Sections

Quality is the best strategy for app store discovery

With both algorithms and curators, Graham McAllister believes that good games are the key to the mobile app stores

So many games, but which ones to play? With around 1.5 million apps on each of Apple and Google's stores, how are users meant to find what they're looking for, or discover games that they didn't know existed?

Discovery is a big problem. To help users uncover the best content, both stores use algorithms to identify the apps that perform the best, and these selected few will surface in one of the more prominent areas of the store. Given the number of apps to be assessed, it's not a surprise that both stores take an algorithmic approach. It would take a small army of trained humans to do the same job, but it does introduce a few issues.

For example, what if the algorithm just doesn't find a best ­in­ class app? Great games could potentially lie hidden from a willing audience. The second issue is a major downfall of many algorithms: they can be gamed. This has resulted in some studios buying their way into the charts via burst marketing and investment in user acquisition. Those installs are picked up by the store algorithm, and translate to a higher chart position. This may briefly increase sales, but research shows that this is generally a poor strategy, as long-­term marketing campaigns prove to be more successful.

”What if Apple's curation team just don't get a chance to see your game? The discovery problem hasn't gone away”

To help combat the issues with algorithmic discovery, Apple recently made changes to the way games are promoted on the App Store. Many of its sections will now rely on curators to select the best content, staff who are trusted to know what good games are, and make us aware of their presence. This change has been met with praise from mobile developers, but it also has problems of its own. For example, what if Apple's curation team just don't get a chance to see your game? The discovery problem hasn't gone away. Instead it has shifted.

Is visibility on the store important anyway? In the Games section of the App Store, there are around 100 games easily visible - that is, 100 icons reachable on the front page via swiping and scrolling only. Let's call these Tier 1 Games. Going one layer deeper - via a tap to Tier 2 - leads to another 1000 games, more or less, split across 20 sections. There are also Tier 3 sections - two taps from the front page - that allow access to around 500 more games. So with relatively minimum friction, an App Store user should have access to around 1600 games across a wide selection of genres, and all should be on the high end in terms of game quality.

With around 400,000 games on the App Store, though, what if your game is one of the 398,400 games that aren't on the surface? Does this mean your game is facing lower discoverability and falling short of its commercial potential?

Getting Discovered

Being visible on the front page, or even one of the lower tiers, is one form of discovery, but is it the most important? Data from several sources, including Google, shows that around 50 per cent of users find apps via the search feature within the store. This means that all other methods - word of mouth, getting featured, reviews, being in the charts - make up the other 50 per cent combined. Being at the top of the charts or getting featured by Apple is not as important as you might think, but being discoverable by search within the App Store is a big deal.

The exact nature of how both Apple and Google return results from a search within their app stores have been closely guarded secrets, but at the I/O conference in 2012 Google revealed how its search algorithm determines which apps to show to the user. They classify searches into two types: categorical, like 'free games', and navigational, which is typically the specific name of a game. On the Google Play store, these two types of search combine make up the vast majority of installs. It's worth pointing out, however, that the structure of the Google Play store is quite different from the App Store, with far fewer apps surfaced to the user, so it's likely that search is used out of necessity rather than preference.

”The user experience begins the moment users first discover your game”

In terms of which results to return, this is mainly a metadata issue. The most important criteria, Google says, is the game's title. It needs to be clear, unique, and also creative. Good metadata is important as users are unlikely to type in the name of your game directly, and even if they know it they may spell it incorrectly. For example, on the App Store, a search for 'running' brings up Temple Run, but it also returns Sonic Dash and Minion Rush, other popular games in the endless runner genre. The technology behind this is most likely powered on Chomp, an app discovery startup that uses natural language processing and sentiment analysis that Apple acquired in 2012. It is also evident that games with high numbers of installs and good reviews will also appear higher in search results.

When it comes to choosing which reviews to display, Google will select those that are most similar to the device that you're on and the country you're in. This goes some way towards presenting the most relevant reviews, and filtering out irrelevant ones. Both Google and Apple have completely rational search strategies, showing the user the best games relevant to their search criteria.

Getting featured

Getting featured is still important. Research shows that around 10 per cent of users are installing apps based on a featured slot on the main page. Those slots are associated with positive game experiences, so games featured in them gain both visibility and status.

Google also revealed how it selects which games to feature, stating app statistics such as installs, uninstalls, long installs (an app that has been installed for months) and engagement as key. It mentioned that long-­term retention is an important criteria in deciding which games to feature, as it's likely a strong indicator of a good game. Google specifically calls out 'engagement and retention' as the key criteria for getting a game featured on the store, concluding with, "There's no substitute for user experience."

So what can developers do to create the best user experience?

First Impressions

The user experience begins the moment users first discover your game. So if this happens to be via search, then the name of the game, icon, description and screenshots all matter. A good example of using screenshots is the new title from EightyEight Games, You Must Build A Boat. In the five screenshots on the Apple App Store, the first begins by showing how you start out, then a few screenshots of gameplay, then finally a screenshot of what you're working towards. Clear, concise, perfect.

