Two advertising API deadlines land within roughly 24 hours of each other next week.

Meta Marketing API v24.0 expires on 6 October 2026.

Google Ads API v22 sunsets on 7 October 2026.

Google is unusually clear about what happens after its deadline: starting on 7 October, v22 API requests begin to fail.

For European agencies and in-house performance teams, the risk is not confined to developers working directly with either API.

An old API version can sit underneath:

  • an internal reporting script;
  • a dashboard;
  • a data warehouse;
  • a bidding tool;
  • a spreadsheet connector;
  • an attribution pipeline;
  • a scheduled client report;
  • a third-party platform.

So the useful question is not:

“Did our developer upgrade the API?"

It is:

“Where does this version still exist inside our measurement stack?"

The deadlines

The immediate calendar is simple.

Platform Version Deadline Risk
Meta Marketing API v24.0 6 October 2026 Version expires
Google Ads API v22 7 October 2026 v22 API requests begin failing

Google published a September reminder telling developers to migrate before 7 October to avoid interrupted API access.

For Google Ads API users, that makes the deadline operational rather than theoretical.

A request still calling v22 after the sunset is not merely “using an older version.”

It can fail.

That can break processes several layers away from the person who originally wrote the integration.

Start with dependencies, not code

Most teams should not begin this audit by opening a code repository.

Begin by listing every place advertising data moves.

For example:

Google Ads / Meta → connector → warehouse → transformation → dashboard → scheduled report → client or internal team

Then identify which component owns the API connection.

Sometimes the answer is obvious.

A custom Python script may visibly specify the API version.

In other cases, a SaaS connector hides the implementation. The agency may have no direct control over the version and must verify the vendor’s migration status instead.

That difference determines the action.

Check internal scripts first

Search internal repositories and automation jobs for explicit version references.

For Google Ads API, Google recommends using the Google Cloud Console to inspect recently called methods. Method names include the API version, which makes it possible to spot calls still using v22.

Do not inspect only production applications.

Also check:

  • cron jobs;
  • notebooks;
  • archived reporting scripts still scheduled;
  • serverless functions;
  • spreadsheet automations;
  • ETL jobs;
  • one-off utilities that became permanent.

The dangerous integration is often not the flagship data pipeline.

It is the small script everyone forgot was still running.

Then audit connectors and vendors

Third-party tools create a different problem.

Your team may never see an API version number.

For every connector, record:

  • vendor;
  • platform;
  • data source;
  • destination;
  • owner;
  • last successful sync;
  • migration status;
  • vendor confirmation;
  • fallback.

This applies to tools feeding:

  • BigQuery;
  • Snowflake;
  • Looker;
  • Power BI;
  • Tableau;
  • Google Sheets;
  • client dashboards;
  • attribution platforms;
  • internal BI systems.

Do not assume a working sync today proves it will continue working after the sunset.

The entire purpose of a sunset deadline is that the old version can work until it suddenly does not.

Build a five-minute failure map

For every affected connection, ask:

If this stops on Monday morning, what breaks?

Classify the answer.

Low impact

A non-essential internal report misses one refresh.

Medium impact

A dashboard becomes stale or an automated weekly report fails.

High impact

Campaign automation, bidding logic, budget controls, conversion ingestion or client-critical reporting is interrupted.

This determines migration priority better than the number of integrations.

One forgotten high-impact script deserves attention before ten low-value dashboards.

The NEMO API migration audit

Use this sheet:

Field Record
Platform Meta / Google
API version Current version in use
Connection Script, connector, app or vendor
Data flow Source → destination
Business owner Who notices if it fails?
Technical owner Who can fix it?
Last successful run Timestamp
Deadline 6 Oct / 7 Oct
Upgrade complete Yes / No
Test completed Yes / No
Fallback Manual export / alternate connector / none
Risk Low / Medium / High

The goal is to turn a platform deprecation notice into an operational inventory.

Test the output, not just the connection

A successful API response does not prove the migration is complete.

After moving to a newer version, compare:

  • row counts;
  • campaign IDs;
  • spend;
  • impressions;
  • clicks;
  • conversions;
  • date ranges;
  • currency;
  • attribution fields;
  • dimensions used by downstream reports.

Version migrations can change fields and behaviour even when authentication and requests still succeed.

For reporting teams, a silent schema or field change can be worse than an obvious error.

An obvious failure gets investigated.

A dashboard that continues refreshing with incomplete data can survive for days.

Meta has another October measurement change coming

The 6 October v24 expiration is not the only Meta API date worth recording.

Meta’s v26 changelog also lists delivery-estimate field removals scheduled for 27 October 2026, including fields such as:

  • daily_outcomes_curve;
  • budget_guardrail;
  • estimate_dau.

That is a separate change and should not be confused with the v24 expiration.

But if the team is already auditing Meta dependencies this week, add 27 October to the same engineering calendar.

One inventory can cover both.

Do not wait for a dashboard to go blank

API sunsets are easy to ignore because they do not create a visible problem until the deadline arrives.

That is exactly why they deserve a short operational audit before the date.

For agencies, the cost is not only technical downtime.

It can become:

  • late client reporting;
  • incorrect pacing information;
  • missing warehouse data;
  • broken automation;
  • wasted analyst time;
  • emergency vendor tickets.

The preventative work is usually cheaper.

Before 6–7 October, identify the connections, confirm the version, test the migration and document the fallback.

That is enough to turn next week’s deadlines from an incident into routine maintenance.

Primary sources

Meta Marketing API version schedule: https://developers.facebook.com/docs/marketing-api/marketing-api-changelog/versions/

Meta Marketing API v26 changes: https://developers.facebook.com/documentation/ads-commerce/marketing-api/marketing-api-changelog/version26.0

Google Ads API v22 sunset reminder: https://ads-developers.googleblog.com/2026/09/google-ads-api-v22-sunset-reminder.html

API schedules can change. Confirm the current platform documentation before performing production migrations.