Your First Paid Ads for an App (On a Budget That Can Lose)

At some point after launch, every founder considers buying users. The logic is appealing: organic growth is slow, ads are instant, and surely a few hundred dollars will tell you something.

It might. But you should know what you are walking into. In 2026, the average cost per install on iOS sits somewhere between $2 and nearly $6 depending on whose data you use and which vertical you are in, and iOS ran roughly 19% more expensive than a year earlier. Android is substantially cheaper, commonly a third to half the iOS price. Those are installs, not paying customers. If one in twenty installs converts to a subscription, your real customer acquisition cost is twenty times the install cost.

Here is how to think about a first paid budget without setting money on fire.

Do not run ads yet if any of these are true

Paid acquisition amplifies what already exists. If the underlying app leaks, ads make you lose money faster and more precisely.

Your retention curve falls to zero. Buying users for an app nobody sticks with converts cash into churn. Fix retention first; it is the cheaper problem.

You do not know your activation rate or conversion rate. Without these, you cannot tell a good campaign from a bad one. You will see installs, feel encouraged, and learn nothing.

You have no attribution set up. If you cannot tell which ad produced which install, you are buying a number with no story attached.

Organic channels are still untouched. If you have an audience, a community, or a content channel you have not worked seriously, those are cheaper and compound. Ads stop the moment you stop paying.

The honest sequence is: retention works for some segment, you know your funnel numbers, attribution is wired, organic is running. Then ads become a way to accelerate something that already works rather than a search for something that does not exist.

What a small budget actually buys

Be realistic about scale. At a $3 blended install cost, $500 buys roughly 150 to 200 installs. If 30% activate and 10% of those subscribe, that is about five customers.

Five customers will not tell you whether your business works. But that is not the point of a first budget. The point is to learn what your acquisition math looks like, so you know whether a bigger budget would ever make sense.

So treat the first spend as a measurement exercise with a fixed, small ceiling: an amount you are genuinely willing to lose, spent deliberately, with a specific question attached.

Start on the cheaper platform

If you are budget-constrained and your app exists on both platforms, start with Android. Install costs run consistently lower, often less than half of iOS, which means the same money buys two to three times more learning.

What you learn transfers: which creative angle resonates, which audience responds, what your activation and conversion rates look like once traffic arrives. You can then decide whether the iOS premium is worth paying, armed with actual numbers rather than hope.

The caveat: if your app is iOS-only, or your ICP is unmistakably an iOS audience, this does not apply. Buy where your users actually are.

Test angles, not audiences

The instinct for a first campaign is to obsess over targeting. In 2026 the platforms' algorithms do most of that work, and they do it better than a founder guessing at interest categories. Your lever is creative.

Run three or four distinctly different angles rather than four variations of the same ad. Different angle means a different reason to care:

• The problem stated plainly, in your users' words
• The outcome, shown rather than described
• The specific audience called out ("if you're a shift worker who...")
• The contrarian take, if your positioning has one

Two notes from what performs now. First, UGC-style creative consistently outperforms polished brand video for app installs, which is good news for a founder with a phone and no budget. A real person talking plainly about the problem beats a slick animation. Second, short vertical video, roughly 15 to 30 seconds, dominates. If you can only make one thing, make that.

Keep each ad pointed at a landing experience that matches it. An ad promising a specific outcome that lands on a generic store listing wastes the click you paid for.

The number that decides everything

One question matters: can you acquire a customer for less than they are worth to you?

Work out two numbers before you spend:

Customer acquisition cost. Not cost per install. Total spend divided by paying customers acquired. If $500 produced five subscribers, your CAC is $100.

Customer lifetime value. Roughly: monthly price times the average number of months people stay, minus store commission. A $10 subscription with average retention of four months nets you around $28 after a 30% store cut.

In that example, CAC of $100 against LTV of $28 means paid acquisition is deeply unprofitable, and no amount of optimization closes a gap that size. That is not a failed experiment; that is a decisive answer for $500. You either raise LTV (better retention, higher price, annual plans) or you grow through channels that are not priced per click.

The rule of thumb many subscription businesses use is LTV at least three times CAC before scaling spend. Below that, the economics are too thin to survive the variance.

How to run the first test

Set a fixed budget you can lose. $300 to $500 is enough to learn. Decide it in advance and do not top it up mid-test because the numbers look "almost there."

Run for long enough to be real. A week or two, not a day. Algorithms need a learning period, and daily numbers early on are noise.

Change one thing at a time. Four creative angles, one platform, one audience setting. If everything varies, you learn nothing about anything.

Track to the money, not to the install. Installs are a vanity checkpoint. Instrument through activation and subscription so you can compute CAC per creative.

Expect most of it to fail. Typically one angle works noticeably better than the rest. That is a useful finding, and it is often reusable in your organic content, your landing page, and your store screenshots.

When ads make sense, and when they never will

Paid acquisition is a good fit when your LTV is high enough to absorb the cost, your funnel converts predictably, and you want to accelerate something already working. It is also the fastest way to test messaging, because you can put four angles in front of strangers in a week.

It is a poor fit for most early consumer subscription apps at a modest price point, where the maths simply does not clear. That is not failure, it is the structure of the category: this is exactly why audience building, content, referrals, and store optimization matter so much for apps like these. Those channels are slower and they compound, which is the opposite trade from ads.

Run the small test, get your number, and let the number decide. Spending $500 to learn that paid does not work for you is cheap, as long as you actually stop when it tells you so.