Store Conversion Basics: Understanding Impressions, Page Visits, and Installs
When installs come in below expectations, the first fix that comes to mind is usually "let's redo the screenshots." But the problem may not be the screenshots at all. Maybe not enough people saw your listing in the first place, or they saw it and nobody tapped through to the page. Staring at the install count alone can't tell you where the repair needs to happen.
This post breaks the path to an install into three steps — impression, page visit, and install — and looks at which assets do the work at each step and how to infer where people are leaving.
Three Steps, Looked at Separately
Inside a store, the path from a user first encountering your game or app usually runs in this order.
- Impression: your icon and name appear on the user's screen in search results, a category list, or a featured section.
- Page visit: some of those people tap the icon or name and land on your detail page.
- Install: some of the people who saw the page tap the install button.
People fall away at every step. So when you look at numbers, don't stop at the install count — look separately at "of the people who saw it, how many came in?" and "of the people who came in, how many installed?" The first ratio answers "did we make them curious?" and the second answers "did we convince them?" Each store names and counts these steps differently, and the definitions change over time, so hold on to the concept and check each store's current help docs for how the actual metrics are defined.
Different Assets Work at Different Steps
The same store asset plays a different role depending on where it's mainly seen. The table below lays out what the user is looking at, and what question they're silently asking, at each step.
| Step | What the user mostly sees | What they're thinking | What you can tune |
|---|---|---|---|
| Impression | Icon, name, and rating in a list | No question yet — just scrolling past | The words in your name and short description, your category choice |
| Page visit | Icon, name, short description, rating, and on some screens the first screenshot | "Is this what I was looking for?" | Icon clarity, whether the name and short description agree, the first screenshot |
| Install | The screenshot set, preview video, the top of the description, reviews | "Is it worth installing?" | Screenshot order and captions, the first sentence of the description, review replies |
Depending on the store and the screen layout, search results sometimes show the first screenshot next to the icon and name, and sometimes only the icon and name. That's why it's safest to treat the first screenshot as an asset that works across both the visit step and the install step. How to plan that first image is covered separately in Planning Your First Screenshot.
The Impression Step: What You Can and Can't Control
Of the three steps, impressions are where you have the least direct control. Stores don't publish how they rank search results or recommendations, and nobody outside can say for certain. Anyone claiming "do this one thing and you'll rank at the top" is best taken with a grain of salt.
There's still work you can do. Check whether the words in your name and short description line up with the search terms people actually use, and whether you've picked the category that fits your game or app. If a hypothetical meal-logging app describes itself only as a "wellness platform," it's cutting off its own chance to meet the people searching for "diet" or "calories." For tips on splitting the work across text fields, see Optimizing iOS App Name, Subtitle, and Keyword Fields.
The Visit Step: Stopping a Thumb with Very Little Information
At the visit step, the user isn't looking at your page — they're looking at a list. Similar icons sit in a row beside yours, and the reason to tap yours has to come from the icon, the name, and a short line of description. Three problems show up most often here.
- You can't guess the genre or purpose from the icon alone
- The name gives no clue about what the app does, and the short description doesn't fill the gap
- The color and composition are nearly identical to the other icons in the same results, so yours gets lost
The check is simple. Search for your main keyword yourself, capture the results screen, cover your icon with a finger while you look at the ones around it, then lift your finger. Does yours catch the eye first? Can you tell right away what it is? Just that much usually points you in the right direction.
The Install Step: Giving People Who Came In a Reason to Commit
Landing on the page means the interest is already there. At this step, the user is trying to confirm "is this really what I expected?" So failures here mostly come from a gap between expectation and reality: an icon with a cute character but screenshots full of dense menus, or a short description that promises "lightweight" while the first screenshot is a packed strategy screen.
The screenshot set carries most of the weight, but the top of the description and the reviews get read too. Front-load the key points above the fold the way How to Write a Store Description That Actually Drives Installs describes, and check that low-star reviews aren't piling up without replies.
How to Infer Which Step Is Stuck
If you can see the ratio at each step, you can narrow down where to look from the symptom, as in the table below. The thing to compare against is not other apps' averages but your own page's previous period. Normal ranges vary so widely by genre, country, and traffic source that comparing against outside benchmarks tends to blur the picture rather than sharpen it.
| Symptom | Suspect first | Try first |
|---|---|---|
| Few impressions to begin with | The words in your name and short description, your category | Write a list of search terms users would plausibly use and compare it against your text |
| Steady impressions, few visits | Icon, name, short description, first screenshot | Compare against competitors side by side on the search results screen |
| Plenty of visits, few installs | Screenshot set, video, top of the description, reviews | Check whether the page delivers on the expectation the icon set up |
| Installs are low only from ads or external links | A mismatch between the ad message and the store page | Put the ad's opening shot and your first screenshot side by side |
| Low only in a specific country | Local-language captions, differences from local competitors | Search for yourself in that country's store |
People arriving from external ads skip the in-store impression step and land directly on the page. Mixing traffic sources together makes it easy to hide the cause, so where you can, look at search, recommendations, and external traffic separately.
A Small Habit for Confirming Your Inference
An inference is only a hypothesis; you have to make the change and check the result to know whether it was right. If you change several things at once, you can't tell which one made the difference, so change one thing at a time — if you swapped the icon, leave the screenshots alone for that period. Keeping the items below in a simple table makes the next decision easier.
- The date of the change and what you changed
- Which step you were targeting (visit or install)
- Outside factors during the same period (updates, discounts, featuring, seasonal events)
- The ratio for that step before and after
For a screenshot experiment aimed at the install step, it also helps to have the comparison versions ready quickly. StoreShot lets you build several versions from one template by swapping only the captions and images, which makes it easy to prepare variants that differ in just one element. How to validate with experiments on Google Play is covered in our guide to Google Play store listing experiments.
Wrapping Up
An install count is the result of passing through all three steps. When the number disappoints, first split it up to see where people dropped off, then start with the assets that work at that step. Just having the map — text and category for impressions, icon and name for visits, screenshots and description for installs — goes a long way toward keeping you from fixing the wrong thing.