To switch mobile app marketing agencies without losing momentum, do three things in order. First, before you give notice, confirm that you own the Meta business portfolio, the ad account, the app ID, the dataset, the measurement partner account and the raw creative files, because some of these cannot be moved later. Second, use the notice period as an overlap: the outgoing agency keeps the account running untouched while the incoming agency audits measurement and freezes a baseline. Third, the new agency should launch new concepts alongside your current winners rather than rebuild the account in week one. Then judge it on the right clock. For a subscription app, the revenue you see for about six weeks after the handover is still mostly the old agency's work.
Most of the momentum in an app account is not in the agency. It is in your app's event history, the ad sets that have stopped learning and are delivering steadily, the winning ads and the engagement on them, the record of what has been tested, and a pipeline of new creative. A switch goes badly when one of those is lost or reset by accident. This guide covers how to keep each of them, and it picks up where our questions to ask a Meta Ads agency leave off: those are for before you sign, this is for when you are leaving one and starting with another.
Key takeaways
- Check ownership before you give notice. Meta says an ad account can only be in one business portfolio, and that claiming one may move it permanently.
- The Meta app ID matters more than the ad account. It is built into your app, and your app events are recorded against it.
- Use the notice period as an overlap. The outgoing agency keeps the account steady; the incoming agency audits and builds a baseline.
- A rebuild in week one resets learning across the account. New concepts should launch beside the winners, not instead of them.
- Subscription revenue lags the handover by weeks. Do not judge either agency on cohorts the other one acquired.
What actually carries momentum in a Meta app account
It helps to be specific about what can be lost, because founders usually worry about the wrong thing. The fear is that a new agency means Meta "forgets" the account. In practice there are five separate things, and each one breaks in a different way.
- Event history. Your installs, trial starts and purchases are recorded against your Meta app ID and any dataset you use. If those belong to you, the history stays with you whoever manages the ads.
- Stable delivery. Ad sets that have left the learning phase deliver more predictably. This is lost through edits, not through a change of agency.
- Winning ads. The specific ads that work, and the likes and comments on them, which sit on your Page's posts if the Page is yours.
- The learning record. Which angles, audiences and formats were tested, and what won. This usually lives in the outgoing agency's spreadsheets and is the easiest thing to lose.
- The creative pipeline. Concepts in production. The outgoing agency stops making them when it gets notice, and the new one needs about three weeks to get its first wave live.
Changing who manages a Meta account resets nothing by itself. What resets momentum is lost ownership, lost files, lost records and an incoming team that rebuilds everything at once.
Step 1: map who owns what before you give notice
Do this quietly, before the conversation with your current agency, because the order of operations matters and some of it is hard to undo. Meta's Business Help Centre page on adding an ad account to a business portfolio is explicit: "An ad account can only be in one business portfolio." It also warns that adding an existing ad account to a portfolio "may permanently move it into the portfolio", after which it cannot be transferred to another one. If your agency created the ad account inside its own portfolio, there is usually no route to move it to yours. The agency can share it with you, but it stays theirs.
| Asset | Who should own it | How to check | If the agency owns it |
|---|---|---|---|
| Meta business portfolio | You | Business settings: whose name is on the portfolio, and who has full control. | Start a portfolio of your own now. Everything else depends on having one. |
| Ad account | Your portfolio, agency as partner | Business settings, Accounts, Ad accounts. Owned or shared? | Keep partner access through the handover, then run from a new ad account in your portfolio. |
| Meta app ID (your app events) | Your portfolio | Business settings, Accounts, Apps. Is the app owned or only shared? | The most serious gap. Request an ownership transfer before anything else, because the app ID is built into your app. |
| Dataset or pixel (web-to-app and site events) | Your portfolio | Events Manager: which portfolio owns the dataset, and which ad accounts use it. | Ask for a transfer, or run a dataset of your own in parallel before switching. |
| Facebook Page and Instagram account | Your portfolio | Business settings, Accounts, Pages and Instagram accounts. | Fix before notice. Ads run from these identities, and the posts carry your social proof. |
| Mobile measurement partner | You, as the app owner | Whose company is the account holder, and who are the admins? | Move the account to your company or open your own and plan a clean cut-over date. |
| App Store Connect, Google Play Console, RevenueCat | You | Agency should hold user roles only, never the account. | Rare, but fix immediately. These hold your revenue and your app. |
| Raw creative files and creator licences | You, by contract | Your contract's IP clause, and each creator licence's term and permitted use. | Negotiate delivery during the notice period. Without the raw files, iteration starts from zero. |
Two rows deserve more attention than the others.
