A mobile app should hire a Meta Ads agency when three things are true at the same time. The funnel after install already converts, so paid users have somewhere to go. The account has enough budget for Meta to learn on an event that matters, which by Meta's own guidance means about 50 results a week per ad set. And the thing stopping you spending more profitably is something an agency can actually fix: creative supply, media buying skill, or the hours the founder no longer has. If the first condition is missing, an agency buys you a faster and more expensive reading of a product problem. If the second is missing, the fee eats the test budget. If the constraint is your onboarding or paywall, no media buyer will move it.
Most advice on this question gives a spend threshold and stops there. Spend is a proxy. Two apps on the same budget can need completely different help, and the honest answer for some of them is not to hire anyone yet. This guide sets out the nine signals we look at when an app asks whether it is ready, the arithmetic that decides whether an account can learn at all, four stages and the right kind of help at each, and the situations where we would tell an app to keep its money. It is about timing. If you have already decided to hire and want to know how to judge the candidates, our guide to what to look for in a mobile app marketing agency covers selection.
Key takeaways
- Hire for a bottleneck, not a milestone. A Meta Ads agency fixes creative supply, media buying and operator time. It does not fix retention, onboarding or a weak offer.
- Meta says an ad set usually leaves the learning phase after about 50 results in the week after its last significant edit. If your budget cannot buy that on a useful event, the account cannot learn and an agency has little to work with.
- Organic conversion is your control group. If users who find you without ads rarely start trials or pay, the ads are not the first thing to fix.
- The strongest case for an agency is an app where paid already works at a small budget but stalls because a handful of ads carry the account and nobody can make replacements fast enough.
- At low spend the fee is a large share of total outlay. Work it out before you sign, and compare it with the cost of your own time.
Hire for a bottleneck, not a milestone
A Meta Ads agency sells three things. It sells media buying: account structure, budgets, optimisation events and the judgement to leave things alone during learning. It sells creative, if it is the kind of agency that makes ads as well as running them. And it sells time, because every hour the agency spends on your account is an hour the founder or the growth lead gets back.
The decision to hire should start from which of those three is actually short. An app with a capable founder-operator, a working funnel and a backlog of strong ideas it cannot produce has a creative problem. An app whose ads are fine but whose campaigns are rebuilt every week by someone learning Ads Manager as they go has a media problem. An app whose founder is spending two days a week on ads instead of the product has a time problem. All three are good reasons to hire. An app whose users install, look around and leave has none of them, and that is the case we see most often on first calls.
One test sorts most apps quickly. Ask what would happen if the agency did its job perfectly and doubled the number of good installs next month. If the honest answer is that revenue would roughly double, the bottleneck is acquisition and an agency can help. If the answer is that twice as many people would churn in their first week, the bottleneck is downstream and you should spend the money there.
The three conditions that must all be true
1. The funnel after install already converts
Paid traffic is only as valuable as what happens after the install. The cleanest check is your organic users, because they arrived with no ad promising anything. If they start trials and pay at a rate you can live with, paid users have somewhere to go. If they do not, better ads will pull more people into the same leak.
For a rough external reference, RevenueCat's State of Subscription Apps 2026 puts the median share of installs that pay within 35 days of download at 2.0% across categories and 2.9% in Health & Fitness. That is a download-to-paid rate, counted on all installs rather than trial starts, so do not compare it with a trial-to-paid figure. Our benchmark guides for install-to-trial and trial-to-paid conversion break the funnel down by category, platform and trial length. Being below a median is not disqualifying on its own. Not knowing your own numbers is, because nobody can then tell whether a change in ad performance came from the ads.
2. The account has enough budget to learn
Meta's Business Help Centre page on the learning phase says ad sets exit learning once they can deliver stably, which usually happens after about 50 results in the week after the ad set's last significant edit. A result means the event you optimise for. That turns budget into a simple test.
