Optimizing The Play Store For Quality

I’ve spent a lot of time thinking about the Play Store over the past couple of months. Previously, I wrote about the terrible quality of apps on the Play Store and presented the small project I started to help solve that problem. But that particular solution isn’t feasible on a scale as large as the entire Play Store.

No, on this scale we’re going to need something different. If you’re dealing with over 1 million apps, that necessitates community. And getting quality input from users requires a platform that allows it…

So, I’d like to explore some potential actionable solutions that Google could take to fix the mess that is currently the Play Store.

An Example To Start

Let’s take an example of the problem first — imagine you’re looking for a quick game to play on the way to work. So you go to the Games -> Casual category. Now you can view the various charts: ‘Top Selling’, ‘Top Grossing’, ‘Top New’, etc. So let’s take a look at a couple of them…

I’ve always wanted to play ‘Remote Control for TV’!

Above we have the Top New & Top Selling lists. So, ignoring the fact that 30% of the apps shown aren’t even games, the question is: how do we decide which one to download?

First, we’ll have to decide which ones are worth investigating further. The ratings are all essentially the same and they’re all in the ‘Casual’ category, meaning all we have to go on are the names and icons. So we’ll have to judge based on that limited information.

As a (mostly) mentally-functioning adult, I’ll have to eliminate all games that look like they’re for children (Dr. Toto, Spongebob, Farm Story, & Heart Surgery Simulator). I’ll obviously have to eliminate the remote control apps and the fake ‘Fingerprint Lock Screen’ too. That leaves me with 3 apps: Game Dev Story, Plants vs. Zombies & Stealing the Diamond. The latter I’m not sure about — the early 2000's Flash-like graphics are throwing me off. So that leaves me with ~2.5 games out of 10, where I’d even consider looking at the description. 25% is a pretty poor percentage, considering these are supposedly some of the ‘Best’ games in the casual category.

The worst bit about all this is that I can’t even refine my search to eliminate these games; there is no further sub-categorization available in the Casual category and there’s no way to filter out games that are specifically not made for my age group. I will have to scroll through endless pages of games, where 3/4 of the apps I see don’t interest me in the slightest, just to find something that I may someday even consider clicking on.

But it’s not just me — if an 8-year-old kid were to search for a game, they’d see the same results as I would. That’s a major problem right there: two totally different target markets seeing the exact same content. Expecting it to appeal to us both is ridiculous.

For the guys who invented quality search, Google are really doing something wrong here. But because we have been dealing with this for so long, we’ve grown used to just scrolling past tons of content that is totally irrelevant to what we’re looking for.

There are 3 parties at play here: the user, the developer, and Google. All of whom benefit from the user finding the content they’re looking for. The user gains ‘utility’ (Economics classes finally coming in handy) from the app, and both Google and the developer make more money. To simplify the steps:

1. Show the user what they want.
2. User spends money on content they like.
3. ?
4. Profit!

Giving The User What They Want

I said at the beginning I’d offer some potential solutions to the problem, so here are three things I think Google could focus on to improve the situation:

1. Better categorization of content.
2. Improved search (using filters).
3. A modernized review system.

That sounds pretty simplistic on the surface, but combining all three should give users the ability to see much more content that interests them. Let’s take a look at them in detail…

1. Better Categorization of Content

Well-categorized content paves the way for searchability and, given the correct tools, users can take control.

1 million apps. Less than 40 categories. Each app has to fit neatly into one category. That sounds pretty ridiculous. Could you imagine if Netflix content was categorized that way? You’d never find what you were looking for!

Taking my earlier example of searching for a casual game — If I could further refine my search by selecting ‘Casual’ & ‘One-Touch’ or ‘Casual’ & ‘Math’, I could eliminate a ton of content I’m not interested in right from the get-go.

But, of course, in order for this to work we actually have to categorize the content.

We can’t trust developers to do it (see: Remote Control apps in the casual games category above).

Google don’t have the resources to categorize over a million apps (and Google doesn’t like manual labor anyway).

The only solution is for us, the users, to categorize it. But we need the tools from Google to do so. User-created tags could get messy, especially without heavy moderation. Allowing users to tag a game as anything they want wouldn’t work, unless you want tags like ‘lol best gAMe EvA!!1’.

A good starting point would be a large set of predefined tags (created by Google).

This would involve having tags like ‘Material Design’, ‘Math Game’, ‘Word Processor’, ‘SMS App’, etc., and allowing users to select some of these tags as part of their review. The most popular selections become the actual tags of the app. This is similar to how Steam and other content discovery platforms work.

So, assuming we have well-categorized content, now we can easily discover it, right? Well, not really. If we can only browse through categories one at a time, then it doesn’t really help. What if I want a Material Design Word Processor? Just going through apps tagged ‘Material Design’ won’t help me.

