Tuesday, August 8, 2023

What’s the Deal With Sensor Tower?

John Gruber:

So, I see three ways Sensor Tower collects usage information for apps and websites that aren’t their own: (1) ad-blocking web browser extensions, (2) screen time monitoring apps for Android and iOS, which on iOS requires access to Screen Time, and (3) the Adblock Luna VPN. (Perhaps I’m underestimating how much data they can collect from users who play Melody Run.)

These apps may well be popular — again, they claim that Adblock Luna has been installed by over 15 million users — but is the data they collect from them representative of the general public?


The news media so badly wants to know usage data that they just accept Sensor Tower and other such firms’ pronouncements at face value, without ever describing — let alone questioning — how they ostensibly know what they claim to know about very private data.


5 Comments RSS · Twitter · Mastodon

Not sure what Gruber is trying here but he’s naive to think Sensor Tower (or data.ai/App Annie) assume their user base is the same as the broad App Store user base.

Of course it’s not, and ofc they know it. They don’t just develop a multimillion business based on sloppy math; they do all kinds of adjustments and benchmarking so that their data is reliable and worth paying money for.

Yeah it’s not like Facebook developed a multimillion ad business and sloppily told the media and advertisers that video was incredibly popular on their platform based on false metrics that said they people viewing videos when they were actually scrolling past them.

I’d take all internet and platform metrics with an unhealthy dose of salt.

Sensor Tower has the same problem any polling firm does—there is no way to do a truly random representative sample, so you always need to take your raw data and adjust it to account for the biases in your sample. This is a hard problem, never 100% correct, and often too wrong, but it's a well-understood problem.

I don't know what Sensor Tower is doing, but Gruber doesn't either—he's just assuming they're not correcting and also assuming the reporting is not questioning it. There is lots of fact checking and validation that goes into reporting that is not part of the written article.

"people viewing videos when they were actually scrolling past them"

...or Twitter, where people have millions of "views" on videos that almost nobody actually watched, because Elon forced millions of people to scroll past them.

Having said that, Twitter, Facebook, Google and other such companies are incentivized to lie, whereas Sensor Tower is presumably incentivized to provide data that's as good as they can get it.

He's trying to throw shade on the "Threads is dying" stories - except Threads related activity on Google search has also nose dived.

Leave a Comment