Weekly budget to leave learning ≈ 50 × cost per optimisation event
Monthly equivalent ≈ weekly budget × 52 ÷ 12
The table applies it to three events. The costs per event are hypothetical round numbers for illustration, not benchmarks and not client data.
| Optimisation event | Hypothetical cost per event | Spend for 50 a week | Per month, one ad set |
|---|---|---|---|
| App install | £2 | £100 | about £430 |
| Trial start | £15 | £750 | about £3,250 |
| Paid subscription | £60 | £3,000 | about £13,000 |
Almost any budget can teach Meta to find installs. Far fewer can teach it to find payers. An app spending £1,000 a month can only optimise for an event it buys hundreds of times a month, which for most subscription apps means installs, and installs are where cheap, low-intent traffic hides. That does not make small budgets useless, but it limits what any operator, in-house or agency, can learn from them. On iOS the reading also arrives late, because attribution postbacks are delayed, which we cover in our guide to SKAdNetwork postback delays. The fuller mechanics are in our explainer on the Meta Ads learning phase.
This is the reasoning behind our own minimum of £3,000 a month in Meta spend. Below it, there is not enough delivery behind each ad to separate a real winner from noise, so every creative decision becomes a guess.
3. The constraint is one an agency can fix
Creative supply, media buying and operator time are fixable from outside. Pricing, onboarding, the paywall, the product itself and app store presence are not, or not by a Meta specialist. Be precise about which one you have before you start interviewing. We do not offer app store optimisation or lifecycle marketing, and an app whose real problem sits there should hire someone who does.
Nine signals that say hire now, or wait
These are the nine things we look at when an app asks whether it is ready. No single row decides it. A run of signals in the left-hand column means an agency is likely to pay for itself. A run in the right-hand column means the money is better spent elsewhere first.
| Signal | Points towards hiring | Points towards waiting |
|---|---|---|
| Current Meta spend | Enough to buy about 50 results a week on the event you optimise for, with budget left to test new creative. | So small that one ad set cannot leave the learning phase, or the agency fee would be a large share of total outlay. |
| Product-market fit | Users who stay, pay and renew without being chased. Retention curves that flatten rather than fall to zero. | You are still changing what the app is for. Paid traffic buys a faster, dearer reading of the same uncertainty. |
| Organic conversion | Organic installs start trials and pay at a rate you can build a CAC target on. | Organic users rarely convert. The ads are not the first thing to fix. |
| Trial and paid conversion | Install-to-trial and trial-to-paid are stable from month to month, so a change in paid traffic quality is visible. | Onboarding or the paywall is changing every fortnight, so no creative result can be read cleanly. |
| Creative capacity | You cannot make enough genuinely different concepts each month to replace the ads that fatigue. | You have a team that already produces range and volume, and media buying is the only gap. |
| Founder time | Ads are eating hours the founder should spend on product, hiring or fundraising, and the work is getting done badly. | The founder is still learning what message sells. That learning is worth doing first hand. |
| Attribution | Installs and in-app events reach Meta and your revenue data, so results can be read by cohort. | Nothing is wired up. Fix tracking first, or make it the first paid job, not an afterthought. |
| In-house expertise | Nobody on the team has run app campaigns on Meta through an iOS attribution cycle. | You already employ an experienced UA manager who needs more creative, not more management. |
| Scaling bottleneck | Paid works at your current budget but results fall apart every time you push spend up. | Paid has never worked at any budget, and nobody can say why. |
Three of these deserve more than a table row. The first is founder time. Early on, a founder running their own ads learns things no agency can learn for them: which promise makes strangers install, which objection stops them paying, what words customers use. That is worth doing first hand, and handing it off too early loses it. The moment to hand it off is when the founder already knows the message and the work has become production and maintenance.
The second is attribution. An agency hired into an account with no working event tracking will spend its first month on plumbing. That is legitimate work, but agree it as the first deliverable rather than discovering it in week three. Check that installs, trial starts and purchases reach Meta, and that someone can see revenue by install cohort, before judging any campaign.
The third is the scaling bottleneck, which is where most apps that should hire find themselves. Paid works at £5,000 a month. At £15,000 it falls apart. Usually the cause is creative rather than bidding: the account is leaning on two or three ads, frequency climbs, and the replacements are variations of the same idea. Our guide to creative fatigue on Meta for mobile apps explains why more versions of one angle do not fix it.
