Google narrows Branded Searches to a 7-day window and drops Performance Max support

Google narrowed the attribution window on its Branded Searches conversion action from 30 days to a 7-day default, and removed Performance Max from the campaign types that support it. Both changes surfaced on 18 August 2026 inside Google's own Branded Searches help documentation, where Search Engine Land found them, and Google published no rollout date and no changelog note to go with either one.

The 30-day measurement is not gone. It is no longer the default. An advertiser can set the Branded Searches attribution window to any value between 1 and 30 days, but an account nobody touches will report on 7 from now on. The supported campaign list now reads YouTube and Demand Gen. Performance Max, which was included when Branded Searches launched in 2025, is off it.

What the Branded Searches conversion action measures

Branded Searches is the Google Ads conversion action that records when someone searches for an advertiser's brand after exposure to that advertiser's ads, and as of the documentation update reported on 18 August 2026 it sits formally under the Consideration goal. It is a primary conversion action for reporting purposes only: it appears in the Results and All Conversions columns, and it is not selectable as a Smart Bidding optimisation target. The August 2026 change left that split exactly where it was.

The metric also does not run by itself. Google's documentation requires the advertiser to configure brand mapping, so the account knows which terms count as the brand, and that requirement applies even to advertisers who are not running a Search Lift experiment. An account that never completed brand mapping has been reporting nothing under Branded Searches, whatever the window was set to.

Where the data can be read is unchanged: campaign level, ad group level, asset level, and Report Editor.

Google set the default view-through window to 7 days

Per the Branded Searches documentation as reported on 18 August 2026, Google's default view-through attribution window for the metric is now 7 days, down from the 30-day window the metric originally shipped with, with an advertiser-controlled range of 1 to 30 days.

Analysis, not a claim from the source: a view-through attribution window is a stopwatch that starts when an ad is seen and not clicked. If the branded search happens before the stopwatch stops, the platform books it. Shorten the stopwatch from 30 days to 7 and every branded search that used to land on day 8 through day 30 falls outside the count. No media changed. No audience changed. No creative changed. The definition changed, and the reported number moves down with it.

How far down is not knowable from outside an account, because it depends on the shape of that account's lag between exposure and brand search. An impulse category where people search the same week may barely register the difference. A category where someone watches a video ad and goes looking for the brand three weeks later loses the whole tail of the curve. Property, insurance, education, medical and high-ticket B2B sit in that tail by nature, because the purchase itself takes months.

The practical trap is that two accounts reporting a Branded Searches decline over the coming weeks can be in completely different situations. In one, real brand interest fell. In the other, only the window did. Nothing inside the metric separates them, which is why the change is worth writing down in the account notes before the next reporting cycle rather than after it.

Performance Max no longer supports Branded Searches

Google's updated documentation, reported on 18 August 2026, lists YouTube and Demand Gen as the campaign types that support Branded Searches, and Performance Max is no longer among them despite having been supported when Branded Searches launched in 2025.

For a Performance Max advertiser this is not a smaller number, it is the loss of a measurement. Branded Searches was, in some accounts, the only in-platform figure that showed Performance Max generating new brand interest rather than harvesting demand that already existed. That figure is gone from that surface. Any dashboard, recurring export or client deck that pulls Branded Searches out of a Performance Max campaign needs checking, because the field it depends on is no longer supported for that campaign type.

Before and after, as the documentation describes it

The table sets what Google's Branded Searches documentation described before against what it describes after the change reported on 18 August 2026. Where the documentation does not state a prior value, the table says so rather than guessing one.

What Google did not say about the Branded Searches change

Google published no rollout date, no changelog entry and no explanation for the Branded Searches change reported on 18 August 2026, and the change was found in the help documentation itself rather than announced.

That absence has consequences an advertiser has to work around. Nobody outside Google can say on which date reported figures started being produced under the 7-day default, whether historical Branded Searches rows in Report Editor were recalculated under the new default or left as originally recorded, whether the window change and the Performance Max removal shipped together or weeks apart, or why either was done. Google stated none of it. Reading a specific effective date off the shape of a chart is guessing, and a guess told to a client in a monthly report becomes a fact in that client's head.

Google also did not publish any expected magnitude of change, so any statement that Branded Searches will fall by a given percentage is invented. The honest version is directional: the number can only stay flat or fall, because a shorter window is a strict subset of a longer one over the same period.

Analysis: rebaselining a comparison when the definition moved mid-flight

A year-on-year comparison assumes both numbers were produced by the same rule. After 18 August 2026 that assumption is broken for Branded Searches, so the comparison needs handling rather than presenting.

The cleanest option is to set the window explicitly. If a 30-day window is what the historical series was built on, an advertiser can set the Branded Searches attribution window back to 30 days and keep the definition stable going forward. Two honest caveats: the account is then running on a non-default setting that a future audit will query, and Google did not state whether changing the window recalculates already-reported history or only applies to data collected afterwards, so nobody should promise that past rows will line up.

The second option is to accept the break and mark it. Annotate the reporting series at the point the change was reported, state in the report that periods before and after were measured under different windows, and stop drawing a trend line across the join. This is less satisfying and more accurate. A chart with a visible break and a one-line explanation survives scrutiny; a smooth line across a definition change does not.

