Mobile App Marketing Services 2026:What a Launch Plan Costs

"What should I budget to market the app when it launches?"

I get some version of that question on nearly every mobile discovery call, usually right after we've agreed the build scope. The founder wants one number. I understand why. A single number fits in a pitch deck, a board update and a bank balance.

Here's the claim, with the date on it. On September 29, 2026 I rebuilt our launch budget template from scratch, using only prices and rules I could read on an official page that day. It came out as nine lines. Two are fixed store fees that add up to $124. Four are tools and services with published prices. And the biggest line by far, ad spend, has no published price at all, because Apple and Google both run auctions and you choose the budget.

So the honest answer to "what does it cost" is a worksheet, not a figure. This article is that worksheet. By the end you'll be able to fill it in for your own app, and you'll know which lines have hard minimums set by the platforms and which ones are pure choice.

To be clear about my position: I run Frenchy Digital, which builds apps and sells some of the services priced below. Where I use our own bands, I say so and link the page they come from. Where I use someone else's price, it's from their own pricing page, checked the same day.

This piece is about money and contents. If you want the tactical how of store listings (keyword research, screenshot sequencing, review management), that lives in our App Store Optimization guide, and I won't repeat it here.

The model most founders walk in with goes like this: find the average cost per install for my category, multiply by the number of users I want, add an agency fee, done. It feels rigorous. It has a multiplication sign in it.

It fails at the first step. I spent part of the research for this piece chasing "average cost per install" figures, and every one I found traced back to a tool vendor's blog or a marketing platform's report with an undisclosed sample. None came from Apple or Google. I cover the specific ones in the refusals section, but the short version is that the input the whole model depends on doesn't exist in a form you should plan around.

The model that works runs the other way. Start from the platform rules, which are published, and from your own economics, which you know better than anyone. Google tells you the minimum daily budget its system needs relative to the target you set. Apple tells you how its two tiers charge and where Basic caps out. Tool vendors publish their prices. Agencies (the good ones) publish bands.

Therefore: you don't estimate the cost of a launch. You decidethe spend lines, and the platform rules tell you whether your decision is big enough to produce any signal. That flip matters, because it turns a question nobody can answer ("what will it cost?") into one you can ("what am I willing to pay per install, and is my budget at least 50 times that per day?").

Think of it like a grocery run rather than a restaurant bill. Nobody asks what groceries "cost" in the abstract. You write the list, some items have fixed prices on the shelf, and the rest depend on how much you decide to buy.

A complete launch plan has nine components. Some are one-off work, some are monthly, and two are pure platform spend. Here's the list I hand founders, with who normally does each piece and how it's paid for.

Three things jump out when you lay it out like that. First, several of the most useful levers (custom product pages, custom store listings, In-App Events, store tests) cost nothing in platform fees. They cost labor. Second, measurement is a one-time build that every later line depends on, which is why I put it before any ad spend. Third, only two lines are ad spend, and they're the only ones where the number is entirely yours.

What's not on the list matters too. Influencer seeding, PR, content marketing and organic social can all help a launch, and if you have an audience already they may be your best channel. I left them off because their costs are too specific to your market to put on a shared worksheet. If social is part of your plan, our social media marketing page publishes its own bands.

Store listing work is the cheapest line on the sheet and the one with the tightest constraints. Both stores publish hard limits, and the limits are small enough that the work is closer to writing a headline than writing a page.

On Apple's side, the product page guidance sets the app name at up to 30 characters and the subtitle at up to 30 characters. The keyword field is 100 characters in total, comma separated with no spaces. Promotional text, which sits above the description and can change without a new app version, runs up to 170 characters. You can show up to 10 screenshots and up to three app previews, each up to 30 seconds.

On Google's side, Play Console Help gives the app name 30 characters, the short description 80 and the full description 4,000, and notes the limits apply whether the characters are full-width or half-width. There is no separate keyword field on Play; the text itself does that job.

So what's actually being bought when an agency quotes store listing work? About 190 characters of Apple metadata that carry most of the search weight, 110 characters of Google headline text, and a set of screenshots. The writing takes hours. The screenshots take days. Almost all of the cost is design time.

The budget line here is labor plus one tool. For keyword research, AppTweak's pricing page lists Essential at $79 a month (500 tracked keywords) and Grow at $299 a month, with annual plans that save two months. A single app at launch rarely needs more than the entry tier. I use AppTweak as the example because it publishes prices; several competitors don't.

First, the name. If a proposal says "Apple Search Ads", it's working from older material. Apple rebranded its advertising business as Apple Ads in April 2025, as 9to5Mac reported at the time, because the product had grown well beyond the single search result slot it launched with.

