Rankgrip: an open Google Play keyword metric

Rankgrip is a 0–100 score for how tightly a Google Play keyword is already held. Zero is an open field. One hundred is a door that will not move whatever you write in your listing.

It exists because app store optimization has a step most keyword research leaves out. Finding phrases people search on the Play Store is the easy half; working out which of them you could realistically take is the half that decides where a week of work goes.

It runs in the direction of difficulty, like the keyword-difficulty scores you already know from web SEO, and for the same reason. "How big is this term" is the question every tool answers. "Can I have it" is the one that decides your week.

Rankgrip is open. Every report on this site scores every keyword with it, and this page is the whole definition — the four factors, their weights, the band cuts and the arithmetic. Nothing is held back and no part of it is proprietary. That is deliberate, and it is the main thing separating Rankgrip from the scores it resembles: Domain Authority, Domain Rating and Trust Flow are all numbers you are asked to trust without being shown how they are made. A score you cannot check is a score you should not trust, including this one.

The three bands

Rankgrip	Band	What it means
0–29	Open	Nothing structural is in the way. A good listing can take this.
30–69	Tight	A real fight, and one a better listing can still win.
70–100	Held	No rewrite moves this. Spend the week elsewhere.

In the August 2026 puzzle report those bands fall exactly where the evidence does: the sixty-one open terms score between 5 and 41, and the eight wall terms between 47 and 80. The two groups were selected by different rules and they do not overlap, which is the closest thing to a check this method has.

The four factors

Rankgrip is a weighted sum of four separate ways a keyword can be held. Each is counted from observed Play rankings, not estimated.

Factor	Weight	What it counts
Slot share	40%	How many of the top ten belong to a developer big enough that outranking them is not a listing problem.
Title lock	25%	How many apps put the phrase in their app name, capped at three.
Maturity	20%	How many of the leaders are established, as against recently arrived.
Tenure	15%	How long the current number one has held the top slot, capped at five years.

Slot share carries the most weight because it is the barrier nothing you write can touch. You do not out-rank a studio with a user-acquisition budget by writing better copy. Title lock is second because it is the strongest on-page signal in Play search and the one an incumbent cannot accidentally give up. Tenure is last, and lightest, because it is the most reversible: a long tenure is evidence of a strong position, never the cause of one.

Why the factors are added, not multiplied

Because barriers stack; they do not gate. A term can be taken from a big developer who never put the phrase in their title; it can be taken from a title match whose app is three months old and small. What cannot be taken is a term where all four are true at once.

A weighted sum says exactly that. Multiplying would say something false — that a single zero makes a keyword free — and the rankings do not support it.

Why there is no search volume column

This is the first thing people look for, and it is missing on purpose.

Google publishes no search volume for Play. No tool has it. Every number sold as "Play search volume" is inferred — from Google Ads data for web search, from autocomplete ordering, from a panel of installs — then scaled by a constant somebody chose. Those are estimates of a different thing. Call them unfalsifiable rather than wrong: nobody outside Google can check them, the vendors selling them included.

The second problem is the one Rankgrip exists to solve. Even a perfect volume number does not tell you whether to write the keyword into your listing. The highest-demand puzzle term measured in August was puzzles, and ten out of ten of its results belong to developers a small studio will not outrank this decade. Volume says chase it. Rankgrip says 75, held. Ranking a report by volume would put that term at the top of the page and be actively harmful.

Demand still appears beside Rankgrip, as a 0–100 score that ranks terms against each other within one report. It is never presented as searches per month, because it is not.

Every filter a keyword passes first

Rankgrip scores a term only after it has survived the gate. In order:

1. It ranks apps in the category. A Play keyword carries no category of its own, so it borrows the categories of the apps holding its top ten.
2. Play's own autocomplete offers it. The demand gate, and the most important filter here. Without it the list fills with stock phrasing lifted from app listings — "their respective owners" scores well structurally and no human has ever typed it into a search box.
3. Every word is vocabulary, not a brand. Each word must appear in at least fifty different app names. A word one studio uses is a product name; a word thousands use is the market's language. This separates block fill puzzle offline from arrowscapes no ads.
4. It is at least two words. Single words on Play belong to whoever bought them years ago.
5. It matches no app or developer name, exactly.
6. Rewordings collapse. ai girlfriend puzzle game and ai girlfriend game puzzle are one row, not two.
7. A person reads what is left. Roughly a third of what survives every filter above is still a phrase no studio should chase — a passing meme, a mistranslation, something off-theme. Those are struck off by hand, with the reason recorded, so next month makes the same call.

That last step is no weakness in the method. It is an honest description of what an automated gate can and cannot know — and any score claiming it needs no such step is either not looking at its own output, or not showing you all of it.

Where the numbers come from

Every number in a report is a measurement of the live Google Play store in the country and language that report names, taken daily. Nothing is modelled from a third-party panel, estimated from web search data, or bought from a data vendor — which is the whole reason the reports can say who holds a slot and for how long, in place of guessing how many people search for it.

The limits, stated plainly instead of buried: the measurement window is young, so month-over-month movement only becomes trustworthy with two full months to compare. Coverage is deepest in US English. A keyword whose competitors have not been measured is marked unmeasured, never scored as open — an absent number is not a good one.

Who made this, and why

Rankgrip was created by Othmane Ettaib and first published in September 2026 by Indie Core Dev, a one-person game studio in France that ships free puzzle games on Google Play.

It exists because of a specific and repeated mistake. Choosing keywords from volume-based tools, the same thing kept happening: a term would look excellent — high estimated volume, low competition score — and the listing built around it would go nowhere. Looking at who was actually ranking for those terms explained it every time. The slots were held by studios with a hundred million installs, or by apps carrying the exact phrase in their title, and no amount of description copy was going to move them. The tools were answering "how many people search this", which is a real question, and treating it as if it were "can you have it", which is a different one.

So the studio started measuring Play rankings daily — who holds which slot, for how long, at what scale — and Rankgrip is that data reduced to the one number that decides the answer. No market forecast, no traffic estimate: it measures how firmly a term is already occupied, which is the part the existing tools leave you to guess at.

The reports on this site are the studio's own keyword research, published rather than kept. There is no product being sold here and no free tier to upgrade from, which is worth saying plainly: it means the incentive to make the numbers flattering does not exist.

Using and citing Rankgrip

Rankgrip is free to use and free to cite. Quote a score, name the report it came from, and link this page so a reader can check the definition:

game lucky arrows has a Rankgrip of 5 (Indie Core Dev, Google Play keywords: puzzle games, August 2026).

The formula is published, so anyone is welcome to implement it. Rankgrip is the name Indie Core Dev uses for this metric — if you compute something different, please call it something different, so a reader who compares two numbers is comparing the same thing.

If you disagree with a weight, the argument is above and the code is in the open. That is the whole point of publishing the method, and not just the brand.