Google Ads API v20 Sunsets June 10 and Your Bid Manager Probably Hasn't Migrated

Google Ads API v20 sunsets on June 10, 2026, after which every request hitting that version returns a hard failure. That gives agencies roughly 40 days to identify which third-party bid managers, ad servers, attribution platforms, and reporting scripts still call v20, and to confirm vendor migrations to v21 or v23.1. The fix is a vendor audit, not a piece of code you ship.

A faster sunset cycle nobody really planned for

This is the first v20-style cutoff under Google's new monthly release cadence, which kicked in this past January. The trade Google made was simple: more major versions per year, but each one only gets about 12 months of support after launch. v20 dropped June 4, 2025. v21 followed. v22 launched in October 2025, and v23 came out in early 2026. The PPC Land breakdown of the new cadence laid out the staggered sunset wall: v19 in February, v20 in June, v21 in August, v22 in October. That is four distinct migrations in nine months, each requiring vendor coordination on most agency stacks.

What changed for agencies is the assumption that "we'll deal with it next quarter" no longer fits the calendar. Under the old three-versions-a-year cycle, you could be one or two majors behind and still have 18 months to move. The runway now is closer to 6, sometimes less depending on when your vendor adopted the latest major.

The audit nobody at the agency is running

Search Engine Land's reporting on the sunset flagged the right thing: when v20 stops responding, automated workflows including reporting, bidding, and campaign management stop with it. The wrinkle is that most working PPC managers have no idea which API version their tools actually call. You log into Skai, or Adalysis, or your in-house ETL job pulling into BigQuery, and there is no version number anywhere on the dashboard.

The audit looks like this:

1. List every product, script, or vendor that touches Google Ads through any kind of automation. Bid managers, reporting platforms, scheduled exports, automated rules outside the Google Ads UI, custom Looker Studio data sources, anything pulling from BigQuery Data Transfer.
2. Email or ticket each vendor with one question: "What Google Ads API version is your platform currently calling, and when is your migration off v20 scheduled to complete?" Get the answer in writing.
3. For in-house code, grep your repo for v20 or check the client library version pinned in your package.json, requirements.txt, or whatever your stack uses.

Most agencies, I'd expect, have a decent answer for the bid manager and zero answer for everything else. The reporting layer is where surprises live. From what I've seen, an agency's third or fourth vendor in the dependency chain (a Looker Studio connector, say) is the one that quietly breaks when an API sunsets, and the symptom is usually data that just stops updating without an obvious error message.

What "migrate" actually means for you

For most agency stacks, you are not the one migrating. Your vendor is. Your job is to confirm the migration shipped before June 10 and to verify it on a live account afterward.

If you do maintain in-house integration code, the destination is v23.1, the version Google released this past February. ALM Corp's v23.1 walkthrough is the most thorough public breakdown, and it includes a backport situation worth knowing about. Google issued v20.2, v21.1, and v22.1 simultaneously with v23.1 to push the EU political advertising compliance field (contains_eu_political_advertising) to every still-supported version. So if you are pinned to v20 specifically, you can move to v20.2 to handle that compliance field, but it does not buy you any extra runway. The v20.x branch still dies June 10.

The realistic upgrade target if you skipped a generation is v22 or v23.1. Google's v22 announcement added the AssetGenerationService, new App campaign bidding goals, Demand Gen TargetCPC, and PMax asset automation expansions, so jumping straight there picks up the recent feature surface as well.

The Wordstream precedent, briefly

When Google retired the older AdWords API in April 2022, the migration cost was steep enough that some vendors chose not to make it. Optmyzr's post-mortem on that period flagged Wordstream specifically: parent company Gannett shut down the self-service Advisor product and pushed customers into managed agency services rather than rebuild on the new API. The lesson is not that Wordstream-scale shutdowns are coming for v20. This is a much smaller migration than the AdWords-to-Ads jump. The lesson is that vendor decisions about whether to keep up are real, and the smaller the vendor, the more likely you find out late.

If a tool you depend on has not announced a v23.1-ready release yet, treat that as a product risk, not just an engineering one. Ask whether their monthly release-cycle support is documented somewhere, not just assumed.

The 40-day shape of this

Working backwards from June 10, this is roughly the calendar I'd suggest:

• Now through May 8: Vendor audit. List, ask, get written answers. If a vendor cannot confirm a v23.1 (or v22) release before June 1, escalate.
• May 8 through May 22: Test in a staging or low-spend account once each vendor ships their migration. Pull a daily report manually and diff it against your usual export.
• May 22 through June 1: Sign off. Document which version each tool now calls so the next sunset (v21 in August) is a 10-minute check, not another scramble.
• June 1 through June 10: Buffer. If anything breaks on June 10, you want at least a week of awareness time to patch.

The only piece of this that needs real engineering is the in-house code path. Everything else is procurement-grade work. It is not glamorous and it is exactly the kind of thing that gets pushed because nothing visibly breaks until something visibly breaks.

We covered a related shift earlier this spring when Google made MFA mandatory across the Ads API on April 21, which surfaced a similar pattern: agencies that hadn't checked their integrations got a surprise lockout. The v20 sunset is a slower-moving version of the same gap, with more warning and equally low odds of being on anyone's calendar.

Where the next 40 days probably go wrong

The most common failure mode is not the bid manager. The bid manager has a roadmap, a status page, and usually a release note within a week of any major API change. The failure mode is the small custom script someone wrote in 2023, scheduled in cron, never documented, that pulls a daily spend export into a finance system. Nobody owns it now. Nobody flags it. It returns blank rows on June 11, and someone notices in mid-July when the budget reconciliation does not match.

The audit catches that. The audit only happens if someone actually runs it. If June 10 ends up costing you a week of finance backfill, it will not be because Google did anything surprising. It will be because the calendar said 40 days and 40 days felt like plenty.

Notice Me Senpai Editorial