
As of November 2026, Google Ads API version v26.2 is a minor release with no breaking changes, published on 11 November 2026. That single fact settles most of the decision for a marketing lead: nothing in your live campaigns breaks, and there is no forced migration. The part that does deserve a reply to your sem agency is narrower and more useful. A default on video asset automation has moved, and reporting can now be asked for at account level instead of only at campaign level.
Google's release notes list four groups of changes for v26.2: assets, campaigns, reports and general items. Read the groups in that order, because they reach different people. Assets and campaigns change what the account can do; reports change what you can see; the general group changes what an integrator can build against.
The release date itself is the first thing to check. A minor version carries no breaking changes, so a well-run account keeps spending and reporting on the previous version while anyone integrating calls it a day for the upgrade. Google published v26.2 on 2026-11-11, three weeks after the v26 major release of 2026-10-14, which did carry breaking changes, and two weeks after the v26.1 minor on 2026-10-28. That major, minor, minor cadence is the rhythm to compare against, because a major version is the one that demands the audit.
| Item | v26 major | v26.1 | v26.2 |
|---|---|---|---|
| Release date | 2026-10-14 | 2026-10-28 | 2026-11-11 |
| Breaking changes | Yes | None | None |
| Action for a live account | Audit and re-test | Monitor | Check the asset default |
This is the one item in v26.2 that changes what a client sees in the ad, not just what an engineer sees in a schema.
Google added a new automation type that overlays text from associated text assets, such as headlines, on top of a video asset. For new Demand Gen video responsive ads created in v26, that setting now defaults to OPTED_IN. In earlier API versions, the same setting defaults to OPTED_OUT for new ads. An account that creates video ads through an API integration, a bulk upload tool or an internal automation is therefore going to produce ads with overlaid text unless someone deliberately turns it off. That is the first question to put to your sem agency.
Two practical consequences follow. The first is a creative check. If the brand rules say the video's own on-screen text must not be doubled by a headline overlay, the account needs an explicit opt-out written into the build script. Default-on is not a bug; it is the new starting point, and the person who owns the ad build has to know it changed.
The second is a reporting check. When the same setting is queried in API versions earlier than v26, the API returns an UNKNOWN value for the automation type. A dashboard that has not moved to the current version will show blank or unknown where a real status now exists, which looks like missing data rather than a version gap.
The reporting change matters most for accounts spending across several campaign types and trying to see overlap.
Unique reach metrics, the counts of users reached one or more times up to ten or more times, could previously be selected only with the campaign resource. In v26.2 they can also be selected with the customer resource, so a single query can return the account's deduplicated reach instead of summing campaign figures that overlap. Google also added companion state metrics that report whether a reach number is suppressed, so a zero can be read as a real zero rather than a missing value.
Segmentation is the second half of the change. These reach metrics can now be segmented by ad format type, ad sub-format type, ad network type and ad sub-network type. For a marketing lead, that turns "how many people did we reach" into "how many people did we reach on each network and in each format, without double counting between campaigns". It is the kind of question an sem agency singapore teams reported on could not answer from campaign-level sums alone.
Demand Gen campaigns now accept bidding seasonality adjustments in beta. The adjustment can be set at campaign scope, naming specific Demand Gen campaign resources, or at channel scope, naming Demand Gen as the advertising channel type.
The coverage limits matter more than the feature. The support applies to the YouTube, Display, Gmail and Maps inventory channels. Seasonality adjustments have no effect on the Discover channel. If a promotion or a peak trading period rides on Discover impressions, the adjustment will not lift bids there, and a plan that assumes otherwise will under-deliver at exactly the wrong moment. Write the channel list into the media plan before the peak, not after.
The general group is small and only matters to accounts using multi-party authorisation, which is common where an agency and a client's finance or legal team both approve changes.
Two new error codes arrive. One is returned when a request combines an operation requiring multi-party approval with an operation on an incompatible resource. The other is returned when a review is resolved against a resource that was modified or deleted after the review was created. Neither changes behaviour; both replace an opaque failure with a message naming the cause. For an account where approval is a recurring monthly chore, that is a real reduction in back-and-forth.
