Cost Per Install by Channel: Meta, TikTok, Snap, Apple Search Ads and Clipping

There is no single cost per install, and the public data shows how wide the range really is. The Liftoff 2025 Casual Gaming Apps Report (data from February 2024 to February 2025) put casual-game CPI at $0.14 on Android and $1.41 on iOS, and casino-game CPI on iOS at $21.03. Same report, same year, a 150x spread. Channel matters, but category and OS matter more. Below: planning ranges by channel for US installs of non-gaming apps, the published data points behind them, and a calculator that converts a CPM (what a clipping network like FindClout actually sells) into an effective CPI, plus the break-even view-to-install rate against your current paid CPI.

This page exists because most "CPI benchmark" content either hides the sourcing behind a gated report or quietly swaps in global averages that make cheap, low-quality traffic look like a bargain. We're doing neither: every figure below says what it measures and where it came from, and every FindClout dollar figure is a ceiling on a verified US view, not a price on a global one.

Cost Per Install Benchmark Table by Channel

Read this table for what it is: planning ranges for US installs of non-gaming consumer apps, not a measured survey. The ad platforms do not publish CPI by channel, and the MMP benchmark reports that do are mostly gated behind a download form. Use the ranges to sanity-check a first budget, then throw them out and use your own numbers after two weeks of spend. A strong creative in a cheap category can beat the low end; a weak creative in finance, sports betting or dating can run several times past the high end.

Published data points

Two patterns hold in the public data: iOS costs more than Android to acquire, and finance, fintech, trading and gambling-adjacent categories consistently pay a premium over consumer and utility apps on every channel. If your own dashboard looks nothing like the table, check category before you blame the platform.

Why CPI Swings So Much Between Reports

Four variables explain almost all of the spread you'll see between any two "CPI benchmark" articles, including this one:

• Category. A finance app and a casual game are not buying the same auction. Finance, fintech, trading and sports betting apps pay a multiple of what utilities and casual games pay on the same channel, because their users are worth more and every advertiser in the auction bids accordingly. The casino-versus-casual gap in the table above is the extreme version.
• OS. Apple's App Tracking Transparency prompt means most iOS users are untracked for ad targeting, so the platforms have less signal to optimize on. That inefficiency shows up as a higher price per confirmed install, not just a higher CPM.
• Geo. "US CPI" and "global CPI" are different products. A campaign that blends in installs from lower-cost geos will report a CPI a fraction of the US-only number, which is exactly why the caveat matters below.
• Attribution window and self-reporting. Networks that self-attribute installs (view-through windows especially) tend to report artificially low CPIs versus MMP-verified numbers, because they're claiming credit for installs that would have happened anyway.

CPM-to-CPI Converter

A clipping network doesn't sell installs. It sells verified views at a CPM ceiling: FindClout's logo and caption placement runs at a $0.20 CPM ceiling and is typically delivered at $0.08 to $0.10 once overdelivery is counted. To compare that with a paid channel that prices per install, you need a view-to-install rate. Enter a CPM, a view count, a rate and your current paid CPI:

We don't publish a view-to-install benchmark, because organic placement does not click-attribute and nobody can measure that rate for your app except you. What the calculator shows is sensitivity. At the $0.20 ceiling:

The more useful number is the break-even: break-even rate (%) = CPM ÷ (10 × your paid CPI). A $0.20 CPM matches a $3.00 paid CPI at about 0.007%, roughly one install per 15,000 views. That is a low bar, and it is also why the next two sections matter: the math only holds if the views are in front of the audience and the country your app is built for. A view that can never convert has no break-even rate.

Where Organic Creator Distribution Actually Sits

Clipping networks sell views instead of installs because the view is the unit they can verify. A creator posting a piece of content that happens to feature your app is not running a direct-response ad; they're making content their audience already wants to watch, with your product placed inside it. Some of that audience installs, most doesn't, the same way most people who see a billboard don't buy the car that week. What you're actually buying is cheap, brand-safe reach inside content real people chose to watch, at a fraction of what the same reach costs as a paid impression.

That's why the input that matters most on the calculator isn't the CPM, it's the rate, and the rate is mostly a function of audience match. On FindClout every page has to clear a 40% US-audience floor, proven by the creator connecting the Instagram account itself so the audience data comes straight from Instagram, and every post in the brand's dashboard carries that creator's audience demographics. A cheap view in front of the wrong audience is the most expensive view a brand can buy, because it can't convert at any rate.

US vs Global Installs: Why Cheap CPIs Can Hide the Wrong Geo

Any benchmark, including this one, needs one caveat attached: a $0.30 blended CPI and a $3.00 US-only CPI can come from the exact same spend, just measured differently. Networks and platforms that let a campaign run open-geo will pull in installs from countries where inventory costs a fraction of the US rate, which drags the blended number down without telling you anything about whether the audience you actually wanted saw the ad. US views and global views are not the same product at any price, and a brand budgeting off a blended global CPI benchmark is comparing numbers that don't mean the same thing. The fix, on any channel, is the same: get the number broken out by geo before you compare it to anything, and don't let "verified" mean anything less than audience-level, not just device-level.

