Pixel Envy

Written by Nick Heer.

Regarding Snappli’s Claim That Only 4% of Their iPhone Userbase Are Using Apple’s Maps

This morning, Snappli, a service I’d never before heard of, posted to their blog that the number of iOS 6 devices using Apple’s new Maps has dropped to absurdly low levels:

  • 64% of Snappli users have migrated to iOS 6 within the last few weeks (UK and US)
  • Before the upgrade to iOS 6, 25% of Snappli users were viewing Google Maps at least once a day
  • Once they moved to iOS 6, that immediately went to 35% of users using Apple Maps
  • However, over the next 5 days that drops down to 4%
  • Summary: before iOS 6 1 in 4 people were using Google maps at least once a day. After iOS 6: 1 in 25 using Apple maps and falling.

A drop from 35% of users to 4% would indeed be a massive blow. But that number seems fishy. For one thing, it’s only representative of people who use the Snappli service, which is a sort of proxy compression utility that reroutes all your traffic through Snappli’s servers. They promise they won’t peek, so why wouldn’t you trust them with your credit card number, date of birth, and nudie pictures?

But later today, competing service Onavo posted an interesting article (hat tip to The Beard wherein they note that Apple’s vector tiles in the new iOS Maps app use far less data than the static images of yore:

Our data experts performed an identical series of activities on Google Maps and Apple Maps that included searching for several US cities, addresses and airports and zooming in and out to locate specific locations. On Google Maps, the average data loaded from the cellular network for each step was 1.3MB. Apple Maps came in at 271KB – that’s approximately 80% less data! On some actions, such as zooming in to see a particular intersection, Apple Maps’ efficiency advantage edged close to 7X.

The pedants at Hacker News would probably prefer if I clarified that by Google Maps, they mean the iOS 5 Maps app. Google has had similar vector tiles for a while now, but Apple wasn’t allowed to use them.

At any rate, this suggests that the new maps app draws far less data than the old one. It gets better, too. According to Guardian writer Charles Arthur1, the new app is astonishingly good at caching data:

That was certainly my experience earlier this week in Korea, where on a visit to Samsung I had an iPhone running iOS 6 which had no data contract. One evening I looked at the overview of North and South Korea (it turns out Apple’s Maps offer more detail than Google does for North Korea; the latter’s is just a white blank). The next day, with zero data coverage, we were taken on a coach trip to a Samsung production facility.

The phone tracked our entire journey, with street-level data including the names of shops, all the way. And all the way back. And then, later, out to the airport. All that, without getting a single extra drop of data.

Reduced data does indeed seem to explain a precipitous drop in apparent maps data usage as recorded by Snappli. John Gruber seems to agree that this makes more sense.

But this doesn’t explain the wording of Snappli’s post. See, they claim that they’re measuring app usage, not app data usage. And in this case, that minor phrasing makes a big difference. Note these three bullet points:

  • Before the upgrade to iOS 6, 25% of Snappli users were viewing Google Maps at least once a day
  • Once they moved to iOS 6, that immediately went to 35% of users using Apple Maps
  • However, over the next 5 days that drops down to 4%

Due to the word choice, I am left with the impression that they’re measuring the number of users opening up Maps on their iPhones. But if that is the case and they’re able to isolate individual launches of the app, that seems like a glaring abuse of reasonable expectations of privacy.

Let us assume, however, that they’re measuring data usage in the aggregate. This doesn’t explain the sharp increase in use followed by a massive decline. What might explain this is a vast user base trying 3D maps for the first time, which uses much more data than street maps (or even satellite imagery). But they don’t say that.

If Snappli is measuring data usage in a bulk form, that means that their wording is off in a big way. If, however, they are measuring the app usage in a more finely-grained way, that seems like a gross violation of privacy expectations.

  1. I really should start an offshoot of this blog featuring all the people I mention with two first names. ↩︎