Smart bidding exploration gets a calendar: who holds the dial in Promotion Mode

Smart bidding exploration gets a calendar: what Promotion Mode is, and what is unconfirmed#

A fox distrusts a rumour that arrives without a trail, so begin with the paper trail. On 29 September 2026, Search Engine Watch reported that Google Ads is testing a beta campaign setting called Promotion Mode, powered by smart bidding exploration, and that it is not yet in every account. The report credits Vivek Gupta with first spotting it on LinkedIn, and lists three controls: dates for the promotion, an adjustable target ROAS tolerance, and an optional extra daily budget. When the dates pass, the campaign returns to its original target ROAS and budget.

Here is the wrinkle that turns a feature note into a story about control. The setting is not brand new. Google announced Promotion Mode in beta for Search and Performance Max in mid-June, when Ginny Marvin, Google's Ads Product Liaison, set out the change, as Search Engine Roundtable recorded on 15 June 2026. Google's own announcement page describes it as a way to schedule temporary changes to your ROAS tolerance and add extra daily budget during peaking periods. So last week's sighting is best read as a slow, uneven rollout reaching more accounts, not a surprise launch, and we found no published list saying which accounts qualify.

The mechanics, as reported, are tidy. Marvin said the very first prompt asks for a start and end date, with a span of 3 to 14 days, so nobody has to remember to switch it off, per PPC Land. The ROAS tolerance change works with daily budgets and can also be combined with campaign total budgets, per the same report. One restriction has already turned up in the undergrowth: Search Engine Roundtable reported on 25 September that Google Ads refuses edits to Target ROAS while promotion mode is active, and offers the ROAS tolerance field instead.

What is being flexed is the target itself. Google's setup guidance says the system lowers your effective target ROAS by the tolerance you choose: with a 200% target and a 10% tolerance, the effective average target becomes 180%. The tolerance can run from 5% to 30%, per Google's best-practice page. The chart applies that rule across the band, and it is the whole argument of this piece in one picture: a calendar and a dial sit in the same hand.

Three things stay unconfirmed in what we read. First, whether the 3 to 14 day window is a fixed ceiling for every account. Second, how Promotion Mode behaves beside a seasonality adjustment, since PPC Land noted that Google's documentation does not address the interaction. Third, how much extra daily budget is permitted, because the reports describe the budget field as optional and give no figure. We found Google's announcement page but no dedicated help article for the setting, so treat every number here as reported, and re-read the setting in your own account before you build a plan on it.

ROAS tolerance is a price decision, not a tuning knob#

Strip away the interface and a ROAS target is a price. A 400% target means each 100 of revenue is allowed to cost 25 of spend. Lower the effective target to 280%, the bottom of a 30% tolerance, and the same 100 of revenue is allowed to cost about 35.7. That is a 43% richer price on the marginal auction, and it is a price you have, in effect, pre-approved for the length of the window. The table below sets out the brush strokes from 5% to 30%.

Read the table as a ceiling on what the system may accept, never as what it will deliver. Smart bidding exploration exists to find conversions you would not otherwise win, and Google says so plainly: in its launch post, the example lender is invited to bid on less obvious but potentially valuable queries more often. Google adds, from its own internal data for March to April 2025, that campaigns using the feature saw on average an 18% increase in unique search query categories with conversions and a 19% increase in conversions. That is a vendor figure from a 2025 sample, so hold it as a claim, not a benchmark for your account.

The marketer's question is whether those extra conversions are worth the richer price, and a promotion is the hardest moment to answer it. Demand is already spiking, so extra conversions may simply be sales that would have happened anyway at a lower cost. Our earlier piece on the 17 August budget-cap change and our read of the bidding change that ended the cheap conversion both turn on this point: Google's help page says budget-limited campaigns now deliver closer to their stated target, and that Google will not adjust your targets or budgets for you.

There is a second, subtler link. Google's best-practice page warns that smart bidding exploration will not be effective if a campaign is limited by budget. A promotion window that adds daily budget therefore removes the very constraint that would have muted the exploration. That is our inference, not Google's wording, but it explains why the two levers travel together: loosen the target, then give the system the room to use it.

The durations in that chart come from two different Google statements, so read them as a contrast, not a like-for-like. The window length is Marvin's, per PPC Land. The other figures are the guidance in Google's best-practice page: allow one to two conversion cycles or 7 days before you evaluate, test for six weeks or more, and expect a further one to two weeks for the system to ramp up. Fourteen days cannot hold all of that, which is why the test comes before the promotion, not during it.

Google Ads promotion mode in a regulated category: where the guard rails go#

• 3 to 14

  days, the reported span of a promotion window

• 5 to 30%

  the ROAS tolerance Google's guidance allows

• 6 weeks

  the suggested minimum test for exploration, plus ramp-up

• 24 hours

  the ideal upload time for backend conversion data

A regulated advertiser does not get to treat a looser target as a free lunch. Promotion Mode cannot widen what policy permits, and Google's rules are blunt about that. Its gambling policy allows online gambling advertising only in listed countries where country-specific certification criteria are met, and requires advertisers to apply for certification. Its healthcare policy says advertisers must apply to serve ads for prescription drug services and check location-specific restrictions for every location they target.