Apple still sells two tiers, and they're priced differently. Apple's own comparison page lays it out, and I've put it in a table because the differences decide which one a launch should start on.

Advanced runs across the four placements Apple lists on its placement options page: the Today tab, the Search tab, search results and the You Might Also Like list on other apps' product pages. Apple lists cost per tap as the pricing model for all four, with a daily budget available on every placement and a target cost per acquisition option on search results only.

The search results placement itself is changing. Apple's search results help page now says an ad can appear at the top of results or further down the page. PPC Land reported in December 2025 that Apple would add those lower positions during 2026, applied automatically to existing campaigns, with no way to pick top or lower yourself. Practically: expect some of your spend to land below the top slot, and watch tap-through and install rates by placement rather than assuming every impression is the premium one.

Apple also makes a vendor claim on that page worth quoting with its label on it. Apple says ads at the top of search results saw more than a 60 percent average conversion rate, defined as taps that led to installs, over November 2024 to October 2025. That's Apple describing its own product, measured by Apple. Useful as a sense of scale; not a forecast for your app.

Which tier for a launch?I start most first-time founders on Basic, bc the pricing model protects them. You pay per install at a maximum you set, Apple handles keywords, and the $10,000 monthly cap is more than most launch budgets anyway. The downside is real: Basic runs only in search results and gives you almost no control over keywords, so you learn less. Once there's a person (in house or agency) who will check the account at least weekly, Advanced is the better tool.

I did not find a published minimum spend for either tier on the pages checked. The minimum is effectively whatever you're willing to set as a budget, which is exactly why the next section's rule matters so much.

Google App campaigns are the other half of paid acquisition at launch, and unlike Apple, Google publishes a rule that effectively sets your floor.

The bidding options first. Google Ads Help's page on App campaign bidding lists target cost per install, target cost per action, target return on ad spend, and Maximize conversions with variants for installs, in-app actions and conversion value. It also carries a warning that I wish every founder read: Maximize conversions campaigns with small budgets may not be able to get any conversions.

Now the rule. Google's App campaign best-practices guide says to set a daily budget of at least 50 times your target CPI when optimizing for installs, and at least 10 times your target CPA when optimizing for in-app actions. It also says to pick an in-app action that at least 10 different users complete each day.

Do the arithmetic out loud, bc this is where launch budgets break. Google's own worked example is a $2 target CPI. Fifty times $2 is $100 a day. Over a 30-day month that's $3,000, and over a 90-day launch window it's $9,000, before a single dollar of management or creative.

The in-app action version is steeper. Say you want to optimize toward trial starts and you set a $20 target CPA. Ten times $20 is $200 a day, or about $6,000 a month. And the action has to be common enough that 10 or more users complete it every day, which a brand-new app often can't hit in week one.

That's the multiplier that should reshape your plan. If your realistic budget is $1,000 a month, the rule says your target CPI can't be much above about 67 cents ($1,000 divided by 30 days, divided by 50). If that target is far below what the auction will give you, the campaign won't learn, and you'll conclude "Google doesn't work for us" when the real finding was "the budget was below Google's published floor".

To be clear, these are Google's recommendations, not hard minimums enforced by the system. You can run smaller. You just shouldn't expect the system to optimize well if you do, and you should read the first weeks of results with that in mind. If you want help, we manage App campaigns under our Google Ads management service, alongside Search, Display, Shopping and Video.

Measurement is the line founders most want to skip and the one I refuse to skip. Every other line depends on it. Without it you can spend the ad budget, but you can't tell which half worked.

On iOS the rules are set by App Tracking Transparency. Apple's user privacy page says that since iOS 14.5, apps must get permission through the ATT framework before tracking a user or accessing the device's advertising identifier. Apple defines tracking as linking user or device data from your app with data from other companies' apps, websites or offline properties for targeted advertising or advertising measurement, or sharing it with data brokers.

In practice, a real share of users will decline the prompt, and for them user-level attribution across ad networks is off the table. The same page points advertisers to AdAttributionKit, which Apple calls its privacy-preserving technology for attributing in-app ad campaigns and web ads on mobile. That's the successor path to SKAdNetwork, and a launch plan in 2026 should describe how it's used rather than promising user-level iOS data.

On Android the picture changed in a direction many plans haven't caught up with. Google spent years building Privacy Sandbox on Android as the replacement for its advertising ID. Then, on October 17, 2025, it announced it would retire several of those technologies, including the Attribution Reporting API, SDK Runtime, Topics and Protected App Signals, citing low adoption. If a proposal still builds Android measurement around those APIs, it's stale.