The app ID is the one that really hurts
An ad account owned by the agency is an inconvenience: you open a new one in your own portfolio and carry on. A Meta app ID owned by the agency is worse, because the ID is configured inside your app's SDK setup and your app events are recorded against it. Replacing it means an app release and a fresh event history. Meta's developer documentation on transferring app ownership sets out how an app can be claimed by a business portfolio, and notes that once the claim is approved "this action cannot be undone". That is a reason to make sure it lands in the right portfolio, yours, and to do it while the relationship is still friendly. Meta's help page on adding apps to a business portfolio also covers the case where an agency owns the app and shares it with you, which is the arrangement to get out of.
Your measurement partner should be in your name
Measurement partners distinguish between the app owner and the agency. AppsFlyer's documentation on agency accounts describes them as measuring campaigns run "on behalf of marketer clients (app owners)", with the marketer deciding which additional data the agency can see. That is the healthy arrangement: your company holds the account, and each agency is a guest in it. If your previous agency opened the account itself, your attribution history and your iOS conversion value setup may sit in an account you do not control. Fix that before the switch, not during it.
If the old agency owns the ad account
This is common, and it is survivable. The sequence we would follow:
- Ask for partner access to the ad account to continue through the handover, so the incoming agency can read the history in place.
- Export everything while you can: performance by ad, ad set, placement and country for at least 12 months, with the ad names intact.
- Open a new ad account in your own portfolio and connect your app and dataset to it.
- If the Page is yours, rebuild the best ads from their existing posts where you can, so the engagement they have built up carries across.
- Re-upload customer-list audiences from your own data. They were created inside the old ad account, so treat them as staying there.
The new ad account starts with new ad sets, which all go through the learning phase. That is a cost, but a bounded one, and it is a cost you pay once. Staying inside an account you do not own means paying it again at the next switch.
Step 2: plan the overlap around two clocks
Every agency switch runs on two clocks at once, and most handover plans only look at the first.
The first is the notice period. It sets how long the outgoing agency stays responsible, and it is the window in which the incoming agency can prepare without yet touching anything. Our own terms on the pricing page are a typical example: an initial three-month commitment, then 30 days' notice. Use all of it. If access is granted in the first week, 30 days is enough for a proper audit and a baseline.
The second is the subscription cohort clock, which runs well past the notice period. RevenueCat's State of Subscription Apps 2026 finds that 19.2% of paid conversions happen in week six or later, so a cohort is only about four-fifths converted once it is five weeks old. On iOS, Apple's conversion windows run to day 35 after first launch, and the final postback can arrive up to 144 hours later, around day 41. We worked through what that means for a new engagement in our guide to the first 90 days with an app marketing agency.
Put the two together and a handover looks like this. The table assumes 30 days' notice and a first wave of new concepts live about three weeks after the handover, which is a realistic pace.
| When | Outgoing agency | Incoming agency | You |
|---|---|---|---|
| Before notice | Business as usual. | Not yet involved, or in sales conversations only. | Map ownership using the table above. Fix anything you can fix without the agency's help. |
| Notice period, week 1 | Keeps running the account. Asked for exports, files and the test log. | Granted read access. Starts the access and measurement audit. | Give notice in writing with a list of handover items and a date for each. |
| Notice period, weeks 2 to 4 | No restructures, no new campaign architecture. Final agreed creative batch delivered. | Freezes the baseline, reads the test log, briefs its first concepts. | Hold the paywall, pricing and onboarding steady. |
| Handover day | Access removed, last of all. | Takes edit access. Current winners left running. | Rotate any tokens the old agency created. Confirm events still arrive. |
| About 3 weeks after | Its last cohorts are still converting. | First new concepts live, read on trial starts. | Judge speed, process and the audit, not revenue. |
| About 6 to 9 weeks after | Last iOS postbacks for its final installs have arrived by about week six. | First cohorts of its own concepts become readable on paid conversion. | First fair revenue comparison against the frozen baseline. |
For about six weeks after a handover, the subscription revenue you see comes mostly from installs the outgoing agency bought. The first fair revenue comparison of the new agency arrives roughly two months after it takes over.
