Most platform deadlines are soft. Something gets deprecated, warnings appear in the response headers, and the date slips twice before anyone loses a night’s sleep. This one is not that.
Apple has set 26 January 2027 as the sunset for the Apple Ads Campaign Management API v5. Its guidance, quoted in the April 2026 announcement coverage, is direct: the v5 API will be sunset on that date, and migration to the Platform API is recommended as early as possible to avoid any disruptions and take advantage of enhanced functionality. Trade reporting of the announcement is consistent on the consequence — after the date the old interface stops working entirely, with no campaign creation, no campaign management and no reporting through it.
Five months on from that announcement, the replacement is no longer a preview. Apple’s developer documentation now lists the Apple Ads Platform API at version 1.0.0, with a full published endpoint surface, OAuth 2 authentication, client libraries, workflow guides for both App Store and Apple Maps campaigns, and a changelog. The thing you have to migrate to exists. The clock that started in April is now visibly shorter than the work.
01The date, and what “sunset” means here
26 January 2027. One date, no phases, no published grace period.
The distinction worth internalising is between deprecation and sunset. A deprecated API still answers; it just warns you that it would rather not. A sunset API is switched off. Coverage of Apple’s announcement quotes the consequence as the API no longer supporting campaign creation, management or reporting functions — and separately notes that after the deadline the existing API stops functioning entirely, meaning campaigns can no longer be managed and reporting can no longer be pulled through it.
For a bidding script, that failure mode is worse than it sounds. Automation that cannot write does not stop spending; it stops adjusting. Budgets keep pacing at whatever they were last set to, bids stay frozen at their last written value, negative keywords stop being added, and the pipeline that would have told you goes quiet at the same moment. An account on Maximize Conversions with a daily budget will happily keep spending up to the 30.4× monthly ceiling while your control plane is dark.
→ THE FAILURE MODE TO PLAN AROUND
Silence, not stoppage
A dead write path does not pause your campaigns. It freezes them in their last configured state while removing your ability to see or change that state. Any migration plan that ends on 25 January is a plan whose failure mode is an unmonitored live budget.
02What the Platform API actually is
Apple describes the Platform API as a way to programmatically manage App Store and Apple Maps advertising campaigns through a single RESTful interface — building campaigns with budgets, creating ad groups with keywords, audience refinement criteria and scheduling, creating and managing ads and creative assets, configuring targeting and bidding, and monitoring performance through reports, insights and recommendations.
Authentication is OAuth 2. Apple’s help documentation sets out the access model: an Account Admin assigns an API user one of four roles — API Account Manager, API Account Read Only, Limited Access API Read & Write, or API Read Only — and organisations can grant a service provider access through Sign in with Apple without sharing credentials, with authorisation scoped and revocable by an Account Admin or Campaign Group Manager.
That access model is worth flagging early because it is the part of a migration that is not an engineering task. Provisioning roles and credentials sits with whoever administers the ad account, which in most organisations is not the person writing the client. It is the step that reliably adds two weeks nobody planned for.
On timing: the April coverage of the sunset announcement did not carry a public availability date for the replacement, and reported only that preview documentation had been published. The documentation now shows version 1.0.0 with a live endpoint reference and a changelog, so the correct present-tense statement is that the Platform API is available and versioned, and that anyone planning around “it isn’t out yet” is working from stale information.
03The endpoint surface, read closely
Reading an API’s table of contents is an underrated way to understand what a platform intends. Here is what Apple has grouped under the Platform API, and what each group signals.
| Group | What it covers | What it signals |
|---|
| Account management, Org & Me | Identity, organisation details, ad account access lists, role-based access | Multi-account and agency structures are first-class, not bolted on |
| Campaigns, ad groups, keywords, negative keywords | Full CRUD plus bulk keyword operations | The v5 core, carried forward — this is the part that ports most directly |
| Creatives, assets, product pages | Creative CRUD, Maps asset upload, product page retrieval with localised detail | Creative is now programmable, which the old campaign API did not meaningfully expose |
| App eligibility | Whether apps qualify to run ads; creative rejection reasons | Rejection handling becomes machine-readable — meaningful under the 4+ asset policy |
| Brands, locations, location groups | Maps-side brand identity and location targeting | The unification with Maps is structural, not a marketing line |
| Reports (Apps and Brands) | Campaign, ad group, ad, keyword and search-term performance for each surface | Two parallel report families under one auth — the thing that makes unification real |
| Insights, recommendations, suggestions | Impression share, search-term popularity, budget and Target CPA recommendations, keyword and category suggestions | Apple’s own optimisation surfaces are now callable rather than UI-only |
| Budget orders, change history | Budget order lifecycle; audit summaries and change detail records | Governance features aimed at teams large enough to need an audit trail |
| Rate limits, client libraries | Quota pacing guidance; Apple-maintained clients | Apple expects sustained programmatic load and is shaping how you apply it |
Two observations. First, this is broader than the object model most teams built against in v5 — creative, eligibility, insights and change history are not a rename of anything, they are new territory. Anyone budgeting this as a find-and-replace on a base URL has mis-scoped it.
