A new report that usage of Apple’s new iOS 6 Maps is plummeting has me pretty steamed. This claim - as is so often the case in the tech universe - is based on a meritless, PR-driven study. 5,000 users of an obscure app for highly technical people is a hopelessly skewed sample of what iPhone users are like.
The app that published the data is called Snappli. It compresses your data while you’re on a cellular network to reduce your monthly usage. The app watches users’ data traffic, so it can tell which services they’re using.
The study says that usage of Apple’s Maps among the 5,000-person sample of Snappli users has dropped to 4%, where 25% of them were using Google Maps before the iOS 6 update. It’s not clear - at least from what GigaOM published - what kind of usage this measures. Presumably, it’s daily usage.
If that’s the case, the study (or the article) should have included info about whether these users had switched to another mapping service. That way, we’d know whether they’ve chosen an alternative, or they just didn’t use any maps that day. Unfortunately, it does not.
Nevertheless, GigaOM thought this was enough information to publish the headline that “Apple Maps usage plunges to 1 in 25 iOS owners.” PandoDaily followed up with a link post claiming that “only 4% of iOS users still using Apple maps.”
This is a reckless extrapolation from flimsy information.
Snappli users do not constitute a statistically meaningful sample of iOS users. Only users who are both aware of the application and concerned enough about data usage to use it are sampled. That limits the sample to a fairly geeky subset. (And it excludes people geeky enough to feel uncomfortable letting some third party track their mobile usage all the time.)
Tech companies can generate valuable insights from their data that are well worth reporting. But the vast majority of what we tech reporters see are shoddy, math-deficient marketing ploys like this. It’s always distressing to see them reported as fact.
Don’t take the Snappli numbers about Apple Maps usage at face value. They’re worthless.