How Uptika measures every number

What does Uptika measure, and where does it come from?

Uptika measures what the stores show publicly, plus the numbers you connect yourself. Nothing is guessed: when a reading fails, that day stays empty instead of being filled with an estimate.

Every storefront (one store in one country, for example Google Play Brazil) has its own readings, charts and score. Your own installs, sources and listing conversion come from App Store Connect and Google Play Console when you connect them; those reports arrive with the delay the stores have.

How is keyword popularity measured?

Uptika measures popularity by how early the store's own autocomplete offers a search while it is being typed. A search people type often is suggested after one or two letters; a rare one only appears when it is nearly complete, or never.

1. Uptika types the term the way a person would: every letter of the first 8 characters, then word by word, then the whole term.
2. After each prefix it reads the store's suggestions for that country: App Store search hints (up to 10) or Google Play suggestions (up to 5).
3. The term counts as offered when a suggestion is the term itself, or a longer search that starts with it: "visa" is offered when "visa appointment" is suggested.
4. The first prefix where that happens sets the score.

popularity = 100 − 11 × min(letters − 1, 3) − 5 × max(letters − 4, 0) − slot penalty − 4 if only a longer search was offered

• letters: characters typed when the term was first offered.
• slot penalty: the position in the suggestion list (0 = top) × 2 on the App Store, × 4 on Google Play.
• An offered term never scores below 2. A term that is never offered scores 1.

Popularity is a relative scale, not a search count: neither store publishes organic search volumes. Uptika keeps the raw reading (letters typed, slot, exact or longer), so the scale can be re-tuned without reading the stores again. App Store hints are read in 28 storefronts: US, GB, TR, DE, FR, ES, IT, BR, MX, CA, AU, NL, SE, JP, KR, SA, AE, IN, ID, PT, AR, CO, CL, PL, RU, AZ, UA and EG.

How is keyword difficulty measured?

Difficulty says how hard it is to pass the apps in today's top 10 for a search. It combines how established those apps are, measured by their rating counts, with how many of them carry the search in their title.

difficulty = 100 × (0.7 × average strength + 0.3 × title share)

• strength of one app = min(1, log10(rating count + 1) ÷ 5.5). About 316,000 ratings or more is full strength; 10,000 ratings ≈ 0.73; 1,000 ≈ 0.55; 100 ≈ 0.36.
• title share = the share of the top 10 whose title contains every word of the search.
• The result is clamped to 1–100 and recomputed with every daily rank reading.

How is search visibility measured?

Visibility is the popularity-weighted share of taps your positions would get across the keywords you track in one storefront, on a 0–100 scale. A visibility of 100 means you are first for every tracked keyword.

visibility = 100 × Σ (popularity × weight of position) ÷ Σ popularity

A keyword without a popularity reading counts as 10. The weights are a fixed model of how taps fall off down a results list, not measured tap data. Example: three keywords with popularity 80 at #2, 50 at #12 and 20 not ranked give (48 + 3 + 0) ÷ 150 = 34.

How is the ASO score calculated?

The ASO score rates one app in one storefront, from 0 to 100, as the weighted average of nine parts. Each part is scored from your own numbers; a part without data is left out and the remaining weights are re-normalised, so a missing number never counts as zero.

Average rating maps in steps: 4.6 and up = 100, 4.3 = 85, 4.0 = 70, 3.5 = 45, lower = 20.

Which keywords count?

Only keywords that fit the app count toward the score. Uptika labels every tracked keyword as fitting, your own brand, a rival's brand (people searching for another app by name), or another kind of app (fewer than 30% of the top 10 are in your category). Rival brands and other-intent keywords stay visible in the panel but are left out of the score. Automatic labels are suggestions until you confirm them.

What do the grades mean?

How is the Uptika score calculated?

The Uptika score rates the whole app rather than one storefront: it averages the ASO scores of all the app's storefronts and adds business signals, again leaving out any part without data.

What does Uptika not do?

• It does not invent search volumes or download counts that the stores do not publish. Rival App Store downloads are not public, so Uptika shows their new ratings per day instead.
• It does not fill a failed reading with a guess. The day stays empty, and any estimate in the panel is labeled as one.
• It reads public results for a country storefront without a signed-in user. What one person sees on their own phone can differ slightly because of personalisation, device and the paid ad slot at the top of App Store search.

Where can I read more?

• What is ASO? covers the basics and the differences between the two stores.
• How keyword rank tracking works explains the readings in more depth.
• ASO glossary defines every term used on this page.
• How to measure ASO impact shows how to use these numbers in a clean before-and-after comparison.
• AI search and app recommendations covers discovery in AI assistants, which has no store rank to measure.

Questions

Sources

1. iTunes Search API (Apple)
2. App Store search (Apple Developer)
3. Get discovered on Google Play search (Play Console Help)