This cuts both ways. A strong month straight after a switch is usually the old agency's cohorts maturing, not the new agency's genius. A weak one may be the old agency's final weeks. Neither should drive a decision about the new team. If you want to see the difference between a cohort read and a blended one, our subscription app ROAS benchmarks explain how ROAS builds across D7, D30 and D60 for a single cohort.
Step 3: freeze a baseline before the handover
The incoming agency should record where the funnel stands before it changes anything: cost per install, install-to-trial, cost per trial, trial-to-paid and early revenue per install, split by platform and country, for the eight to twelve weeks before it arrived. Write it down and agree it with them. Without it, every later conversation about whether the switch worked becomes an argument about memory.
Freeze the product too. If the paywall, the price and the onboarding all change in the same month as the agency, nobody will know which change did what. Log every product change with a date for the first two months. Our guide to CPI, CPA and CAC for subscription apps sets out which of those numbers to treat as the headline.
Step 4: do not rebuild the account in week one
New teams like to restructure. Sometimes the structure genuinely is wrong, but a rebuild on day one is the single fastest way to turn a change of agency into a loss of momentum. Meta's Business Help Centre on the learning phase says the delivery system re-learns when you "create a new ad or ad set or make a significant edit to an existing one", that ad sets usually leave learning "after about 50 results in the week after the ad set's last significant edit", and that frequent budget changes can send an ad set back into learning. A full rebuild puts every ad set back to the start at once, at the moment nobody yet has a baseline to compare against.
The safer pattern is to leave the ad sets that carry most of the spend running untouched for the first few weeks, and launch the incoming agency's concepts alongside them. Structural changes can follow once the new concepts have a read, one change at a time and with a written reason. We cover the mechanics in our guide to the Meta Ads learning phase.
Do not change measurement in the same month
The same logic applies with more force to tracking. Moving to a new measurement partner, changing your iOS conversion value mapping or switching the optimisation event in the same month as the agency switch makes the data impossible to read. In Ads Manager a change in measurement looks exactly like a change in performance. If the audit finds tracking that genuinely needs fixing, fix it, but date it, tell everyone, and expect the numbers either side of the fix not to be comparable. Our post on SKAdNetwork postback delays explains why iOS results move around even when nothing has changed.
Step 5: keep the creative pipeline full
The gap that hurts most in practice is creative. The outgoing agency has little reason to brief new work once it has notice, and the incoming agency needs research and production time before its first concepts are live. Left alone, that is six or seven weeks with nothing new entering the account, which is long enough for your best ads to tire. Our guide to creative fatigue in mobile app campaigns explains how quickly that can happen at scale.
Three things close the gap:
- Agree a final batch in the notice terms. If you are paying for the notice month, ask for a defined set of finished ads to be delivered within it.
- Get the raw files, not just the exports. Project files and unedited creator footage let the new team make variations of proven winners in days, which bridges the gap until its own concepts land.
- Check every creator licence. UGC usage rights are usually time-limited and sometimes tied to the agency that negotiated them. Know which of your live ads lose their rights in the next three months. Our guide to UGC agency pricing covers how usage rights are priced.
The most valuable single document in a handover is the test log: which concepts ran, what each was meant to prove, and what happened, losers included. It stops the new agency spending its first two months re-testing ideas you already paid to rule out. If the old agency never kept one, the new agency should reconstruct a rough version from the ad-level exports and map it against the angles and customer types already covered, which is how we use the TSO Creative Framework on a new account.
The access handover checklist
Access changes are where accounts break by accident. The rule is simple: add and verify the new access first, remove the old access last.