Four stages, and the right help at each
Most apps sit in one of four stages. The right kind of help changes between them, and an agency is only the best answer in one and a half of them.
| Stage | What is true | Right help | Wrong help |
|---|---|---|---|
| 1. Funnel unproven | Retention is weak or unknown, onboarding and paywall still changing. | Product work. Small founder-run tests to learn which message lands. | A Meta Ads agency on a retainer. |
| 2. Funnel works, paid untested | Organic users convert. Tracking is partial. Spend is below the level where an ad set can learn on trials or purchases. | Founder or a freelance media buyer to set up attribution and run first tests. | A full-service retainer that eats most of the test budget. |
| 3. Paid works, but stalls | A few ads carry the account. Fatigue arrives faster than replacements. The founder cannot keep up. | A Meta Ads agency that makes the creative and runs the media, so testing and reading happen in one loop. | More budget on the same ads. |
| 4. In-house team at scale | An experienced UA lead owns the account. Creative range is the constraint. | A specialist creative partner alongside the in-house team, or a hybrid split. | Handing the whole account to an agency and losing the internal knowledge. |
Stage three is the clearest case for a Meta Ads agency, and specifically for one that makes creative and runs media in the same team. The account has proved it can work. What it lacks is a steady supply of genuinely different concepts and someone reading the results properly enough to brief the next round. Stage four can justify an agency too, but usually as a creative partner working alongside an in-house UA lead rather than a replacement for them. We costed that trade-off in agency versus in-house creative teams.
When a mobile app should not hire a Meta Ads agency
There are five situations where we would tell an app not to hire us or anyone like us.
- Before product-market fit. If users do not stay, paid acquisition measures that faster and at a higher price. Spend the money on the product.
- When the leak is onboarding or the paywall. An agency can send better-matched users, but it cannot make a confusing onboarding convert. Our own Steps & Beasts case study lists an onboarding overhaul next to heavy creative testing and AI UGC; the case study credits the onboarding changes with lifting install-to-subscriber conversion. The +1112% revenue came from the ads and the funnel together, which is the point.
- When the fee would swamp the media budget. See the table below. At very low spend most of your money should be in the auction.
- When the problem lives on another channel. If your growth is limited by store ranking, retention messaging or a channel other than Meta, a Meta specialist is the wrong hire.
- When you cannot give access to the data. An agency that cannot see revenue by cohort ends up optimising to platform-reported numbers, which is how cheap installs get mistaken for good ones.
What the fee costs you at different spend levels
The fee looks very different as a share of what you spend in total. The table uses our own published starting fees from our pricing page to show the effect. They are starting prices, so treat the shares as the best case.
| Monthly Meta spend | Our monthly fee | Total monthly outlay | Fee as share of total |
|---|---|---|---|
| £3,000 | from £1,500 | £4,500 | 33% |
| £10,000 | from £1,500 | £11,500 | 13% |
| £20,000 | from £1,500 | £21,500 | 7% |
| £50,000 | from £3,250 | £53,250 | 6% |
At our £3,000 minimum, a third of the total goes on the fee. That can still be the right call if the alternative is a founder losing two days a week, or a budget spent on ads nobody is reading properly, but it needs to be a deliberate decision. At £20,000 a month the same fee is about 7% of the total and the case is far easier. For how other agencies charge, including percentage-of-spend models, see our breakdown of Meta Ads agency pricing.
Put your own time on the other side of the scale. Count the hours a week you spend making ads, building campaigns and reading reports, and price them at what that time would earn the business elsewhere. If the result is close to the fee, the decision comes down to whether the agency would also do the work better.
The cost of hiring too late
Waiting has a cost too, and it is easy to miss because it shows up as nothing happening. The usual pattern is an app that found two or three winning ads early, rode them for months, and now spends the same budget every month at a slowly rising cost per trial. Nobody decided to stop growing. There simply was no system for replacing the winners, so the account settled at the level its oldest ads could support.
