As the Games Industry turns its attention to unlocking the power within Big Data, it is worth taking a step back to contemplate what the correct vision is to truly make use of data in order to drive decision making. It raises the question of what levels of investment and commitment are required to deliver a analytics environment that gives an acceptable ROI?
The danger is that companies are tempted to collect a vast amount of data at great expense and then are not able to deliver the levels of insights and direction that make the whole effort worthwhile in the first place. On successful games, vast amounts of data can be warehoused each day to record each player’s experience. It is easy to default to collect everything – every action, every click- and store it for later ‘just in case’.
Instead of starting from the ‘data up’ let’s try starting from the ‘business down’. The key question is‘What decisions do I want to make from the data?’, and then create the data environment to support these requirements.
There are certain levels of sophistication in using data to optimise and manage games.
|1||Storage||Having access to the data to answer very specific tactical questions such as ‘how many swords were bought yesterday?’|
|2||BI Reporting||Providing dashboard reports to track key game metrics such as MAU, DAU, Retention Rates and Revenue measures. This information is game performance focused and reports on the average player.|
|3||Player Segmentation||In level 2, we will learn that the average session per user is 1.22, for example. In level 3, the focus re-orientates to player-centric information. So, how many players have only one session and not returned? How many players are engaged and socially active? How many players have potential to spend? By recognising that the game user base has a wide range of different player types with varying levels of competency and focus to their game play, it is possible to start to adjust the game design to respond better to the player types that are most important to retain and monetize. Well balanced games come from good judgement on behalf of the designers but they are also delivered by a clear understanding of the experiences that actual players are having and responding to this information.|
|4||Individual Player Experiences and Predictive Modelling||Once the broad player behaviours are established using segmentation, it is then desirable to unlock more detailed behavioural information to understand the specific reasons why players leave the game and why players buy. A more detailed level of event data is required to support this analysis and often predictive models are used to understand the combination of events that result in players being likely to leave or likely to spend. In level 2 historical performance is being reported. Now the future is being predicted so that player management can be proactive and timely.|
|5||Communications Strategy Delivery||Once a clear understanding of player behaviours has been achieved to unlock retention issues and purchase opportunities the natural outlet for this insight is to devise a contact strategy in the game and around the game that interacts with the players on the basis of their individual behaviours and playing styles. By personalising the playing experience in this way, it is possible to optimise engagement and revenues, making the game reactive to individual players.This is where the ultimate payback of your investment in analytics can be achieved.|
But the data is complex, and although the analytics objectives are easy to state, it is less easy to find the golden nuggets in the noise which make analytics actionable.
Let’s take slot machines as an example. There can be large amounts of data but actually very few descriptive dimensions. Try plotting the balance fluctuations for all active players in a slot machine game and you get a black splurge all over the chart. And this is only one variable…well two if you’re feeling generous and count time as well.
By refining the cohort of players being investigated it is possible to pick out common behaviours.
For example, looking at multiple payers only shows their common behaviours before the first payment event and their typical balance at each transaction. Now the players are forming into coherent groups which can be understood from a player psychology point of view and actioned in terms of in-game messages.
Analytics is a journey and the rewards have to outweigh the effort and investment required to deliver added value. Making the journey as direct as possible with the minimum of detours and dead ends does significantly influence the ROI delivered by analytics departments.
And so, many CEOs have a dilemma.
The cost of running a credible analytics department for a medium sized publisher is not insignificant once all software, data storage and staff costs have been accumulated.
So what are the analytics priorities that will deliver a multiple of this sum back in terms of additional game income? Improvements to game design and the implementation of targeted in-game messaging can increase revenues by 20-30%. Therefore across a game portfolio with a selection of successful and not so successful F2P games this additional income could be worth at least $1m.
But there is a lead time for a analytics departments to get up to speed and that is why organisations are now favouring a combined approach with in-house resource being complemented by external software and experience to fast track delivery and improve ROI.
There are a number of benefits to this approach:
- Access to analytics experience across a vast range of games, platforms and genres
- Take the motorway not the B road to optimised games
- Embedding analytics alongside game development as a part of the organisational culture with analytics becoming a key support to the creative team in steering the development of the player experience
- Investment is matched to delivery with some suppliers offering revenue share on additional income raised through analytics.
- Pre-defined service levels where analytics moves at the same speed as the business
- Inbuilt flexibility to focus in-house analytics on supporting the game teams with fast turnaround insights whereas the outsources supplier concentrates on implementation of the strategic plan
The ROI for this approach is typically 4-7 times ROI which looks much better on the balance sheet. It is true that games companies can suffer from a ‘not invented here’ mentality but as analytics becomes a mainstream activity within publishers and developers, strong partnerships will be built between suppliers and their clients to facilitate accelerated analytics capabilityand strong returns to the bottom line.