Category Receipts: What Real Campaigns Have Delivered

Two receipts from FindClout's own network, at the view level rather than the install level, because that's the honest unit for organic distribution: a sobriety app ran a campaign that delivered 25.7 million views, and a trading app delivered 160 million views in a single month while growing its follower base roughly 5x in six weeks. Neither number is an install count. They are the top of the funnel that a brand's own attribution stack (promo codes, branded search lift, App Store source reports) has to convert into installs on the brand's side, which is exactly what the organic install attribution guide below walks through in detail. The honest way to read a receipt like that is as proof the audience and the volume are real, not as a promise of a specific CPI. That number is yours to measure once the views land.

How to Blend Organic Reach With Paid Capture

The two channels aren't competing for the same job. Paid UA on Meta, TikTok, Snap, Apple Search Ads and Google is built for cost-per-action targets you can dial in and scale predictably; organic creator distribution is built for cheap volume and brand-safe reach that a paid auction can't touch at the same price. The pattern that works in practice:

• Run organic distribution wide and cheap to build the top of the funnel: brand awareness, App Store search volume, and a base of content that keeps circulating after the campaign ends.
• Retarget the warm audience with paid. Someone who watched a creator's video about your app and searched for it in the App Store a week later is a far cheaper paid conversion than someone who's never heard of you, because branded search and retargeting CPMs are a fraction of cold prospecting CPMs.
• Use Apple Search Ads on your own brand terms as the closer. Brand-term taps usually cost far less than category terms because you are the most relevant advertiser for your own name, and this is the channel that catches the search demand the organic content created.
• Keep the cap disciplined. A per-post view cap keeps a single viral clip from blowing a budget the same way a runaway CPC campaign can. FindClout's model pays creators up to a fixed view cap per post and lets overdelivery run free for the brand, which is the organic-channel equivalent of a hard CPA ceiling.

For app marketers building this out end to end, our app launch guide and Snapchat CPM vs meme page CPM breakdown go deeper on sequencing an organic-first launch against paid capture.

Frequently Asked Questions

What is a good cost per install in 2026?

Good is relative to what a user is worth. For orientation, public 2025 data put casual games at $1.41 on iOS and $0.14 on Android, and casino games above $20 on iOS; most non-gaming consumer apps land between those, higher on iOS, and finance, fintech and subscription apps pay toward the top because their customers are worth more. The only real test is your own LTV-to-CAC ratio: a $6 CPI is cheap for an app whose users are worth $60 and expensive for one whose users are worth $15.

Is clipping cheaper than Snapchat or Meta app-install ads?

Per view, yes, usually by a wide margin, because a clipping network is buying attention inside content someone chose to watch rather than buying an ad slot inside someone else's auction. Per install it depends entirely on the view-to-install rate a specific piece of content gets, which varies far more than paid CPI does. Clipping isn't a replacement for a campaign with tight cost-per-action targets; it's a way to get a large volume of cheap, brand-controlled impressions in front of a real audience, some of which convert.

Why is iOS CPI higher than Android CPI?

Three reasons stack together: iOS users historically spend more, so advertisers bid more aggressively for them; App Tracking Transparency removed the IDFA for most users, so targeting on iOS is blunter and less efficient than on Android, which pushes up cost per converted install; and SKAdNetwork's delayed, bucketed reporting makes it harder for algorithms to optimize spend in real time, which tends to raise the effective price of a confirmed install.

Do organic installs retain better than paid installs?

Often, though little public data isolates it: a user who installed because a creator they already watch made the app look interesting tends to be a more intentional install than one who tapped an ad mid-scroll. That said, retention depends heavily on onboarding and product fit regardless of channel, and a badly targeted organic push (wrong audience, wrong geo) can retain just as poorly as a badly targeted ad campaign.

What conversion rate should I use in the calculator above?

Your own measured rate, once you have one. Before that, run the calculator at 0.01%, 0.05% and 0.10% and compare each result with your paid CPI. If the campaign only beats paid at the optimistic end, treat it as an awareness spend and measure the lift, rather than budgeting it as user acquisition. Content that shows the app working (a gameplay clip, a trade, a before and after) will usually convert better than a placement that only shows the name.

Does a CPM ceiling protect brands from runaway spend?

Yes. A CPM ceiling with a guaranteed floor means the brand pays a fixed maximum per thousand views up to a per-post view cap, and anything the post does beyond that cap is delivered free. That's structurally different from a paid auction, where cost can climb with demand and there's no equivalent free-overdelivery mechanic.

Why don't clipping networks quote a CPI directly?

Because view-to-install conversion is a function of the specific content, audience and CTA, not the network. Quoting a CPI without controlling for those variables would be a guess dressed up as a benchmark, which is exactly the problem with a lot of the "CPI benchmark" content already out there.

FindClout is a curated creator distribution network that has generated 3.3B+ views for brands across sports, prediction markets, AI, fintech and more. Reach us at [email protected] or book a call. Brands can start a campaign at findclout.com/advertise; creators can apply at findclout.com/join.