Clipping for app installs: the CPM math vs paid UA

Snapchat's app-install objective has a reported median CPM of $23.40. Clipping campaigns run $0.50 to $2 per 1,000 organic views industry-wide. That is roughly a 15x difference in the unit price — but the units are not the same, and the honest comparison happens at cost per install, not cost per thousand. Converted properly, clipping lands somewhere between $0.33 and $3.00 per install depending almost entirely on how legible your app is in the first two seconds of a clip.

Paid user acquisition is a mature, well-instrumented machine and this article does not argue it should be replaced. It argues that most app teams are paying mature-machine prices for the top of their funnel, where a $1 CPM channel with fuzzier attribution would do the same job for a fraction of the money. Here is the arithmetic.

The unit problem, stated plainly

A paid ad CPM buys 1,000 impressions served into a feed by an ad system, in a unit that contains a store button and was targeted at people the system believes install apps. A clipping CPM buys 1,000 organic views on a clip a creator posted to their own account, where the viewer has to tap out to a bio link or remember your name and search for it.

The paid impression converts better. That is not a small caveat — it is usually a 2x to 5x advantage per unit — and any comparison that skips it is misleading. The reason clipping still wins in most models is that the price gap is larger than the conversion gap, and the creative production is inside the rate instead of billed on top.

Channel CPMs converted to cost per install

Using 2026 industry-reported figures, with illustrative impression-to-install and view-to-install rates applied consistently:

Channel	Reported CPM	Assumed unit-to-install	Implied cost per install
Snapchat (app-install objective)	$23.40 median	1.5%	$1.56
Snapchat (conversion objective)	$27.10 median	2.0%	$1.36
Meta / Facebook (blended average)	~$8.60	1.0%	$0.86
Meta (high-cost verticals)	up to $42.17	1.5%	$2.81
TikTok ads	~$3.50 avg reported	0.8%	$0.44
Clipping — weak case	$1.50	0.05%	$3.00
Clipping — mid case	$1.50	0.19%	$0.80
Clipping — strong case	$1.50	0.45%	$0.33

Two things jump out. First, TikTok ads are genuinely cheap and are the paid channel clipping competes most directly against — the reported $3.50 average CPM (other sources put the range at $4 to $13) is only a few multiples above a clipping rate, and the ad unit converts better. If your paid UA is already concentrated on TikTok and working, the arbitrage is narrow.

Second, the clipping range is enormous. The spread between $3.00 and $0.33 per install is one variable: what fraction of viewers tap out. That is a creative problem, controllable through the brief, and it is where a pilot budget should be spent learning rather than scaling.

Where each side actually wins

The row worth pausing on is creative volume. A paid UA program is bottlenecked on how many ad variants your team can produce and test. A per-view campaign inverts that: dozens of creators each make their own interpretation of the brief, and the ones that work are visible in the data within days. You are effectively running a creative tournament where you only pay for the entries that got watched.

The row that cuts the other way is attribution. Paid UA gives you an MMP dashboard and SKAN postbacks. Clipping gives you a landing page and a baseline. If your growth org cannot make a decision without deterministic attribution, that is a real organizational blocker and it will not resolve itself.

Measuring installs without deterministic attribution

Clips are organic posts on creators' own accounts. There is no click ID threading through to the store, and no campaign identifier in a SKAN postback. What works instead:

A campaign-specific landing page with a store redirect. Tap-through is countable even when the install is not, and it gives you the click-through rate variable that dominates the whole model.

Organic install lift against a held baseline. Freeze paid spend for the test window. Measure organic installs weekly against the four weeks before. Crude, but the effect size from a campaign doing millions of views is usually well outside the noise.

Branded search volume. The fastest indicator, and it moves within days of clip volume arriving. Track it in both app store search and web search.

In-app source prompt during onboarding. Self-reported and biased, but at thousands of installs it reliably separates "nothing happened" from "something happened," and it catches the viewers who searched your name rather than tapping a link.

Cohort retention comparison. The real question is not just how many installs but whether they retain. Tag installs arriving in the campaign window and compare D7 and D30 against your paid cohorts. If clipping installs retain better — which self-selected interest often produces — the CPI comparison understates the channel.

Getting the click-through rate up

Because one variable dominates the model, here is where it comes from:

• The first two seconds must show the app doing something. Not a logo, not a talking head introduction, not a problem statement. The output first, the explanation after.
• Vertical source footage. Landscape screen recordings cropped to vertical lose the interface and read as low effort. Ship 1080x1920 captures in the asset pack.
• Name the app on screen, spoken and written. A large share of installs come from viewers who search rather than tap, and they cannot search a name they did not catch.
• One capability per clip. Feature tours do not convert. The strongest clips are built on a single surprising thing the app does.
• Let creators pick the framing. Prescribing a word-for-word script produces fifty identical clips that the algorithm stops serving. Constrain claims, not creativity.

The campaign brief guide has the full structure, and how to run a clipping campaign covers approval workflow and review cadence — which matters more than most teams expect, because a slow approval queue drives creators to other campaigns faster than a low rate does.

When to use which

Use paid UA when you need a specific audience, when you are retargeting, when you need to scale spend up or down inside a week, or when your reporting requires deterministic attribution.

Use clipping when you need cheap top-of-funnel volume, when your creative pipeline is the bottleneck, when your app demos well in three seconds, or when your paid CPMs have been climbing and the auction is no longer your friend.

Use both when the app is working. They do different jobs: clipping creates awareness at organic prices, paid UA harvests it with precision. Running clipping alongside a paid program usually shows up as improved paid performance too, because a warmer audience clicks ads at better rates.

The full revenue model for a subscription app is worked end to end in growing a mobile app to $10k MRR with clipping, and the free-to-play version is in clipping for mobile games. When you want to price a pilot against your current CPI, launch a campaign on Vues — you set the rate, the budget is capped, and you pay only on tracked views.

Frequently asked questions

What is the cost per install from a clipping campaign?

It depends almost entirely on click-through rate. At a $1.50 CPM the model ranges from about $3.00 per install in a weak case to $0.33 in a strong case, with the mid case around $0.80. The variable to optimize is what fraction of viewers tap out or search for the app.

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

Per 1,000 units, yes by a wide margin — Snapchat's app-install CPM is reported at $23.40 median against $0.50–2 for clipping. Per install the gap narrows because paid ad units contain a store button and precise targeting, so they convert two to five times better per impression.

How does clipping compare to TikTok ads for app installs?

TikTok is the cheapest major paid social channel at roughly $3.50 CPM reported, so it is the closest competitor to clipping on price. If your paid UA is already concentrated on TikTok and performing, the cost arbitrage from clipping is narrower than against Snapchat or Meta.

Can you use an MMP to attribute clipping installs?

Not deterministically. Clips are organic posts with no click ID threading to the store and no campaign identifier in a SKAN postback. Most teams use a campaign landing page, organic install lift against a held baseline, branded search movement, and an in-app source prompt.

Do installs from clipping retain as well as paid installs?

Measure it rather than assume it. Tag installs arriving in the campaign window and compare D7 and D30 retention to your paid cohorts. Self-selected interest often retains better than network-targeted installs, which would mean the cost-per-install comparison understates the channel.

How much creative do you have to produce for a clipping campaign?

One asset pack of vertical source footage. The creators produce the finished clips as part of earning the CPM, which inverts the usual paid UA bottleneck where creative production capacity limits how much you can spend.