promotional margin. If the sale is already marginal, wider tolerance accelerates the loss. Recalculate an acceptable cost per conversion at discounted margin first.
- Fragile landing pages or thin stock. Buying more traffic for an item that sells out on day three converts budget into bad reviews.
One more scenario worth naming: if your business also sells offline, part of the promotion’s effect lands in stores and never appears in the account. Pulling that result back into reporting is covered in store visit conversions in Google Ads.
What this says about where Google Ads is heading
Promotion mode is a symptom of a broader shift: Google keeps taking day-to-day bid control away and handing back “scenarios” — pre-declared situations in which the algorithm behaves differently. The same direction is visible in the August change to how target strategies behave under budget constraints, covered in target-based bid strategies and what changed in August 2026.
For a team, the practical consequence is that value moves from manual edits to preparation: clean signal, correct conversion values, landing pages that hold, and promotional economics calculated in advance. With that base, promotion mode adds. Without it, the mode just spends faster.
If you need account infrastructure to run volume like this, see Google Ads agency account rental and the rest of the PPC Rebels services — though the mode itself requires nothing but preparation in any account that has beta access.
FAQ
Is promotion mode available to every account?
No. At the time of writing it is a beta for Search and Performance Max campaigns, rolling out gradually, so it will not appear in every account or region.
What are the minimum and maximum window lengths?
Three to fourteen days. Start and end dates are the first thing the setup flow asks for.
Do I have to turn it off manually after the sale?
No — that is its main advantage over manual target edits. When the window closes, the campaign reverts to its original settings by itself.
Can I combine it with seasonality adjustments?
Technically yes; they control different parameters, one signalling an expected conversion-rate change and the other changing ROAS tolerance and budget. Google does not document the interaction, so measure them separately on a first run.
Does it reset the learning period?
It is designed as a temporary relaxation within an existing strategy rather than a target change, so the effect is gentler than editing tROAS by hand. Any material shift in auction economics still takes the system a few days to absorb — build that into your dates.
What should happen to ROAS during the window?
It will most likely fall. That is expected: you authorised more expensive traffic. Evaluate gross profit for the period instead.
Does it work for Shopping campaigns?
The announcement names Search and Performance Max. Shopping and feed-based PMax received an expansion of Smart Bidding Exploration separately — related, but a different feature.
What if the campaign uses a shared budget?
Shared budgets distribute spend across campaigns and can absorb the effect of extra daily budget. Before a promo window, it is usually worth moving the campaign onto its own budget.
How much extra budget should I allow?
As a rule of thumb, 30–50% above your projected spend increase for the period — a buffer for the fact that wider tolerance buys pricier traffic than your plan assumed.
How do I know it actually engaged?
Three signs: total impression share up, lost IS (rank) down, and new converting phrasings in the search terms report. If none of the three moved, the window was wasted — check whether budget or a period total is the real constraint.
Is it useful for lead generation rather than e-commerce?
Yes, provided lead value is passed back correctly and you know which leads qualify. Otherwise wider tolerance mostly delivers more cheap, unqualified enquiries. Getting primary and secondary conversions right is the prerequisite.
What should I do immediately after the window closes?
Nothing dramatic for 3–5 days. The campaign returns to its original target and restabilises. Compare against baseline only once your conversion window has closed.