”Both the Apple and Google stores seem only to offer the bare minimum to help users find and discover content”

This is just to encourage the user to click the 'Get' or 'Buy' buttons, but once the game is installed the user experience then has to be sufficient to engage and retain the user. What do you do about that? Are you helping yourself?

If Google is telling us that its algorithms look at variables relating to engagement and retention (length of time played, number of times played, installs, uninstalls, time on device etc), then the user experience of your game is the key to getting featured. But getting users to play and enjoy your game, then to keep playing and enjoying your game, is no small task. Most games don't do this successfully, and there are many reasons why users don't engage with a game or fail to play long­-term. Developers need to ask if they have done enough to identify the potential issues that cause users to stray from the intended journey. If the answer is 'yes', then the next question they should ask is how they know that.

The move to a more curated app store is meant to benefit both developers and users, and it's a step in the right direction. Both the Apple and Google stores seem only to offer the bare minimum to help users find and discover content, and a Pfeiffer report looking into the usability, friction, and effectiveness of the app stores draws the same conclusions. Could they be further improved to help users discover the content they want? Certainly. But while we wait for those changes to happen, it's worth noting that the one constant in all of this,­ whether it's algorithms or curators, the best way to get discovered remains to make a great game.

Related stories

Keywords Studios acquires Player Research in deal worth £1.3m

Brighton-based consultation and user testing service joins burgeoning empire

By Dan Pearson

Whose Game is it Anyway?

Graham McAllister finds that too many devs don't know who they're designing for

By Graham McAllister

Latest comments (13)

Paul Johnson Managing Director / Lead code monkey, Rubicon Development3 years ago
As a mobile developer myself, I'd love to agree with this. Sadly it's just bollocks.

The best strategy, borrowing from the title, is to spend tons of money on UA.
Next best is to have strong marketing skills and get it going viral.
A good quality title with neither of those is probably going to fail.

If you want a perfect example, consider "Coin Dozer" with "Jungle Coin Falls". (There are a zillion other comparisons you can make, I just know about that one.)

The one you'd (hopefully) think the inferior is about the 100th top grossing title on the store. The other one hasn't yet paid for its title screen.

Edited 1 times. Last edit by Paul Johnson on 8th June 2015 5:42pm

5Sign inorRegisterto rate and reply
Anthony Chan3 years ago
The headline is missing "Ideally:" at the beginning. While I do appreciate the optimism, this article is definitely not doing any favors to anybody who would have decided start-up shop today. The truth is, getting to the top of the charts and getting 5* ratings is the way to get visibility. Getting visibility will in turn create traffic. Yes the game being a great game is VERY important in the long run, however visibility is the first impression.

It's like getting past the first date or the first interview. A guy with a great character and personality and long-term potential does not necessarily get the girl or the job. Games are the same. You need to get the user to download the game and try it - and hopefully play it for more than the time it will take to refund the game on Play Store for example.

Great games builds long term player base. But great marketing and advertising builds the player base that will give you the mechanism to go viral. Unfortunately all of this is needed to make a profitable mobile video game. No matter how much we hate it, it is reality that is not about to change.
1Sign inorRegisterto rate and reply
Klaus Preisinger Freelance Writing 3 years ago
How do you define quality on such a subjective matter? Does a game have the most quality when it appeals to those with the highest standards? Assuming those standards were objective. Or is quality in this context merely a sign attributed to the games making the most money, since they have the quality of appealing the most to the people spending the most?

I am sure, there is no shortage of literature professors who fail to see the quality found within bestseller lists for books. No shortage of musicians wrenching their gut when listening to the top 10. Then again, Google has algorithms and who wouldn't agree that everything is better when based on an algorithm? Especially when you could just as easily paraphrase Google's algorithm criteria as: creates more sales at our store.
1Sign inorRegisterto rate and reply
Show all comments (13)
Robin Clarke Producer, AppyNation Ltd3 years ago
If you want a perfect example, consider "Coin Dozer" with "Jungle Coin Falls". (There are a zillion other comparisons you can make, I just know about that one.)
Coin Dozer has been on the App Store since 2010 and has pretty firmly established itself as the known name brand for that type of game. Five of the top ten search results for Coin Dozer are other versions by the same developer. So yes, trying to compete with that directly (especially with a less obviously compelling proposition) is going to be an expensive uphill slog.
4Sign inorRegisterto rate and reply
Paul Johnson Managing Director / Lead code monkey, Rubicon Development3 years ago
It got in that position through expansive UA advertising. I know, in a nicely ironic twist a lot of my income comes from showing their adverts.
0Sign inorRegisterto rate and reply
Jamie Firth Video Games Production 3 years ago
UA will *aid* discoverability (like all marketing has, ever), but only quality will retain users.
High quality means you have to work less hard marketing it as word of mouth makes up for the shortfall in marketing. But there is also a risk that it will just go undiscovered unless given a nudge.
A combination of both gives you a good chance of success.