The third is to stop leaning on a single platform-defined number. If the only proof that upper-funnel spend works is a metric the platform can redefine without notice, the proof was always fragile. Pairing the in-platform figure with a measurement Google does not control, and watching whether the two move together, is what makes the next silent redefinition visible instead of invisible.

Analysis: what a Performance Max advertiser can look at instead

None of the following is equivalent to Branded Searches, and anyone presenting them as a replacement should say plainly that they are not. They are different instruments pointed at the same question.

• Brand Lift studies. These work by surveying an exposed group against a control group and reading the difference in recall or consideration. That is a designed experiment with a control, which is stronger evidence than an attributed count, but it is a study with its own eligibility and spend conditions, it returns a result for a period rather than a live daily column, and the Branded Searches documentation change says nothing about it.
• Search Console branded query impressions. Impressions and clicks on the brand term set are an organic-side signal of how many people went looking for the brand. They are not attributed to any campaign, so this is correlation read against a flighting calendar, not attribution. Its value is that Google Ads cannot redefine it, so a rise during a flight and a fall after it is evidence that stands on its own footing.
• Direct and branded session trends in analytics. Sessions arriving with no referrer, or through a branded organic query, tracked across flight and non-flight periods and split by geography, show the same underlying behaviour from a third angle. The weakness is well known: direct traffic absorbs a lot of unrelated noise, so it only reads cleanly when the flight is large relative to baseline and geographically contained.

The mechanism behind all three is the same one Branded Searches was measuring: a person sees an ad, does not click, and goes looking for the brand later. Losing one instrument for Performance Max does not mean the behaviour stopped, only that the platform stopped counting it in that campaign type.

What to check in your own Google Ads account

• Whether brand mapping is configured at all. Without it, Branded Searches reports nothing, and a zero here is a setup gap rather than a media result.
• The attribution window value on the Branded Searches conversion action itself. The setting is described by name in Google's documentation, and it is the value that determines what the column counts.
• Any recurring report, dashboard or data export that reads Branded Searches from a Performance Max campaign, since that campaign type is no longer supported.
• Whether a client KPI, incentive or budget case was written on Branded Searches volume. If it was, renegotiate the baseline before the number changes underneath it.
• The account's own historical lag between exposure and brand search, while data recorded under the longer window is still there to look at. That lag is what decides whether the shorter window costs a rounding error or a third of the metric.

What this means for Thai marketers

Thai advertisers lean on YouTube for awareness more heavily than many markets, and branded search lift is frequently the only number that justifies that line item to a Thai CFO. That makes the 7-day default a budget-defence problem before it is a measurement problem. A branded-search figure that falls in September because Google changed a default in August will be read in the meeting as a media failure unless somebody has already explained the change.

The 7-day window is a reasonable fit for the way Thai e-commerce actually runs, where a large share of spend is compressed into double-date sale bursts and the gap between seeing an ad and searching the brand is measured in days. It fits badly in property, insurance, private education and health, where a Thai buyer commonly sees a video ad, discusses it at home, and searches weeks later. Those categories should look hard at whether to set the window back to 30 days rather than accept the new default.

For Thai accounts running Performance Max as the main always-on structure, which is common where a team is small, the practical loss is a reporting argument. If the awareness case for Performance Max rested on Branded Searches, that case needs rebuilding on something else before the next quarterly review. Teams that run separate YouTube advertising campaigns keep the metric, which is itself an argument for not folding every objective into one campaign type.

FAQ: Branded Searches after the August 2026 change

Is the 30-day Branded Searches window gone?

No. Google changed the default to 7 days, and advertisers can still set the Branded Searches attribution window anywhere from 1 to 30 days, so a 30-day measurement remains available to anyone who configures it. The difference is that an account nobody adjusts will now report on 7 days instead of 30.

When did this take effect, and is it live in Thailand?

Google did not publish a rollout date, a changelog note or any regional detail, so the honest answer is that nobody outside Google knows the effective date. The change was found in Google's help documentation on 18 August 2026, and the documentation is global, but the source did not state a Thailand-specific timing or any staged rollout.

Do I have to do anything?

Only if Branded Searches is in your reporting, in which case check the window setting and warn whoever reads the number before it moves. Advertisers who never configured brand mapping have never been collecting this metric, and nothing changes for them; advertisers running Performance Max should stop expecting the metric from those campaigns.

Can I bid on Branded Searches?

No. Branded Searches is a primary conversion action for reporting purposes only, showing in the Results and All Conversions columns while remaining unavailable as a Smart Bidding optimisation target, and the August 2026 update did not change that.

Will my Branded Searches numbers drop?

They can only stay level or fall, because a 7-day window counts a subset of what a 30-day window counted over the same period, but Google published no expected magnitude and no honest figure can be given. How much an individual account loses depends on how long its own audience takes to move from seeing an ad to searching the brand.

If your upper-funnel reporting rests on Branded Searches, this is a good moment to check the window setting and the campaign types it still covers before the next report goes out. Our Google Ads team can review how the change lands in your account and what to rebaseline.