Second, the presence of impression-share and search-term-popularity endpoints alongside recommendations is the most interesting line in the table for anyone running a serious account. Those were the Insights capabilities Apple shipped to the web interface in March 2026; having them behind an API is what turns share-of-voice from a screenshot in a monthly deck into a series you can trend and alert on.
04Who is exposed — including the risk you inherited
If your Apple Ads account is managed entirely through the web interface, the sunset does not touch you. Almost nobody spending meaningfully is in that position.
The exposed list, roughly in order of how quietly each one fails:
- Warehouse reporting jobs. The scheduled pull that lands Apple Ads cost, taps and installs next to your revenue tables. Fails silently into a gap in a dashboard nobody checks over a weekend.
- Bid and budget automation. In-house scripts that adjust CPT bids or pace daily budgets. Fails into frozen configuration and unmonitored spend.
- Keyword pipelines. Search-term harvesting into new keywords, and the negative-keyword hygiene loop that keeps brand and category spend from bleeding into each other.
- Agency and client dashboards. Reporting layers built once, maintained rarely, and often by someone who has since left.
- Third-party bid management and MMP integrations. The inherited risk, and the one worth a specific email this week.
That last category deserves emphasis. A vendor still on the old API is your outage, not theirs, and you cannot see their migration status from inside your account. The remedy is unglamorous: write to every vendor holding an Apple Ads credential, ask for a dated commitment on Platform API support, and record the answer. Vendors who have migrated will answer in a day. The ones who don’t answer are your actual risk register.
05The Maps half, and whether app advertisers should care
The Platform API’s documentation carries a dedicated workflow for advertising a business on Apple Maps: identify your brand, upload creative assets, build location groups, then create campaigns, ad groups, creatives and ads. There are separate Brands report endpoints, Brands data objects, and Maps-specific enumerations.
Timing lines up. Apple opened Maps ad buying to businesses in the United States and Canada on 14 August 2026, with delivery described only as reaching users in the coming weeks. The launch carried a limited-time promotion — 15% back as a monthly statement credit for up to a year, capped at $1,000 per month, applied to subsequent spend, for eligible businesses booking by credit card by 11 October 2026. Two placements: Suggested Places on the search screen before a query, and search results when relevant to one. Apple cites over a billion relevant business searches per month in Maps across the two countries.
Derived figure: a $1,000 monthly cap on a 15% rebate means the full percentage only applies up to roughly $6,667 of monthly spend, above which the effective rate declines — and the theoretical maximum value across twelve months is $12,000. That arithmetic is ours, not Apple’s; Apple publishes the cap and the percentage, not the implied break-even.
For most readers of this article, none of that is addressable. Maps ads are sold to businesses with a physical location and an existing listing; an app with no storefront cannot buy them, there is no keyword bidding, and there is no product page or install attribution surface — the argument we made in full when the format was still pre-launch in the Maps ads analysis, and the launch has not changed it.
So why does the unification matter? Because it tells you what Apple is building. A single control plane, a single auth model and parallel report families across two surfaces is the architecture of an ad platform that expects to add more surfaces — Apple News premium sponsorships already exist, and Search Engine Land reported in December 2025 that Apple planned to add inline App Store search ad inventory mixed into organic results during 2026, with placement determined primarily by relevance. Data vintage: that inline-inventory report is from December 2025 and describes a plan, not a shipped state. The API you migrate to in 2026 is the one those surfaces will land in.
→ FREE 10-POINT AUDIT
Find out what breaks in your account on 26 January
We map every automation, integration and reporting job touching your Apple Ads account, flag what’s still on the retiring API, and hand you a sequenced migration plan. One hour. Senior strategist. No pitch.
Book my free audit →
06The migration sequence that doesn’t break spend
The order below is chosen so that every step before the last one is reversible and none of them can move money by accident.
1. Inventory every credential, not every script
Start from the ad account, not the codebase. List every system holding an Apple Ads credential — including vendors — and record for each whether it reads, writes, or both, and who owns it. Teams consistently find one integration nobody remembered. Finding it in September is a task; finding it in January is an incident.
2. Provision OAuth 2 access early
Have your Account Admin create the API users and assign roles now. This is a queue problem, not a technical one, and it blocks everything downstream. Where a vendor needs access, use the Sign in with Apple delegation path rather than sharing credentials, and confirm you know how to revoke it.
