We need better ways of measuring app stores

There are app stores everywhere but how do we know which are the best ones? It's time to rethink how we measure success.
Written by Mary Branscombe, Contributor

Here's one thing I know: the sheer number of apps in an app store is a pretty useless measure of how good it is.

You have how many thousand apps? That doesn't tell me how many of them are useful, well written, actually functional, not just copies of each other, checked for malware, not pumped out from a cruddy template, going to get updated or actually useful.

In other words, how good is what's in your big number?

The democratisation of development that comes with mobile apps, umpteen mobile development platforms, and the race for big numbers have combined to make finding the good stuff hard.

Early developers on PlayBook were delighted to be in a small marketplace where buyers could find them. Early Windows Phone developers had the same advantage; reasonable size fish, tiny pond, profit; Wordament started on Windows Phone and parleyed success there into a profitable business on all the major phone platforms.

Once you get over a certain size in an app store, the size of the store isn't a useful guide. That's why phone OS makers have started touting how many of the top 40 or 50 apps on other platforms they have for their system. But your top 50 may not be my top 50 (my interest in SnapChat and Instagram would require a telescope of intergalactic accuracy to detect).

And if you have a Twitter or Tripit or FourSquare or Untapped app for your platform, is it any good? That's not just a question of how good the official app is, but — assuming a service has good APIs — whether there are good third-party clients that do interesting things in ways that take advantage of the platform. Like the live tile for My Trips, which shows the details of your next flight or the sleek Cal app that puts your whole day in an Android widget, for example.

It's hard to measure how good apps are in a store without trying out hundreds of them. But some of the measures that would actually tell us more about the value of an app store are things like usage, perceived quality and freshness.

What percentage of the apps have ever been downloaded? Is that the same across all the categories of apps in the store, suggesting a broad spread of apps, or are there only a couple of categories with popular apps, suggesting it's a one trick pony? What percentage of the apps that have been downloaded were still installed a month later and six months later? What percentage of those apps have been launched more than once and how many are getting used regularly?

What is the ratio of apps with four- and five-star reviews to apps with one and two stars? Is the quality high in all categories or just a couple? How many developers are there with multiple apps that have only low ratings, which suggests they're running poor-quality app farms?

How many abandoned apps are there? What percentage of apps have ever been updated by the developer? What percentage have been updated in the past six months? How many of the apps that have been updated by the developer have had those updates downloaded and applied by users? How many developers have created an app, abandoned it to create another one and done that again and again, pumping up the numbers with the app equivalent of tumbleweeds?

You can't get those stats for any app stores, because the owners don't collate and publish them or provide APIs for third parties to get them.

And not all stores even expose the same minimal stats for comparison; the Apple App Store and the Windows Phone store both show when an app was last updated but the Windows Store shows nothing but a copyright date.

Users don't want you tracking which apps they keep or use frequently for privacy reasons, so most usage figures come from ad networks and may not be representative of the spread of apps in a store.

But we need to start looking at more than just meaninglessly large numbers to learn anything useful about the health and success of different app stores and platforms.

Further reading

Editorial standards