IPM vs CTR Playable Ads: Which Metric Actually Moves

IPM, CTR and CTI measure different stages of the same funnel, and a playable changes each of them in a different way. Here is how to read them together and decide what to change next.

By the PlayableRun team··7 min read

Key takeaways - IPM (installs per thousand impressions) is the outcome metric. CTR and CTI are the two stages that multiply into it. - A playable mostly changes the quality of the click, so it tends to show up in CTI and IPM more reliably than in raw CTR. - A rising CTR with a flat or falling IPM usually means low-intent clicks, not a better ad. - Compare variants inside one network, placement and geo, and change one thing at a time. - Use IPM to pick a winner, CTR and CTI to explain why, and retention or revenue to confirm it.

The three metrics, defined plainly

CTR (click-through rate) is clicks divided by impressions. It answers: of the people who saw this ad, how many took the action the network counts as a click?

CTI (click-to-install rate) is installs divided by clicks. It answers: of the people who clicked, how many finished the install?

IPM (installs per mille) is installs per thousand impressions. It answers: for every thousand times this ad was shown, how many installs did it produce?

They are linked. IPM is roughly CTR multiplied by CTI, scaled to a thousand impressions. That relationship is the most useful thing to hold in your head. If IPM changes, either CTR moved, CTI moved, or both did. Looking at the three together tells you which.

Why playables behave differently from video

With a video ad, the viewer watches passively and decides whether to tap. The click is a reaction to a pitch.

With a playable, the viewer has already done something. They have tried your core loop, however briefly, before they ever see the install button. That changes what a click means:

  • Someone who clicked after playing has seen a rough version of the product.
  • The click usually comes at or after the end card, not mid-ad.
  • The ad has already filtered out a share of people who would have clicked on curiosity alone.

Because of this, a playable often trades some raw click volume for better click quality. The effect is not guaranteed and varies by genre, network and placement. But it explains why judging a playable by CTR alone can mislead you.

What each design choice tends to move

Think of the playable as three zones: the first screen, the interaction, and the end card. Each zone has the strongest influence on a different metric.

First screen: the gate to everything

If people do not start interacting, nothing downstream happens. A clear first screen with an obvious first action (a hand cursor, a highlighted object, a short prompt) determines how many impressions turn into real play. This is not one of the three metrics, but many networks report an engagement or play rate alongside them. Check what yours exposes.

A weak first screen drags down every metric because it shrinks the pool of people who reach the later zones. If you localize, this is also where text matters most. See How to Localize Playable Ads: Start With Screen One.

Interaction: where CTI is mostly earned

The play section sets expectations. If it shows a satisfying slice of the real game, the people who click know what they are getting. If it shows a mechanic that does not exist in the product, some will click anyway and bounce at the store or uninstall quickly.

Changes here, such as the mechanic you demonstrate, how long play lasts and how hard the challenge is, tend to affect CTI and downstream quality more than CTR. A fail state that nudges "try again in the full game" is one example of a change that can raise intent.

End card and CTA: where CTR is mostly earned

The end card is where the click happens. Button size, wording, contrast, timing and whether the whole card is tappable all matter. Changes here are the most direct lever on CTR.

They are also the easiest way to inflate CTR without improving anything. A CTA that appears immediately, or a card that triggers the store on any tap, can raise clicks while lowering intent. That is the pattern to watch for.

Reading the three metrics together

Here are the common combinations and what they usually suggest:

  1. CTR up, CTI down, IPM flat or down. More clicks, worse quality. Look for accidental taps, an early or oversized CTA, or a playable that overpromises. This is where the risks in Misleading Playable Ads: Fake-Play vs Real-Play Risk become a performance issue, not just a policy one.
  2. CTR flat, CTI up, IPM up. The click is better qualified. This often follows an honest, more representative play section.
  3. CTR down, CTI up, IPM flat. The ad filters harder. Whether that is good depends on your downstream numbers. If retention or revenue per install improves, it may be a win even though IPM did not move.
  4. CTR up, CTI up. The best case, and worth a second look to make sure the test was clean and not just a lucky stretch of traffic.
  5. CTR fine, CTI poor, across every creative. The playable may not be the problem. Check the store page. Your screenshots, icon and description affect conversion after the click. For help with that side, see App Store Listing to Ad Creative: A Practical How-To.