If that sounds familiar, the signals are visible in the account: the same ads at the top of spend for more than a couple of months, frequency climbing on your core audience, new tests that are versions of old winners, and a budget you could increase but do not because each increase makes results worse. Our guide to the Meta Ads testing budget shows how much of the spend should be going on new concepts, and our creative refresh calculator estimates how many your account needs each month.
What to have ready before you talk to an agency
The best first calls are the ones where the app brings numbers. You do not need all of them, but the more of these you can answer, the faster a good agency can tell you whether it can help and a bad one runs out of things to say.
- Monthly Meta spend over the last three months, and the event each campaign optimises for.
- Install-to-trial and trial-to-paid, ideally split into paid and organic users.
- What you can afford to pay for a subscriber. Our guide to setting a target CAC shows how to work it out.
- How many new ad concepts you launched last month, and how many are genuinely different from each other.
- Who owns the ad account, the tracking and the raw creative files.
Then use our list of questions to ask a Meta Ads agency on the call itself, and if you are still building a shortlist, our comparison of the best Meta Ads agencies for mobile apps ranks the specialists on stated criteria, including where each one is the wrong choice.
How we decide whether to take an app on
We run the three conditions above on every application. We look for a funnel that converts, a budget of at least £3,000 a month on Meta, and a constraint we can move, which for us means creative and media together. Our Meta Ads service is built for stage three: an app where paid works but stalls, and where one team making the ads and reading the results closes the gap faster than two teams passing briefs back and forth. If the unit economics do not work, we would rather say so than take a retainer to scale a leaky funnel.
If that describes your app, see how we approach consumer and subscription apps, and apply to work with us. If you are not there yet, the honest advice is to keep the money in the product or the auction and come back when the account can learn from what it runs.
Frequently asked questions
When should a startup app hire a Meta Ads agency?
When three things are true at once. Users who arrive organically already start trials and pay at a rate you are happy with, the account has enough budget for Meta to learn on the event you care about, and the thing stopping you spending more profitably is creative supply, media buying skill or your own time. If the funnel after install is the problem, fix that first, because an agency will only measure the leak faster and at a higher cost.
How much should an app spend on Meta before hiring an agency?
Enough for the account to learn on a meaningful event. Meta's guidance is that an ad set usually exits the learning phase after about 50 results in the week after its last significant edit, so the floor depends on what each result costs you. At a hypothetical £15 per trial start, 50 trials a week is roughly £3,250 a month for one ad set. Our own minimum is £3,000 a month in Meta spend, and below that we tell apps to put the money into the auction rather than into fees.
Is a Meta Ads agency worth it at £3,000 a month in ad spend?
Sometimes, but be honest about the maths. With our published Starter fee from £1,500 a month, an app spending £3,000 is paying a third of its total outlay to the agency. That only makes sense if the agency saves more than that in wasted spend or founder time, or produces creative the app cannot make itself. At £20,000 a month in spend the same fee is about 7% of the total, which is a much easier case to make.
Should I hire a freelancer or an agency first for Meta Ads?
If the main job is setting up the account, tracking and a first round of tests, a strong freelance media buyer is often the better first hire because the scope is narrow and the cost is lower. An agency starts to make more sense once the bottleneck is producing enough distinct creative every month and reading the results properly, because that is a team's worth of work rather than one person's.
How long does a Meta Ads agency take to show results for an app?
Plan on at least a quarter before judging it. Our own pricing page says most accounts see a meaningful signal within 30 to 60 days and a compounding effect from month three, and our engagements start with a three-month commitment for that reason. The first month is research, creative and a clean baseline; the gains come from each round being briefed off what won in the last one. Nobody can honestly guarantee a ROAS figure in advance.
The short version
Hire a Meta Ads agency when the funnel after install works, the account can buy about 50 results a week on an event worth optimising for, and the thing holding you back is creative supply, media buying or your own time. Do not hire one to fix retention, onboarding or a paywall, and be honest about what share of your total outlay the fee represents at your spend. The right time is usually the moment paid acquisition has proved it can work and then stops growing, not the moment you first decide to try it.