FunnelHound / Guides / Keyword search volume
App Store keyword search volume: what the popularity score really means
Uncomfortable truth first: Apple publishes no search counts. Anywhere. For anyone. The only demand signal from inside the house is the Apple Search Ads popularity score, 5–100 — and every "12,400 monthly searches" you've ever seen in a tool is a model's estimate of that score wearing a confident font.
Reading the 5–100 scale without getting fooled
Three properties matter. The floor is 5, not 0 — a 5 spans everything from "nobody has ever typed this" to "a dozen people once did", so a 5 tells you almost nothing except don't bet the title on it. The scale compresses upward — behavior is logarithmic-ish, so 60→70 represents far more additional searches than 20→30, and the 90s are a different universe occupied by "instagram" and "whatsapp". It's relative, not absolute — perfect for ranking two candidate keywords against each other, useless for forecasting installs. The practitioner band worth knowing: sustainable indie targets usually live around 10–50 — real demand, beneath the giants' radar — the same popularity-versus-difficulty math the keyword tools automate.
Where the "monthly searches" numbers come from
Third-party tools take Apple's score, blend their own panels, ad data and category models, and print an absolute number. Some models are decent; none are measurements. The tell: different tools confidently disagree about the same keyword by 5× — which is fine once you treat the numbers as the score, re-costumed, and compare only within one tool. What you never do is arbitrage between tools ("tool A says 3,000, tool B says 15,000, so B's keyword is better") — that's comparing two costumes and calling it research.
The free demand checks that don't lie
- Autocomplete: type three letters in the App Store — Apple's suggestions come from real queries, per storefront. Phrase completes early = demand exists. Doesn't complete at all = every estimate for it is fan fiction.
- The popularity score itself, free in the Apple Search Ads console, no campaign needed — per country, which matters enormously for localization since the same concept scores wildly differently across storefronts.
- Your own impressions — the ground truth. Rank on a term, then watch search impressions. Rank #8 with impressions flat? The demand was never there, whatever the estimate said. This is the check that retires debates: volume estimates predict; impressions confirm.
Demand is a hypothesis until the funnel agrees
The whole game in one sentence: popularity scores and volume estimates choose which keywords to try; your own funnel decides which ones were real. A keyword that looked like a 40 and moved nothing gets replaced at week six; a modest 15 that quietly feeds installs earns its slot forever. FunnelHound sits on the confirm side — search impressions, installs and purchases from App Store Connect, milestone on the day the new keywords shipped — so the estimates stay accountable to the only counter that pays.
Estimates propose. Impressions dispose.
FunnelHound shows whether the keyword actually delivered — search impressions to installs to money, from App Store Connect.
Get FunnelHoundData notes: popularity score behavior per Apple Search Ads as of 2026; the logarithmic-ish reading and the 10–50 indie band are practitioner consensus, not Apple documentation. Third-party volume models change methodology without notice — compare within one tool only.