rankcusp vs Appfigures

Stop overpaying for ASO tools. Faster results, clearer priorities: rankcusp shows you the 11-30 cusp band the day you connect your app, and tells you the next step.

Where the suggestion comes from

An AI suggestion or a measured move? The difference is in what stands behind the suggestion.

When a tool tells you to "add this keyword", there is one question worth asking: what is that suggestion resting on?

Suggestions from a language model read fluently and convincingly, but there may be no measurement behind them. If the model does not know what rank your app holds for the phrase, its suggestion is a guess - and a guess presented without saying so gets read as a measurement.

rankcusp's suggestions come out of measurement. For a phrase to be suggested you have to rank 11-30 for it, its popularity has to have been measurable, and the strength threshold has to be reachable. If any of the three fails, no suggestion appears.

There is a language model inside the product, but in a bounded job: processing review text and extracting phrase candidates. Everything it produces then goes through the same measurement. The model invents neither a rank nor a popularity figure.

In practice that turns into this: every suggested move carries why it was suggested - what rank you hold, what the popularity is, which field is free. There is no suggestion you cannot verify.

The only honest question in a comparison

Tools get compared by putting feature lists side by side: this one has keyword tracking, that one has rival analysis. The list says nothing, because very different work hides under the same heading.

There is one question worth asking: where does this number come from? Three answers are possible and they are not alike. It may be read straight from Apple - rank, category charts, metadata, ratings, reviews. It may be derived from Apple's signals; popularity works that way, a relative value that does not claim to be absolute. Or it may be estimated from a model: downloads, revenue, a rival's keyword field.

No tool has a private arrangement with Apple. Every one of them reads the same open sources. The difference is not access; it is coverage and honesty.

Coverage is how many countries are measured, how often, and how long the history is kept. Honesty is whether what could not be measured is presented as if it had been - and that second one decides how wrong a tool will let you be.

Figure	Where it comes from
Rank	Apple - read straight from the search result; you can count it yourself.
Rating score and count	Apple - public on the app page.
Review text	Apple - public.
Popularity	Derived from search suggestions; relative, not absolute.
Download count	Estimate - computed backwards from rank.
Revenue	Estimate - derived from the download estimate, so error twice over.
A rival's keyword field	Estimate - Apple keeps it private to the developer.

What rankcusp does differently

The whole product rests on one observation: the highest return does not come from finding new keywords, it comes from the ones you already rank 11-30 for. Apple has already associated you with those phrases; what is missing is a few places of rank, and that is usually one metadata edit away.

So the screens hand you a work list rather than a keyword list. The band is separated out, ordered by popularity, the phrases above the strength threshold are set aside, and every remaining row carries one move: which field, and what to write in it.

The second difference is the pool's fifth channel. Everyone's pool feeds from three places - search suggestions, category charts, app titles. rankcusp adds rival reviews: where users describe the product in their own words. Those words are in no tool's pool, because nobody looks there.

The third is the unit of measurement. Measurement runs per phrase, not per row: every app tracking the same phrase is read out of one search. Adding a rival does not grow your bill - what grows it is phrases multiplied by countries.

The fourth is how an unmeasured value is shown. A phrase whose popularity could not be measured carries a short dash, not a zero. Zero is a reading and people decide on readings; writing zero drops the phrase to the bottom of the list and makes it look worthless.

The 11-30 band is separated and every row becomes one metadata move.
The pool feeds from five channels; the fifth is rival reviews.
Measurement is per phrase - adding a rival costs nothing.
Search runs across 175 countries and regions; phrases fold by the country's own letter rules.
An unmeasured value stays blank; it is never written as zero.

Ten questions to ask while trialling any tool

Questions to ask during a trial. The answers should be findable in the interface or the documentation; if they are not, that is an answer too.

How many countries does it measure rank in?Apple is open in 175 countries and regions. Easy to test: try to add a country that is not on its list.
How often is rank measured?Daily or weekly? Weekly measurement can miss a metadata move's arrival entirely.
How much history is kept?Three months is not enough to see seasonality. A year lets you compare against the same period last year.
How is an unmeasured value shown?Zero or blank? That single screen tells you how the tool thinks about data.
Are downloads and revenue labelled as estimates?Apple does not publish either. Shown unlabelled, a user reads an estimate as a measurement.
Is a rival's keyword field shown?If it is, an inference is being made. Is it said to be an inference?
Does adding a rival raise the price?If it does, measurement runs per row; if not, per phrase. The second is both cheaper and more accurate.
Is there review data?Few tools extract phrases from rival reviews, and users' own words are in no other pool.
Can you export the data?The measurement history is yours and should be portable; a history that cannot move is reset when you change tools.
Is there an API or a command line?Can you wire your own measurement into your own process? It tells you whether the tool is a closed box.

If you are thinking about switching

Moving from one tool to another costs more than it looks, and most of the cost is on the data side.

The first loss is rank history: the new tool starts measuring from zero. If you can export the old one's history you can carry it; if you cannot, months of accumulated comparison points are deleted.

The second is a difference in method: two tools may normalise the same phrase slightly differently - especially if their letter-folding rules differ. Then the old and new measurements cannot be compared directly.

The third is the time of day: if one measures in the morning and the other in the evening, an artificial jump appears on the switching day. Writing that jump down as the effect of a move is the most common switching mistake.

So running both for a month during the switch is sensible. Comparing what the two show for the same phrases both measures the difference and validates the new tool.

Frequently asked

Which tool has the best data?: None of them has a private arrangement with Apple; they all read the same open sources. The difference is coverage (how many countries, how often) and honesty (labelling estimates as estimates).
How reliable are download and revenue estimates?: They vary widely by category and the error bar is usually not printed. Both are computed backwards from rank and rating count; a decision should not rest on them.
Can I see a rival's keyword field?: No. Apple keeps it private to the developer. Tools claiming to show it are inferring from rank data; the inference is sometimes right and often incomplete.
What happens to my history if I switch?: If you can export it, it moves; if not, it is deleted. Running both for a month during the switch measures the difference and validates the new tool.

Where a rank is won

The keywords you rank 11-30 for are your cheapest growth

Reaching the first page does not take winning a rank from nothing. A keyword in this band is already a few places away - and the first page is where nearly every download happens. rankcusp sifts that band out and puts what costs you least effort at the top.

The words in competitor reviews appear in no keyword tool

Users search for your app in their own words, not in your marketing language. Those words are written down in your rivals' reviews, and no tool looks there. rankcusp does: it sweeps the reviews, pulls out the phrases that look like searches, and hands you the ones that can rank.

Half of your 100-character keyword field goes to waste

Repeating a word already in the title, leaving a space after a comma, leaving country fields empty. Each burns characters quietly. rankcusp shows every one and tells you how many characters you get back.

Whether the change you made worked is clear the next day

You changed the title, you rewrote the keyword field - then what? rankcusp records your rank every day. You see, in hindsight, which move carried which keyword how many places. No guesswork, a record.

It looks not at the keyword tools everyone looks at, but at the user reviews nobody does.

Every keyword you track has its position written down each day, in each of the 175 storefronts Apple opens. After a metadata move you can look back and see which phrase moved, by how many places, and on which day - a record rather than a guess. The charts on this page carry no figures, because nothing measured them; the ones in your panel are nothing but measurements.