Building a UA Strategy Around 'Affinity'

In this guest post by the CEO of the mobile game insights and market data specialist GameRefinery, Markus Ramark considers a new way to build your targeting, and looks at why understanding player ‘affinity’ with given games might unlock new levels of success.

 

At its heart, the mobile games industry is a multi-billion dollar battle to find the highest spending, frequently playing, most loyal new players. Top-performing games are supported with marketing spend in the tens of millions, and app stores are full to overflowing with competing games. It’s no wonder that the twin challenges of discovery and user acquisition are do-or-die for most publishers.

Better targeting and efficiencies of scale have helped to keep acquisition costs from spiralling out of control, and have led to programmatic advertising – the automation of the buying and selling of ad space – becoming the dominant form of mobile advertising. The ability to market games in a way that’s affordable and at scale is also a key reason why the F2P model exists, as it only works if the price is right. 

With many mobile games, it only takes a few minutes to see the problem with the current state of programmatic, with a high volume of poorly targeted, intrusive mobile ads. Improvements in ad creative and the introduction of formats such as playables and rewarded video has somewhat changed the ad experience for consumers, but it remains far from perfect.

But what if you could take a mobile game, deconstruct its elements and build your targeting around an audience that finds those very same elements enjoyable in other games? User acquisition teams could then focus their efforts on players that would find their game not just appealing to download but more importantly enjoyable to play. That’s what we at GameRefinery believe should be the future of user acquisition in mobile games. It’s a model based around player affinity to a game rather than just targeting around players’ personal data and behaviour.

GameRefinery is all about understanding what makes a successful mobile game. A chart-topping game obviously needs to have great design and really compelling gameplay, but it also needs to have elements that make it appealing to many different kinds of players. Our platform has been built to automatically analyse games and categorise hundreds of different features and design elements, so it’s a logical next step to use this methodology as a way to use affinity as a way of better marketing mobile games.

 

This data from Chartboost shows that install costs for mobile games have stayed largely static – but even a CPI of a few dollars is a significant marketing cost when scaled across thousands or millions of players

 

Behind all the industry jargon of ARPPU and LTV, we must remember that the vast majority of consumers make choices based on their affinity to a given product, which means there is a personal and emotional element to their preferences. 

Take the example of Netflix’s much-vaunted recommendation algorithm. This uses a viewer’s affinity to certain themes and genres, tailoring hundreds of recommendations to each individual user. Netflix is able to do this because it has built a detailed picture of individual viewing habits, and used this to construct a granular breakdown of sub-sub-genres of movies using more than 250 data points. That’s why you’ll get hyper-specific suggestions such as “Suspenseful movies starring Denzel Washington” if, for example, you are a fan of films like The Book of Eli or The Equalizer.

This is something that current programmatic campaigns are not designed around. Game publishers know very well that finding the right games for their ad spend has had a major impact on UA performance – the mobile gaming equivalent of finding a needle in a haystack. 

The challenge is that it’s hard to find and target your ads to high-affinity games. This list of programmatic targeting parameters – courtesy of programmatic platform Pocketmath – illustrates this point quite nicely. Whilst the targeting options are granular, they are something of a blunt instrument, as they are designed more around historic user behaviours and manual targeting for similarity.

 

This list of programmatic targeting parameters from Pocketmath shows the typical options for finding similar mobile gamers

 

There are a number of key reasons why targeting based on affinity is the key to finding higher quality players. Firstly, our research has shown that high-affinity scores correlate with better user retention, which is a good indicator of high LTV users. Or, to put it more simply, offering a consumer another game that shares an affinity to something they already know and like means they will likely play the new game for longer, and be willing to spend more whilst they play.

Secondly, using affinity as a means of more tightly targeting an ad campaign to higher quality players means that publishers naturally see a better return on their ad spend. This benefits both the ad network and the publisher, as less waste means more budget can be channelled towards those high-affinity games that will generate the best response. 

Third, this improved targeting means that those games with the highest affinity will naturally attract a higher eCPM, due to the higher bids they will attract from publishers. For the ad networks, this means more revenue even if the conversions stay the same.

 

Using an affinity approach to targeting means including titles that have similar meta, art style or core gameplay similar to AdVenture Capitalist

 

Take the example above of the F2P idler, AdVenture Capitalist. With normal targeting we would see adverts targeted at other games from the same genre or category. But by taking a deeper look at the underlying game design and themes, we can build a picture of other games in different genres and categories which nevertheless have a strong affinity with the game to be promoted. For example, a game meta that is about base building and construction opens many more potential matches beyond just tycoon-style games.

It’s early days, but game publishers we have worked with on initial test campaigns have reported significantly better results by using our platform. New users acquired through this affinity-based approach showed a 30 per cent increase in day one retention and over 60 per cent at day seven. The biggest gain was in the lifetime value of these new users, which more than trebled – proof enough that this approach has merit in conjunction with existing programmatic targeting. 

Whilst it is certainly possible for a publisher to do their own manual targeting for similar titles, the challenge becomes defining ‘similarity’ in a more quantitative and ‘scientific’ way than just based on their own judgement. That’s why we think there is clearly a need for a solution that can streamline this targeting process, and most importantly base it around a consistent and scalable dataset that has the potential to become industry-wide. Ideally, we would then see programmatic platforms using this affinity data feed alongside existing data sets.

You may wonder why the industry needs another targeting methodology on top of the wealth of data already in use? The reason is that there are plenty of warning signs that existing approaches to programmatic targeting may be breaching GDPR, which could well cause massive upheaval to the advertising industry, and by extension, mobile games companies around the world. 

There has been an element of uncertainty since the introduction of the GDPR legislation in Europe in 2018, but in recent months that there has been a noticeable shift in sentiment. The UK’s ICO recently warned companies about ongoing issues with the way in which consumer data is being obtained and used, and the US Congress has spent much of the second half of 2019 looking into data privacy in the light of recent Facebook data breaches. Plus, California’s own privacy law, the California Consumer Privacy Act (CCPA) comes into force in January 2020. 

So there is the potential for the restrictions around data sharing and security to become a significant problem for game developers sooner than they expect. The core problem is around how programmatic platforms handle personal data, which under GDPR requires a user to opt-in whenever their personal data is used for a specific purpose. If the legislation on GDPR keeps tightening up, the majority of the ad targeting business will likely move to contextual methods.

The result should be that you get higher spending, longer-lasting players – without using any personal data that could be subject to GDPR.

Even if the worst case scenario with GDPR doesn’t happen, there is massive untapped potential for affinity-based targeting as part of programmatic campaigns. It’s perfectly feasible for programmatic platforms or publishers themselves using contextual data alongside regular programmatic audience targeting. 

In the coming months, we plan to share the results of the test campaigns we have been running with a programmatic partner, with the ultimate goal of our affinity model becoming part of any mobile publishers’ UA strategy.