Which brings us to the next point…

2. Improved Search (by ditching search and using filters!)

Filters give users the power to find what they’re looking for. Search takes it away from them and gives it all to an algorithm.

Okay, maybe don’t ditch the search, but combine it with a powerful filtering system. Let’s go back to the earlier example again. Using a filter system that gives the user more control, I could search for games tagged ‘Casual’ and ‘Math’, and exclude results where the price is too high or the age recommendation is ‘8 & below’.

Once the content is categorized, adding a filter system isn’t rocket science. It’s how virtually every shopping app/website in the world does it, and it’s how I’ve implemented it in Curated.

Search is great for the web because most content can’t be easily sorted into buckets, whereas apps are the perfect candidates for this type of categorization. They fit neatly into genres and there are plenty of defining attributes such as price, in-app-purchases, support for various devices, age recommendations, etc.

Filters give users the power to find what they’re looking for. Search takes it away from them and gives it all to an algorithm. And in this case the algorithm appears to be extremely rudimentary. Basically, whoever spams the title of their app with the most keywords comes out on top. Kind of like the web before Google came along. Ironic.

3. A Modernized Review System

If everything is rated highly, it kind of defeats the purpose. It’d be like if everyone at school got an A — great and all, but how do you know who’s the best?

This one isn’t such an easy problem to solve. The current review system clearly leaves a lot to be desired (see: Why You Shouldn’t Trust App Ratings on Google Play).

But let’s just take a look at the star system, where you can give ratings of between 1-5 stars. Next time you’re searching for something, take a good look at the ratings of all the apps. Pretty much every app you’ll see has between 3.5 & 4.5 stars, with most falling into the 4.0–4.3 range.

You might think, “Wow, that’s great, so many quality apps out there!”. But if everything is rated highly, it kind of defeats the purpose. It’d be like if everyone at school got an A — great and all, but how do you know who’s the best?

Further to that, if you read the article above, you’ll notice that most ratings are given without a review, and the quality of reviews tends to be quite poor.

There are a couple of options here, and it’s hard to tell which is better. My only recommendation would be to test a combination of the following:

• Require a Written Review with any rating — this should lower the amount of spam/low effort feedback. A rating without a review provides absolutely no meaningful data for future users. It also encourages spam reviews from sites like fiverr.com.
• Threaded Reviews System — So that users can reply to each others reviews (just like reddit, YouTube, Facebook, Amazon, Steam, and virtually every other comments system in the world). Users & developers can ask for more information about individual issues that way and problems can be resolved transparently. Right now, only developers can respond and there can be no further dialogue — the user usually ends up editing their original review to say something like ‘Fixed. Thanks’, meaning future users have no idea what the problem was or how to fix it.
• Transparent Voting — At the moment it is possible to upvote/downvote reviews, but we can’t see how many times a particular review was upvoted/downvoted, so it kind of defeats the purpose. Almost every other review system has a transparent way of doing this (reddit, YouTube, Facebook, Amazon, Steam, etc.), so why not the Play Store?
• Pros & Cons Reviews — Slant have a good approach to user reviews: they’re broken up into good points and bad points, which requires more work for the user, but could be interesting to try out.

Conclusion

I don’t know why Google are so relaxed about the poor quality of search results in the Play Store. I assume it’s because they’re making a bajillion dollars as it is, so there isn’t much incentive to optimize further.

But that seems quite short-sighted because it’s not really about the money here. It’s about solidifying your hold on the market. As a company, weaknesses in your product give other companies a strategic opening to fill that market gap. This is exactly how Google was formed! They saw flaws in existing products and decided to exploit them. And where are AltaVista, Yahoo Search, et. al. today? Exactly.

But maybe Google are working on this problem and it’s just that the Play Store infrastructure is so huge that pushing changes like this requires years of work. A valid reason of course. But I would contend that it’s in Google’s best interests to hurry this up. Right now, people are asking how to install the Play Store on Amazon Fire OS, but what if in the future the reverse is true? (I’m not implying that the Amazon App Store is any better, just that there’s a gap there for somebody to fill).

Another important reason is individual developer/publisher satisfaction. A large part of that satisfaction is revenues. When most companies are thinking about creating apps, they always hear the same ‘Android has more downloads, but iOS has more revenue’ story. So when you’re on a limited budget, the logical thing to do is to put more resources into developing the iOS app, thus lowering the quality of the Android app in comparison.

This kind of creates a vicious cycle — low quality apps, meaning users are less likely to pay for them, leaving less resources to put into Android development, meaning… you see where this is going.

I’m not saying that the only reason iOS revenues are higher than Android’s is because of low quality content on Android, but it’s going to be pretty hard for a user to pay for something they can’t find (because it’s drowning in a sea of ‘Remote Control for TV’ apps).

P.S. If you’re looking for something to check out in the meantime while Play Store gets its act together, I’m working on project called Curated, where I’m attempting to collect the best Android games around into one simple app.