- Grant the incoming agency partner access to your portfolio, ad account, app and dataset, read-only at first if you prefer.
- Add it as a user to your measurement partner, App Store Connect, Google Play Console and your subscription revenue tool.
- List every server-side integration and API connection, and find out where each one's access token was created. Replace any made under the old agency's portfolio, or by a person who is leaving, before you remove their access, then confirm events are still arriving.
- Check the payment method on the ad account is yours, not the agency's card with spend re-billed to you.
- On handover day, remove the outgoing agency's partner access and any individual user access, then check that every event still arrives in Events Manager the next morning.
When switching agencies is the wrong fix
A switch is expensive even when it goes smoothly, so it is worth one honest check first. Look at the baseline. If cost per install is fine but install-to-trial or trial-to-paid has fallen, the problem is more likely onboarding, the paywall or the product than the agency, and a new agency will inherit it. If the current agency has been in place for less than about three months, the cohort clock above may mean you are judging it on incomplete data. And if the real issue is that nobody owns creative direction, our comparison of agency and in-house creative teams shows why that follows you whatever the arrangement. If you are not sure the app is ready for an agency at all, start with when a mobile app should hire a Meta Ads agency.
Taking over an account from another agency
Most of our new app accounts were run by someone else first. Our engagements start with a creative and account audit, and our pricing page describes the first cycle as research, coverage and a clean baseline, with a meaningful signal in most accounts within 30 to 60 days. In a takeover that first cycle is exactly the handover work above: ownership, measurement, the baseline, the test log and new concepts launched beside the existing winners. Our Meta Ads service puts creative and campaign management in one team, which removes one of the two handovers you would otherwise be managing.
If you are planning a switch and want the transition run this way, see how we work with consumer and subscription apps, look at our case studies, and apply to work with us or book a call before you give notice, so the ownership check happens while it is still easy. For where paid acquisition fits in the wider funnel, our subscription app marketing strategy guide is the place to start.
Frequently asked questions
Can I move my Meta ad account from my old agency's Business Manager to mine?
Usually not. Meta's Business Help Centre says an ad account can only be in one business portfolio, and that adding an existing ad account to a portfolio may permanently move it there so it cannot be transferred to another portfolio. If your old agency's portfolio owns the account, the practical options are partner access during the handover and a new ad account in your own portfolio afterwards.
Will switching agencies reset the Meta learning phase?
Changing who manages the account does not reset anything on its own. What resets learning is what the new team does: Meta says new ads and ad sets, and significant edits to existing ones, put an ad set into the learning phase, which usually ends after about 50 results in the week after the last significant edit. Launching new concepts alongside your current winners, rather than rebuilding everything at once, keeps the stable ad sets delivering.
How long should the overlap between two agencies be?
Long enough for the incoming agency to finish its access and measurement audit before it takes control, which in practice means using most of the outgoing agency's notice period. A 30-day notice period is usually enough if access is granted in the first week. The outgoing agency should keep running the account without restructuring it until the handover date.
How soon can I judge the new agency after switching?
Not on revenue for about two months. For roughly six weeks after the handover the subscription revenue you see mostly comes from cohorts the outgoing agency acquired, because paid conversions keep arriving for weeks after install and the last iOS postback can arrive around day 41. Judge the first month on the audit, the baseline and how quickly new concepts went live, and the second on trial starts.
What should I ask my old agency to hand over?
Exports of campaign performance by ad, ad set and country for at least the last 12 months, the raw project files for every ad that ran, creator contracts and usage licences with their expiry dates, the test log showing what was tested and what won, naming conventions, notes on the measurement setup, and a list of every integration or access token it created.
The short version
Check you own the portfolio, ad account, app ID, dataset, measurement account and creative files before you give notice, starting with the app ID. Use the notice period as an overlap in which the outgoing agency holds the account steady and the incoming one audits and freezes a baseline. Hand over access by adding first and removing last. Keep the winners running, close the creative gap with a final batch and the raw files, and launch new concepts beside the old rather than rebuilding. Then wait about two months before judging the new agency on revenue, because until then you are mostly looking at the old one's cohorts.