What does the measurement line actually buy? An event plan (install, sign-up, the one action that predicts revenue, purchase or trial start), those events wired into your analytics and ad platforms, subscription revenue flowing into the same view, and a dashboard someone looks at weekly. Our analytics setup service publishes $1,500 to $6,000 for that work, depending on scope.

For subscription apps, the revenue side usually comes from a subscription platform. RevenueCat's pricing page lists its Pro plan as free up to $2,500 in monthly tracked revenue and 1 percent of tracked revenue above that. For most launches that line is zero for months. That's a nice property: the tool starts costing money roughly when the app starts making it.

The fixed store costs are the only lines on the sheet with no range. The Apple Developer Program costs $99 per membership year, with fee waivers for eligible nonprofits, accredited schools and government entities. Google Play charges a one-time $25 registration fee. Together that's $124 in year one and $99 a year after.

Commissions aren't a marketing cost, but they belong in the plan bc they decide how much you can afford to pay per user. And they've moved more in the last 18 months than in the previous decade.

Apple's program page states a standard 30 percent commission on digital goods and services, 15 percent for qualifying subscriptions and for members of reduced-rate programs such as the App Store Small Business Program, and notes regional variations in places including the European Union, Japan, South Korea and Brazil. In the United States, the court fight over purchases linked out of apps is still live: as Courthouse News reported on August 11, 2026, the district court is determining what commission Apple may charge on linked-out purchases, after Apple was found in contempt in April 2025 for imposing a 27 percent one, while the Supreme Court reviews the contempt finding. That status is pending, not settled.

Google's service fee page still describes 15 percent on the first $1 million of annual earnings and 15 percent on subscriptions for most markets, and a new structure starting June 30, 2026 in the US, UK and EEA that splits fees by new versus existing installs and separates a billing fee. It's detailed enough that I won't paraphrase every row; read the table for your app type before you model margins.

Why does this belong in a marketing article? Because your maximum affordable cost per install is a function of what you keep per user. A subscription app that keeps 85 percent of revenue can bid differently from one that keeps 70. Get the fee right before you set a target CPI.

Here's the whole worksheet in one table. Every price is from a page checked September 29, 2026. Three lines use Frenchy Digital's own published bands, and I've labeled them.

A few notes on reading it. The management band on our Google Ads page runs from $1,000 to $1,500 a month for accounts up to $5,000 of spend, $2,000 to $3,500 up to $15,000, and $4,000 to $5,000 and up above that. Management fees in this industry are often tied to spend in some way, so ask any agency how its fee moves when your budget does.

The conversion optimization line is optional at launch. It becomes worth paying for once paid traffic is steady enough to test against: store page tests, onboarding changes, paywall changes. Before that, you're paying people to optimize a funnel with too few visitors to tell anything.

And the two spend lines are yours. The Apple one has a ceiling if you use Basic ($10,000 per app per month). The Google one has a floor if you follow Google's guidance (50 times your target CPI per day). Everything else on the sheet is small next to them.

A worked example, not a client result.Consider a founder launching a subscription app on both stores in the United States, with a 90-day launch window and no existing audience. Suppose they choose Google's own illustrative $2 target CPI, put $1,500 a month into Apple Ads Basic, use an entry-tier ASO tool, and hire someone for measurement and campaign management at the bottom of our published bands.

So about $18,400 to $20,900 for the quarter. Of that, $13,500 is ad spend, which is roughly 65 to 74 percent of the total depending on the end of the range. Tools and store fees are about $361, less than 2 percent. That ratio is the thing to remember: the parts people argue about (which tool, which agency) are small, and the part people skip deciding (spend and target CPI) is most of the money.

Now the multiplier. If the founder's real budget is $10,000 for the quarter instead of about $19,000, they need roughly 0.53x of this plan. The fixed lines don't shrink, so the cut has to come from spend. Drop Apple Ads to $500 a month and cut the Google target CPI to $1 (which halves the rule's daily floor to $50, or $4,500 over 90 days), and the total lands near $10,860 at the low end. Whether a $1 target is achievable for that app is exactly the thing the first four weeks will tell you.

Two warnings about the scenario. The $2 figure is Google's example in its documentation, not a benchmark for any category, and I'm using it only because it's the one published number I can attribute. And the scenario doesn't include creative, PR or social, which can each be larger than everything else for some launches.

I won't cite a client's install numbers or ad results here; none of our case studies publish them, and I won't invent them. What our shipped work does show is how much of the launch budget gets decided during the build.

