Sensor Tower Says 5 Million US Downloads in 22 Days, So Which Is It?

Two numbers about Muse's download performance have circulated since early October, and on their face they read like they're disagreeing with each other. Sensor Tower, via reporting from Forbes, estimated Muse crossed 5 million US downloads in its first 22 days. Hunterbrook Media's TickerTrends report, covered in Did Muse Downloads Really Fall?, says daily downloads fell 21.1% from their late-September peak by October 6. A reasonable reader could look at those two headlines and conclude someone's numbers must be wrong. They're not — they're measuring different things, and the actual lesson here is about how easy it is to manufacture a false contradiction out of two real, non-contradictory data sets.

What each number actually measures

Sensor Tower's figure is a cumulative total: all US downloads added together across Muse's first 22 days after the September 8 launch. It's also geographically scoped to the US specifically. A cumulative total, by definition, can never go down — every day adds to it, even a day with fewer new downloads than the day before. Five million downloads in 22 days averages out to roughly 227,000 downloads a day over that stretch, which is actually a useful anchor point: it's in the same neighborhood as TickerTrends' reported peak of roughly 215,000 daily downloads on September 24, which is a reasonable sanity check that the two data sets aren't wildly inconsistent with each other even though they come from different firms using different methodologies.

TickerTrends' figure, by contrast, is a daily rate tracked over time, and its scope isn't stated as being US-only — the DAU figures in that report appear to be worldwide. A daily rate can rise, peak, and fall within a window where the cumulative total kept climbing the whole time, because a slowing rate still adds a positive number to the total every single day, just a smaller one than before. Those two shapes — a cumulative total that only grows, and a daily rate that can rise and fall — aren't in tension. A download rate can decline from its peak while the cumulative count it's feeding into keeps setting new highs, for exactly the same reason a car can be decelerating while still moving forward and covering ground.

Why this matters more than the specific numbers

The temptation, when two reports about the same product produce numbers that sound like they're in conflict, is to treat it as a "which one is right" problem and pick a winner — usually whichever number confirms what you already expected to be true about the product. That's the wrong frame here, and probably in most cases like it. The right question isn't "which report is correct," it's "what is each report actually measuring, and are those measurements even comparable in the first place." Once you ask that question about Muse's downloads specifically, the apparent contradiction dissolves: a healthy cumulative US install base and a cooling worldwide daily download rate can — and, on this evidence, probably do — coexist.

The actual failure mode to watch for isn't citing either number. It's citing both numbers in the same sentence without stating that they measure different things, which is the move that manufactures a contradiction neither original report actually makes.

Doing the arithmetic check

It's worth actually running the sanity check rather than just asserting the two numbers are compatible. Five million downloads across 22 days averages to roughly 227,000 downloads a day over that full window — a number that sits just above TickerTrends' reported single-day peak of about 215,000 on September 24, which makes sense given that an average over 22 days, including the slower early days right after launch before word-of-mouth built, would be expected to land somewhat close to, not dramatically above, the eventual peak day. If Sensor Tower's average had instead come out to, say, 500,000 downloads a day, that would be a real red flag suggesting the two firms' numbers couldn't both be describing the same product in the same window. Instead, the two independently-produced figures turn out to be in roughly the same neighborhood once you do the unit conversion — which is exactly the kind of cross-check that should make you more confident in both reports, not less, even though neither firm set out to validate the other's work.

What's still genuinely uncertain

None of this resolves the open questions from the previous two articles in this series. Sensor Tower's estimate doesn't carry a disclosed financial conflict of interest the way TickerTrends' report does — covered in A Short Seller's Numbers — which is a point in its favor as a neutral data point, but Sensor Tower's methodology for estimating app downloads is itself a modeled estimate rather than a number Meta has confirmed, the same caveat that applies to essentially every third-party app-analytics figure cited anywhere in this series. Meta has not published its own download or DAU figures since the initial "topped app charts" claims at launch, which means both of these numbers are filling a gap the company itself left open, and neither should be treated as an official figure regardless of which firm produced it or what its incentives are.

There's a broader habit worth building from this specific case: whenever two numbers about the same product seem to clash, the first move should be checking their units and their scope before checking which source seems more credible. Cumulative versus rate, US-only versus worldwide, firm-estimated versus self-reported — any one of those mismatches alone is enough to produce headlines that look contradictory while describing the same underlying reality. Muse's download story happens to be a clean example of exactly that pattern, which is as much the lesson here as the specific figures themselves.

What the two reports together do support, read fairly, is a picture of Muse's launch month that's genuinely strong by raw volume — five million-plus US downloads in three weeks is a large number for any consumer app — while also showing real signs of a cooling growth rate as the initial curiosity-driven wave passed. Those aren't contradictory facts. They're the same story, told from two different angles, and the only mistake would be treating either angle as the whole picture on its own.

Part 42 of 46 in the Meta Muse series. Previously: A Short Seller's Numbers: Why Hunterbrook's Disclosure Matters Here. Next: The Personal-Messages Complaint: What Is Confirmed and What Isnt. Browse more builder insights or explore AI skills at aiskill.market.