App Discovery and Ratings

  Pond Scum  Texture X (CC-BY 2.0)
Pond Scum Texture X (CC-BY 2.0)

(Note that this is from the perspective of a consumer. It's probably also far more rambly than it needs to be.)


One of the common problems of app delivery stores is discovery. Particularly in today's mobile app stores, the sheer volume means that discovery tends toward to drift towards the "scumbag" end of the scale very quickly in much the same manner as SEO does with web page ranking algorithms.

Since most of my media intake looks at the iOS ecosystem, I hear (and experience) a lot of complaints about Apple not minding the store for both iOS and the Mac, but some searching indicates that this is a problem that Android has as well. The Windows store might also suffer, but it has bigger market share issues.

Is it actually a problem?

Charles Perry (one half of the Release Notes podcast) made a blog post about the App Store not being a marketing tool, but in fact being a delivery truck (with a followup post), responsible for nothing more than content delivery and payment processing. I thought this made a lot of sense. At the scale that mobile app stores operate, the idea of good discovery and marketing for anyone, let alone indie developers, doesn't seem to work, at least in the info-sparse modes that stores operate.

What did everyone do before Apple launched the App Store (really it began with iTunes, but I'll talk about that further on)? Well, we relied on recommendations and search engine results. Companies with big advertising budgets sold lots and some people sold next to nothing. Sounds familiar, right?

Content Discovery vs Application Discovery

There are platforms that have the discovery problem down to a non-traumatic level. There's the music side to iTunes: we've had music genres forever, and even with all the kids-these-days new genres and sub-genres, I can generally find things I like based on other things I like. Netflix recommendations (related: the Netflix Prize, and associated story on not using the prize winning algorithm) are actually pretty good, I've discovered plenty of things I like with very few things cropping up that I hated.

The one platform which I really dislike when it comes to discovery is Steam, and mostly due to it seemingly ignoring the trove of information that it gets from its Discovery Queue tool. For those not familiar with this, Steam gives you a daily (or more if you ask for it) queue of games and software to browse though. Each game has a list of user tags for genre or features, and you have the option to follow the game in your activity feed, add it to your wishlist, or hide it in future (obviously as well as just buying it). The problem seems to be the criteria for showing something in the queue:

  • The game is popular
  • The game has high user reviews

I fill out the queue most times I launch Steam in the vague hope that one day it will actually take into account something more than that I don't want to see this version of Euro Truck Simulator. Perhaps that I never want to see another one of these Simulator games ever again. It's simple: pay attention to the user tags or your own profile of the game when I say "don't show me this" and learn from it. Learn from the fact that I tend to follow sneakers and RPGs. Stop showing me things just because people in general like them and look at what I like. It makes me sad because I feel that Steam is actually in the best possible position of all the platforms I use to actually learn from my behaviour.

At the end of the day though, all of these platforms are not in the application discovery business, they are in the content discovery business. I use applications, but I consume content. It's worthwhile to show me content which I have liked before, because I will very likely enjoy it again in the future. Once I find an application that satisfies a particular need well, I will very likely not want to replace it until it stops doing its job.

Same as it Ever Was

So application discovery sucks. It was actually always like this as markets grew, and it isn't limited to digital goods. How do we find good books to read? We rely on recommendations of people we share tastes with, curated lists, and a dose of judging by the cover. How do we discover appliances? In one of his numerous excellent podcast episodes, John Siracusa (of Hypercritical, ATP and more) and (I think) Marco Arment talked about the issues with buying a new appliance:

  • people don't buy them often and so don't get an objective opinion when recommending or criticising them
  • appliances have crazy release cycles and so most of the collections of reviews you read are actually reviewing somthing that is no longer available
  • reviews that are tied to the same model you're looking at (e.g. on Amazon or similar) suffer from the problems in the first dot point and so don't give you a good spread of criticism

So mostly we choose appliances by what's advertised to us, and shop floor attendants who probably have a vested interest in upselling you.

At the end of the day, this isn't an app discovery problem, it's a marketing problem. For all the individual success stories we see in the App Store through promotions by Apple, there are going to be many others which languish in obscurity in the "work wanted" section of the back pages, while we get another equivalent of the full page spread from the new Clash of Clans spin-off.

So what do we as consumers do? Ironically, given my general stance on ad-blocking, I think the answer lies in higher quality advertising in niche markets, and not the carpet bombing of lowest common denominator rubbish that we tend to get subjected to. Then as people who actually want quality software we need to make it worthwhile to do better marketing by being prepared to spend money on software. (Tangentially, this would be made easier by Apple making trial software and decent upgrade paths a reality on the iOS and Mac app stores.)