3. Migrate read paths first, and run both in parallel
Reporting is the safe half: you can pull the same date range from both APIs and diff them without touching a live campaign. Do this for a full reporting cycle before you trust anything. Expect small discrepancies at the edges of attribution windows and decide in advance what tolerance counts as a match — that decision is easier to make calmly in October than at 2am in January.
4. Port write paths one entity at a time
Campaigns, then ad groups, then keywords and negative keywords, then creatives. Keep the old integration live but idle behind a flag until each entity type has produced identical results across a full cycle. Bulk keyword operations are their own endpoint group and deserve their own test pass; a bulk write that half-applies is the worst outcome in this entire migration.
5. Cut over with a buffer, and not in peak season
Pick a cutover date months ahead of 26 January, not weeks. Q4 is the wrong time for anyone with seasonal spend, which argues for landing this in autumn and leaving the holidays clear. Then keep the old path decommissioned but documented, so a rollback is a decision rather than an excavation.
6. Re-point your alerting before you decommission
The failure mode from section 01 is silence. Whatever alerts you on stale data, missing rows or budget drift must be watching the new path before the old one goes away — otherwise the migration removes your automation’s ability to fail loudly at precisely the moment it is most likely to fail.
07Why automation earns more on this channel than any other
It is fair to ask why a migration this involved is worth doing well, rather than doing minimally. The answer is about what a marginal unit of control is worth on this particular surface.
Apple Ads is where you buy declared intent. A user types a query into App Store search and taps a result; the intent is stated rather than inferred, and the auction sits at the point of decision rather than upstream of it. That is also why the audience is the highest-value one in mobile — the App Store’s consumer-spend-per-download runs several times Google Play’s on 2025 aggregates, a gap we derived in full in the ARPU piece. Derived from published store aggregates rather than reported directly; the direction is robust across sources, the exact multiple is not.
Two consequences. First, precision compounds faster here: on a channel where a single keyword can carry a materially different downstream value than its neighbour, the ability to act on search-term data hourly instead of weekly is worth more than the same ability on a broad-targeting channel. Second, the mistakes are more expensive, which is the whole thesis of the expensive-user paradox — the users worth most are the ones you are most likely to overpay for, and overpaying at scale is exactly what an unattended automation does best.
The Platform API’s insights, recommendations and change-history groups matter in that light. Impression share and search-term popularity as callable series let you see competitive pressure moving before it shows up in cost. Change history gives you a machine-readable record of what your own automation did, which is the prerequisite for ever diagnosing it. Those are not new features you can ignore until 2027 — they are the reason to treat this as a rebuild rather than a port, and to stop bidding against the wrong signal while you are in there, per the case against CPI.
08What Apple hasn’t said
The honest inventory of what is not knowable from public sources today:
- No published feature-freeze date for v5. The sunset date is fixed; whether the old API stops receiving new capabilities before then is not stated in the coverage. Assume new platform features land on the new API only.
- No public parity matrix. Apple has published preview documentation described as showing how requests and responses differ, but there is no authoritative list of v5 capabilities with no Platform API equivalent. If your integration depends on something unusual, you have to test for it rather than look it up.
- No stated migration tooling beyond documentation and client libraries. There is no announced bulk-migration utility or credential-transfer path. Plan for hand-written work.
- Third-party readiness is opaque. No vendor-by-vendor public register of who supports the Platform API exists. You have to ask, one vendor at a time.
- Maps ads have no published pricing model, auction mechanics or benchmarks. Buying opened in August 2026; the pricing model, ranking rules and reporting metrics were not disclosed at launch. Any Maps ads CPC benchmark you encounter right now is not sourced from Apple.
- Deadlines have moved before, but plan as if this one won’t. Nothing in Apple’s language suggests flexibility, and a migration finished early costs you nothing.
09The takeaway
There is a fixed date, the replacement is already shipping at 1.0.0, and the replacement is a broader system than the one it retires. Those three facts together mean the cheap version of this project — swap the base URL, ship it, move on — is available to almost nobody.
The teams that come out of this ahead will treat it as the rebuild it actually is: audit what touches the account, migrate reads before writes, get the credentials provisioned before anyone writes a line of client code, and use the new insight and change-history surfaces rather than porting a five-year-old script into a newer wrapper. The teams that come out behind will discover in the third week of January that a vendor they never thought about is the thing standing between them and a frozen budget.
A dead write path doesn’t stop your spend. It stops your ability to change it — and takes your visibility with it.
Five months of comfortable work, or three weeks of uncomfortable work. The date is the same either way. And while you are in there, it is worth reading the other fixed-deadline story of this month — different regulator, same lesson about lead time.