Game design by data

[This feature was published in the July 2013 edition of Develop magazine, which is available through your browser and on iPad.]

Good data. Bad data. Big data. In today’s connected age, data is the buzzword that’s being bandided around from engineering to politics and everything in between.
Knowing how to actually get at it and, more importantly, what you can take from it, is what’s desired.

Tracking and monitoring is something that is becoming more frequent in our culture, and it’s something that isn’t always welcomed. But, what if your journey time across a busy city on public transport could be reduced by ten per cent? That could be the difference between having to make a mad dash straight from your desk to catch the bus that will arrive in time for your connecting train. This is precisely what a team of IBM researchers did in the Ivory Coast by monitoring mobile phone data.

Recognising tangible benefits can dispel reservations about technology invading privacy. Data can be taxing. Data can be tiring. And data it can be outright addictive.

When it comes to games, pretty much anything can be tracked to some degree. Nowadays, telemetry is being used to inform game design long before a game comes out. But for developers, on the face of it, the cost and resource requirements for data collection and analysis appear too heavy for the benefits they could produce.

Whether you are creating social games, mobile games or full-fat console titles, data and analytics are becoming a bigger part of the development process, and they need to be understood.

DATA DELUGE

“So many of the developers that we speak with are finding it very difficult to skill-up and to invest the resources, time and money into really getting to grips with generating and leveraging actionable insights through analytics,” Clemens Wangerin, co-founder of analytics firm Setgo, tells Develop.

“A lot of that stems from the fact that they have other priorities, as many developers are seriously struggling with the notion of discovery for their titles in the digital world much more than they are with understanding their audience.”

Where data first started to get noticed in the games was in social space. Early adopters such as Zynga and Playfish made huge gains thanks to acting on the information they collect, and making incremental adjustments to retain more players as well as adopting mischievous methods of attracting new users. And now the focus is on mobile, though even in this area things aren’t moving as quickly as some would wish.

“As mobile games move heavily towards the freemium model, the use of analytics is becoming more and more important. VisionMobile’s recent study ‘Developer Economics 2013’ found that user analytics are used only by about 28 per cent of developers in mobile apps, which is still pretty low,” says James Kaye, director of Dimoso, an integrated marketing and PR agency that specialises in digital, mobile and social.

Kaye explains that analytics in the mobile games industry can really be split into two halves: the app stores (Apple, Google, Amazon and so on) and the data analysers, the third-party firms that supply data services to developers.

And Kaye believes that where data’s potential is going unused is with smaller mobile developers, who are in need of guidance.

“The biggest disadvantage now is not ‘design by data’ but the fact that small and medium mobile game developers may not have the skills to properly analyse and interpret the data they have in order to introduce new features or make key design changes” he says.

“Whilst many of the large and sophisticated mobile publishers may have the skills and knowhow to create ‘funnels’ and ‘events’ in Flurry Analytics [a popular free-to-use analytics tools for mobile apps] and extract maximum insight, smaller developers are still in the dark. Rather than be analysts and marketers, all they want to do is make a good game.”

MAKING DATA WORK FOR YOU

There are few developers in the mobile space or elsewhere that would disagree with Kaye’s view about wanting to produce the fun before anything else.

Mark Robinson is COO of GamesAnalytics, which aims to provide games developers with better tools to understand the types of players that are playing their game. It offers services such as Engage, which provides intelligent messaging based on scoring players in real-time into treatment groups for targeted messages to be delivered in the game. In other words, this means the aggressive players get the ammo packs, while impatient players gets boosters.

Robinson says it is all about understanding player behaviour and personalising the experience. Its service displays data to users in the form of dashboard diagrams, which show you game performance, retention and monetisation rates.

But deeper analytics are needed in order to understand why and when players are leaving, or, on the other hand, why they’re buying. Using its technology to gain an insight into players actions, GamesAnalytics claims to have increased retention rates by 50-to-70 per cent and revenues by 20-to-30 per cent for developers its worked with.

“In a recent collaboration we analysed the game economy to understand how the reward structures where affecting engagement in a deathmatch game,” Robinson explains.

“We found that aggressive players who had high death rates were being very well rewarded; whereas more expert players who had better outcomes had high defection rates as the reward system was not orientated towards this behaviour.

“By balancing the rewards mechanic across different player segments we were able to dramatically increase retention in the expert players and also monetise persistent and aggressive players at higher levels. Making sure the grind economy is well balanced against real money economy is one of the most important design features that analytics can fundamentally address.”