Use this sequence rather than reading the whole release note.
An account that answers "no" to the middle of that list has nothing to do with the release. That is the common case, and saying so plainly is more useful than turning a minor release into a project.
The release itself is a poor signal. A better signal is whether the agency knew it shipped, and how fast it can answer three questions about it.
Ask which API version the account's tooling runs against, and when it moves. Ask who owns the asset automation defaults for video ads, and whether they are set deliberately or inherited. Ask whether the reporting layer can query reach at customer level, and what it does with suppressed values. A team that maintains integrations as versioned engineering services, tested and deployed on a schedule, will answer all three from a ticket queue. One that treats the API as a one-off setup will answer from memory.
The same test applies to market cost. A Singapore SEM engagement typically starts from S$2,500 a month in management fees after a free first month, with a recommended media budget of S$3,000 to S$5,000 a month, and paid management on a 3-month minimum term after the free month. At that price, a version audit is not an extra line item a client should have to buy separately; it is part of what the retainer already pays for, and asking for it is the normal test of an sem agency.
Topkee Media's Hong Kong team works as an official Google Ads partner, and the Performance Google Ads service is built around the path from impression to click, landing page, enquiry, proposal and close rather than clicks alone. That framing is why a minor API release is filed as a maintenance item, not as news: the service measures the revenue end of the funnel, and the funnel only stays measurable if the data plumbing behind it stays current.
The plumbing is explicit in what Topkee builds. Custom development and API integration connect ads, landing pages, CRM and analytics into one connected system, so qualifying conversions are fed back to Google and Meta and smart bidding can optimise against revenue rather than clicks. Those integrations run as versioned engineering services, built and tested in edge containers, pushed through a build pipeline and deployed on Kubernetes, rather than as a one-off script that quietly expires. Topkee Cloud positions the same integration layer for SMEs that need their existing systems connected in days rather than months.
The asset side is the other half. Topkee's ad creative production groups ad groups by keyword and generates headlines, descriptions, images and extensions in volume with human selection and editing, which is exactly the kind of build where a changed default on video text overlays would otherwise slip through unnoticed. Add the creative refresh practice, which replaces ageing assets to keep response rates up, and the account has a standing reason to read every asset-level release note. The knowledge base holds no record of how any specific version is adopted, so treat this as what the service is built to do, not a claim about which version runs where.
Before the next monthly report, send your agency the five-step check above and ask for answers in writing. If you are still choosing an sem agency, use the same five questions in the pitch. They separate a team that maintains accounts from one that merely launches them. For a closer look at how a Singapore team runs paid search against a revenue target, see Google & Meta Ads Agency in Singapore and the Google Ads ROI guide.
Usually not. It carries no breaking changes, so live campaigns keep running, and a competent sem agency will say so rather than raise a work order. The one client-visible item is the new default for video text overlay automation on Demand Gen video ads, which only matters if the account creates those ads through an API or bulk tool.
It adds options rather than replacing anything. Unique reach metrics can now be queried on the customer resource as well as the campaign resource, and segmented by ad format, sub-format, network and sub-network. Existing queries keep working.
Only if the account runs Demand Gen, which you can confirm with your sem agency in a single message. The adjustment support covers YouTube, Display, Gmail and Maps, has no effect on Discover, and is in beta, so it is a planning input rather than a fix for a delivery problem.
Nothing directly, and that is the point. Quality Score is a keyword-level diagnostic scored from 1 to 10 and is not an input in the ad auction, so a version change cannot "raise" it. Use it to decide which ads, keywords and landing pages to fix first, and read version notes for the things versions actually change.
On the cadence shown by this cycle: a major version on 14 October 2026, a minor on 28 October 2026 and another minor on 11 November 2026. Majors are the ones that carry breaking changes and deserve an audit; minors deserve a check of the items that touch your build.