On Fighter Cut, the weight-cut app for combat athletes, the case study describes a fully headless iOS pipeline (xcodebuild plus the App Store Connect API, no clicking through Xcode), TestFlight distribution, and RevenueCat Web Billing with Stripe for payments. That matters to marketing because a release pipeline you can run from a script makes it cheap to ship store metadata, screenshots and small fixes often, which is what store testing needs.

The Corner, Fighter Cut's companion app for gyms and coaches, sits on its own bundle id under the same Apple Developer team. One $99 membership covers both apps. Small, but it's the kind of line that doubles for no reason when two apps get set up by two different people.

And on the Drink Caveman funnel (a Shopify store, not an app), the case study describes signed links per channel, a 30-day token for Meta ads and a 365-day one for email, each tagged with its campaign, and a Meta Pixel hooked so it fires exactly once per event. The lesson carries straight over to apps: if you can't tell which channel a user came from, or your events double-count, your cost per acquisition is fiction no matter how much you spend.

The downside I'd name: analytics added after launch always costs more, because the first weeks of paid data are the ones you can't read. Put measurement into the build scope.

Several figures dominate this topic online, and none of them made it into the worksheet. Here's what I chased and why each one stayed out.

• Average cost per install by category or country. Every version I found came from a tool vendor, ad network or marketing platform blog with an undisclosed sample, and none from Apple or Google. I don't print any of them as a benchmark.
• A US average cost per tap for Apple Ads. One vendor blog states a specific US figure; it gives no sample, date range or method I could check, so it stays out.
• Claims that Apple Ads Advanced delivers a fixed percentage lower cost per install than Basic. Found on a subscription tooling vendor's blog, with no published method. Not used.
• A single average cost to market an app. Totals quoted in listicles range widely and never show their line items. The worksheet above is the replacement.
• Launch week budget splits between Apple and Google presented as best practice. These are one agency's preference written as a rule. Your split should follow where your users are.

Two figures I did use come with a vendor label on them: Apple's own 60 percent conversion claim for top search results, and Google's own $2 example CPI. Both are the platform describing itself. I treat them as scale, not as forecasts.

If you're comparing agency proposals for a launch, these are the patterns I'd push back on, and what to ask for instead.

The account ownership one deserves an extra sentence. Your Apple Ads and Google Ads accounts carry the learning history your campaigns depend on. If they live in an agency's name, you lose that history when you change agencies. Own the accounts; grant access.

Arguing against my own worksheet for a moment, because it has failure modes.

It can make paid look mandatory.Some apps shouldn't buy installs at launch at all. If your revenue per user is unknown and your onboarding is untested, the cheapest plan is store listing work, measurement, and a small Apple Ads Basic budget to learn from. Spend comes after the numbers exist.

The budget rule can be misread as a guarantee.Meeting Google's 50x guidance means the system has room to learn. It doesn't mean you'll get installs at your target. If the auction won't deliver at your CPI, you'll see low volume, and the right response is to rethink the target or the creative, not just add money.

Platform rules move.In the last 18 months Apple renamed its ad product and added search placements, Google retired most of Privacy Sandbox on Android, and both stores changed fee terms in some markets. A plan written from last year's notes will be wrong somewhere.

Why is the worksheet still the right call? Because the worst case is bounded. If you follow it and the first month fails, you've spent a known, chosen amount and you have clean data explaining why. The alternative (a package price and a hopeful CPI) fails in a way you can't diagnose.

What I couldn't verify, stated plainly. I found no published minimum spend for Apple Ads Basic or Advanced on the pages checked; if one exists in the account interface, I didn't see it. I found no official Apple or Google average cost per install or cost per tap, and I used none.

I didn't confirm an exact date on which Apple's additional search result positions went live in every market; Apple's help page describes them and a trade outlet dated the announcement, but I couldn't find an Apple rollout date. The US linked-out commission question is pending in court as of the August 2026 report cited, and could change after this is published. Google's new fee table is detailed enough that you should read it directly rather than rely on my summary.

Tool prices are list prices from public pages on September 29, 2026 and exclude tax, discounts and enterprise terms. Frenchy Digital's bands are our own published ranges, not quotes.

You can turn this into a real plan in about a week. Here's the order I'd do it in.

1. 1.Work out what you keep per paying user after the store's fee for your app type, then set a maximum target cost per install you can live with. Write it down.
2. 2.Multiply that target by 50 for Google's daily floor and by 90 days for the quarter. Add an Apple Ads budget, the $124 in store fees, a tool and measurement. If the total is more than you have, lower the target or the channels, not the measurement.
3. 3.Open both developer accounts and both ad accounts in your company's name this week, draft your 30 character app names and subtitle, and list the three audiences that would each get their own custom product page.

Do those three and any proposal you get will be measured against your numbers, not the other way round. Time to fill in the sheet.