Wangerin’s company Setgo released Pingflux, a data analytics tools that enables you to see many events in many different contexts. For developers who use analytics, Wangerin says the advantage is that you are “listening” to what your audience does once they’ve brought the game.

“We always liken it to an entertainer who stands on stage and uses laughter and applause as a way of measuring how their material is going down with the crowd. In my mind, any game developer should be naturally curious about how people are playing their game, even if they do not intend to iterate on this particular title. It would still inform their thinking when they’re designing their next game,” he says.

While some companies focus on core game metrics, other, such as GameGenetics, are involved with player retention.

“The bottom line is that games developers nowadays need to heavily focus on increasing the cost-effectiveness of their marketing mix to survive and secure a competitive position in the long-run. This requires a thorough understanding of their game-specific tracking and conversion funnel, which is a core service offered by GameGenetics,” says the company’s head of games David Mohr.

A NUMBING EFFECT

Improvements to the overall quality of games is, or at least should be, what data analytics is about. However, criticisms levelled at the likes of Zynga and other companies that have made a name for themselves in the social and mobile space is that this is ‘game design by numbers’, readied, rinsed and repeated to unduly squeeze more cash out of the hands of consumers, even if it takes the creativity out of the craft.

We’re seeing this more of late with freemium mobile games, which Kaye says have the danger to become “cynical money machines”.

For instance, the launch of Real Racing 3 from EA saw a shift from Real Racing 2’s premium mode to freemium. The game was initially launched in Australia and New Zealand to test the payment model, which would have been driven by analytics.

However, following the game’s global launch it received damning reviews and caused uproar from it’s the most vocal players, which is a sharp reminder why the freemium model can push things too far.

“This might be true to a certain extent when it comes to actual in-game metrics. There will always be a need to balance cold, hard data with the creative decisions of game makers. However, when it comes to metrics that measure the success of external influences like marketing campaigns, smart analytics should be your top priority,” offers Mohr.

But Setgo’s Wangerin argues that this is an overly simplistic way to judge analytics, and is something that certainly isn’t reflect of the companies at the forefront of the sector: “Analytics can be utilised in so many different ways that I don’t think it’s right to say that anyone who uses them operates in a ‘developing by numbers’ way. I actually think that’s archaic thinking and does not reflect the sophistication of how the most successful developers make use of these tools.

“Your fundamental ideas and game design still need to be attractive, be fun to play and find an audience. Analytics won’t help you with that – but if you get those things right, you certainly put yourself in a good position to ‘tweak by numbers’, which means you are able to make changes with a map and not blind, as you would without the insights data will provide.”

Kaye agrees with Wangerin: “I can understand the sentiment, but would argue strongly that top earning freemium mobile games such as CSR Racing, Hay Day, Clash of Clans and Subway Surfers are very high quality titles to begin with – both in terms of their production values and their playability and design. These are games that have the X factor.”

GOD IS IN THE DATA

Too much of anything can be a bad thing. But data and analytics have much to offer developers at a time when audiences are fragmenting and flitting between platforms so readily.

Names such as Supercell and King are currently riding high on their titles that have attracted the attention of millions of players. Part of their success comes from a judicious approach to data collection.

“For those who have been able to integrate analytics and data-driven thinking into their organisations, I think we’ll continue to see a trend of utilising cluster analysis, user profiling and predictive analytics to drive commercial opportunities via direct-messaging and to help identify other ways to monetise the audience, perhaps outside the game,” says Wangerin.

“On the other end of the scale, where the more lightweight, casual users sit, I think it won’t be long before we’ll see a really simple, hosted analytics solution out there that will cover 95 per cent of the most common KPIs to developers at the click of a button, which delivers an almost instant health check of your game.”

For the analytics specialists themselves, keeping up with developers needs means adapting their offerings, as Wangerin’s Setgo has had to endure: “We launched Pingflux a while ago, which has seen some good initial uptake, but it’s a pretty crowded space now. For many developers, the ones that are free to use are almost the default choice. The important thing is to start with identifying what questions you have and then looking at which solution will help you get those answers.”

GameGenetic’s Mohr is similarly resolute about how his firm can aid developers: “Our challenge is to educate and support our clients when it comes to tracking and improving their user acquisition campaigns. We try to integrate as closely as possible with our client’s system, to ensure maximum efficiency and optimisation.”

Even as the sector itself is become more competative, ultimately, it should lead to better exploration of how data and analytics can be used to improve games.

As Robinson concludes: “There are only opportunities to understand players better. This is a great position to be in and many other sectors would love to have this close relationship with their consumers. We are in a great place to use analytics to help deliver entertaining gaming experiences, but let’s extract knowledge from the data that changes the game.”