Where the numbers lie to you

IPM looks clean, but several things can distort it.

Click definitions differ. Some networks count any tap on the playable as a click. Others count only taps on the install action. A CTR from one network is not directly comparable with a CTR from another. Read each network's reporting documentation before building a comparison.

Placement mix changes the result. The same playable can behave very differently in a rewarded placement than in an interstitial, or on one app inventory versus another. If one variant happens to get more of a favorable placement, its IPM will look better for reasons unrelated to the creative.

Attribution is not uniform. Install counts depend on the network's and your measurement partner's attribution rules, and these can include different windows or view-based credit. On iOS, reporting may also be delayed or aggregated. Apple documents its attribution framework on the SKAdNetwork page. Know what your numbers include before you treat a small gap as real.

Early data is noisy. A few good days do not make a winner. Wait for the difference to persist across days, and treat anything within your usual day-to-day variation as a tie.

IPM is not the goal. It is a proxy. An install that churns on day one is not worth much, so a variant that wins on IPM but loses on retention or revenue is not actually the winner.

A practical way to use all three

A workable routine for a small team:

  1. Pick IPM as the decision metric for the test, within a single network, placement type and geo.
  2. Change one thing per variant. If you change the first screen, play length and CTA all at once, you will not know which one moved CTR and which moved CTI.
  3. Log CTR and CTI beside IPM for every variant, so a result always comes with an explanation.
  4. Decide in advance what counts as a tie, so you do not rationalize a small gap afterward.
  5. Check downstream numbers such as retention or revenue before scaling a winner.

The structure of a good test, including how many variants to run and when to stop, is covered in Playable Ad Creative Testing: A Practical Framework.

Variants are cheap to make when you can generate them quickly from existing store assets. PlayableRun, for example, builds a playable from a store listing, so testing different end cards or first screens does not need a developer for each one. The tool does not replace the discipline of testing one variable at a time.

A quick example

Say you run a puzzle game and test two end cards on the same network and geo. Variant B has a much larger, brighter install button that appears as soon as the player finishes the first level.

You see CTR rise in B. CTI falls. IPM ends up about the same.

The reading: the larger button captured more taps, but many were from people who were not ready to install. The decision is not "B wins on CTR." A better follow-up is a variant that keeps the improved button but shows it after the player has seen a more rewarding moment, then checks whether CTI recovers.

That is the pattern. CTR tells you where the click happened, CTI tells you whether it was worth anything, and IPM tells you whether the net result moved.

Summary

Treat IPM as the scoreboard and CTR and CTI as the explanation. A playable mostly changes what a click is worth, so a result that looks good on one number alone deserves a second look at the other two. Keep tests clean, keep comparisons fair, and confirm with downstream quality before you scale.

If you want to see how different playable structures look before you test them, browse the live demos, or create a free account and build one from your own store listing.

Frequently asked questions

›Is IPM or CTR the better metric for judging a playable ad?

IPM is usually the better headline metric because it combines click behavior and install behavior per impression. CTR is still useful as a diagnostic, because it tells you whether the problem is before the click or after it.

›Why can a playable have a high CTR and a poor IPM?

Often the clicks are low intent. Accidental taps, a misleading mechanic, or a CTA that appears too early can produce clicks that never become installs. The store page can also be the cause if it does not match what the playable promised.

›What does CTI measure?

CTI is click-to-install rate: installs divided by clicks. It reflects how well the click was qualified and how well the store page converts the people who arrive.

›Should I compare playable metrics with video metrics?

Only with caution. Placements, click definitions and user intent differ between formats, so compare within one network, one placement type and one geo wherever you can, and use downstream metrics like retention or revenue to settle close calls.

›How much data do I need before trusting an IPM difference?

There is no universal number. Wait until differences hold steadily across days and are larger than the normal day-to-day swing you see on a single creative. Check your network's reporting for how it attributes installs.

Sources

Turn your store listing into a playable

Paste an App Store or Google Play link and get an HTML5 playable you can test in minutes. The first one is free.

Keep reading