Defining quality is hard, yes, but a simple look at the consumer reviews is a good start. That's a good way to turn something that is largely subjective into an objective opinion.
3Sign inorRegisterto rate and reply
Robin Clarke Producer, AppyNation Ltd3 years ago
It got in that position through expansive UA advertising. I know, in a nicely ironic twist a lot of my income comes from showing their adverts.
It got there through being a solid free game in a market that had a few thousand games in total at the time of release. Trying to do the same thing five years later (with a less compelling proposition) isn't going to work.
1Sign inorRegisterto rate and reply
Paul Johnson Managing Director / Lead code monkey, Rubicon Development3 years ago
Trying to do the same thing five years later (with a less compelling proposition) isn't going to work
Not sure what you're hinting at with this (less compelling propostion) thing, but that's the whole point isn't it. Quality is meant to be the key factor here, not age on the store and if nothing else, coin dozer is certainly showing its five years of age.

A more modern version that has more coins, more gameplay, better physics, more progession, a lot more visual appeal on the store (imo) is doing bupkis. That's because it wasn't UA'd, it's not old, or in fact posessing any other property beyond just being a halfway decent product. According to this article, that's all it really needed to have to win.

It only takes one fail to render a theory in need of either refinement or abandonment and I've just provided one. I'm not bitter about it, that's just the way it is.
2Sign inorRegisterto rate and reply
Seb Long User Researcher, Player Research3 years ago
Disclosure: I work with Graham the author
If you want a perfect example, consider "Coin Dozer" with "Jungle Coin Falls". (There are a zillion other comparisons you can make, I just know about that one.)
Thanks for the examples - I downloaded both these games to better understand your point. Resultantly, I don't agree with you, and I hope you don't mind if I explain some reasoning here.

With Coin Dozer not having the 'pachinko' fall in JCF, and in having more control thanks to the ability to place the coins directly on the board, Coin Dozer offers a more tactical and (in my opinion) more interesting experience for the coin-push-game player.

Nice touches like CD limiting the frequency of placing coins to avoid players spamming as a tactic, and a more OTT and rewarding visual feedback for both wins, power-ups and prize drops. Having a thinner board (~8 coins wide in CD versus ~10 in JF) makes it easier to keep track of the coin interactions, and it also seemed that in JF the coins were more likely to be on top of one another, which makes it harder to again to understand and play tactically. Coins in JF are around half the actual size of those on CD (played on iPad), which contributes to this visual clarity issue also. I don't think that having 'more coins' has made JCF a better experience - quite the opposite, in fact.

Although I only played 5-10 minutes of both, the sheer amount of screen takeovers (advertising and tutorialisation) during the FTUE in JF was overwhelmingly higher in JCF than CD. Despite playing on an iPad Air 2, the framerates in JF dropped very low indeed.

Both games have quite a few issues that will affect perception of quality; from technical through understanding, usability and the core experience - exactly the issues that Graham describes - but I fear that JCF seemed to come off worse on most.
4Sign inorRegisterto rate and reply
Paul Johnson Managing Director / Lead code monkey, Rubicon Development3 years ago
I'm certainly not going to address that on a point by point basis, but the feedback is appreciated thanks. (The adverts frequency is something I thought we did better at though, so will have that checked myself.)

I'm paying the price here by using a game of my own as an example, but I don't think the above has proven my position false at all. Coin Dozer is not a premium product (I accept that you don't find mine to be either), yet it has had great success brought about by other means. Mainly "grandee" status and a very aggressive advertising regime. In fact it has had more success (measured in terms of dowloads, user score or income) than very many games that would be called "a quality title" by anyone. Because it got the actually important parts right better than those. Advertising!
0Sign inorRegisterto rate and reply
Robin Clarke Producer, AppyNation Ltd3 years ago
Paul - You seem to be assuming that conditions are static. Coin Dozer established itself when the bar was a lot lower. It is now the incumbent successful game of its type - money spent on user acquisition is to maintain that position.

Meanwhile user expectations (not to mention efficiency of F2P design) have moved on several generations. In the last few years, the heavyweight publishers like King or Zynga could have made a super-polished Coin Dozer clone and thrown money at unseating it, but they haven't. Why do you think that is? (The ones that have tried - Sega and Miniclip's versions - have substantially changed and extended the formula, hybridising the game with slots and monster battling.)

Aiming your sights at trying to 'top' what the state of the art was five years ago is not a particularly representative example of applying 'quality'.
2Sign inorRegisterto rate and reply
Paul Johnson Managing Director / Lead code monkey, Rubicon Development3 years ago
And I think you're missing my point so hard I'm sure you're doing it deliberately. Please reread it from the top without assuming I'm saying "my game's better than coin dozer. waaaah haaaa wheres the money the owe me"
0Sign inorRegisterto rate and reply
Jeremy Glazman Programmer 3 years ago
How do you define quality on such a subjective matter?
He defines it at the end of the article, quality is gauged in terms of "engagement and retention". My suspicion, however, is that there's also some metric of 'critical mass' that must be met as well before your app's metrics can be taken seriously.

If your sample size of users is too small to start with, how does that affect the algorithms? I agree with the other comments here that you still very likely need good marketing out of the gate, regardless of the 'quality' of your game.
1Sign inorRegisterto rate and reply

Sign in to contribute

Need an account? Register now.