What the flex can change is the thicket of queries inside the footprint you are already allowed to serve. Smart bidding exploration bids on less obvious searches more often, and the less obvious the search, the more likely it is to sit near a line your compliance team drew. The extra budget then feeds that exposure. A looser target in a calm category is a tuning choice. In a category where a wrong query can mean a disapproval, an account suspension or a regulator's letter, it is a risk decision, and someone senior should sign it.

That line, from a daily industry recap on Reddit, is easy to skim past and expensive to ignore. If you spot a bad batch of queries on day two of a window, the quick fix of nudging the target is closed to you. Your levers become the tolerance, the budget, the negative keywords and the pause button, so decide before the window opens which of them you will pull first.

Our own guard rails for a window in a sensitive category are short enough to pin to the wall. Set a tolerance at the low end of the band for a first run, and write down the highest spend per 100 of revenue you will tolerate, using the arithmetic above. Fix a budget ceiling in advance. Refresh your negative keyword lists and brand-safety exclusions the week before, not the day of. Name the person with authority to pause. And agree the stop condition in plain numbers, not in a mood.

These are the same instincts we bring to iGaming marketing and healthcare marketing, where the cost of a loose rule is rarely the media bill and usually the account. They are also why our recent look at demand gen and sensitive categories argued for smaller doors and slower trust.

Google Ads Black Friday, seasonality adjustments and the plumbing beneath the flex#

A google ads black friday plan is the obvious home for this feature, and the obvious trap. A fortnight at most is long enough for a sale and short enough that a poor start eats a large share of the window. If your promotion has a hard commercial deadline, treat the first 48 hours as a measurement period with a pre-agreed exit, not a launch party.

Do not assume Promotion Mode replaces google ads seasonality adjustments. Marvin explained that seasonality adjustments address conversion rate expectations, while promotion mode adjusts ROAS tolerance and budget, and that both can run together on Search and Performance Max, per PPC Land. The same report notes that Google's documentation does not say how the two signals interact when they overlap. Until it does, run one at a time on your first test, so that you know which lever produced which result.

Then comes the plumbing under the flex, and it is the part most teams skip. A looser target amplifies whatever the conversion signal says, true or false. Google is documenting a multi-source conversions beta that lets advertisers connect a website tag to backend data from a CRM or order system, as Search Engine Watch reported on 28 September 2026. Google's help page says uploads must carry a transaction ID and the conversion date and time, plus at least one attribution identifier, and recommends uploading within 24 hours of the conversion.

Google's multi-source help page carries a warning that is worth more than the headline. When an uploaded record shares a transaction ID with the tag event, the uploaded conversion value becomes the source of truth and permanently overwrites the value the tag recorded, and Google says the tag and the data source must use the same units, such as dollars versus cents. It also notes that value updates are disabled during the initial 14-day trial period for a conversion action.

Put those together with a 30% tolerance and you can see the hazard: a unit error becomes an inflated ROAS, and an inflated ROAS becomes a licence to overspend. Our guides to measurement data strength and enhanced conversions are the place to start.

Target ROAS vs target CPA, and a test to run before the window#

Which campaigns can use any of this? The reports describe a ROAS tolerance, so a campaign built on target CPA has no ROAS tolerance to flex, and we found nothing saying otherwise. That is our reading, not a Google statement, and it reopens the old target ROAS vs target CPA question for any advertiser planning a promotion. Our honest brief on target ROAS sets out when value-based bidding earns its keep, and the answer depends on whether your conversion values are trustworthy, which is where the multi-source work above comes in.

A sensible sequence therefore starts weeks before the sale. Use Google's guidance as the spine: run an experiment with smart bidding exploration as the only difference between control and trial, hold your Target ROAS steady during the test, and wait out the conversion cycle before you judge. Our note on why smart bidding now waits for fifty conversions matters here, because a thin account feels every edit and a flexed target is a large edit.

Measurement should be as dull and disciplined as the guard rails. Count incremental conversions and margin, not platform-reported conversions alone. Hold back a comparable campaign or region where you can. Record the spend per 100 of revenue each day, since that is the number the tolerance is silently renegotiating. And watch query reports for the new, odd terms exploration brings in, because that is where a regulated advertiser meets surprises first.

If you run paid search in a category Google watches closely, this is a good week to look at your windows, your approvals and your data. Our PPC service builds the quiet scaffolding around those windows, and our fintech marketing work shows how the same discipline travels. The fox that prowls patiently will outfox the one who simply presses the button on the first moonlit morning.

Frequently asked questions#