It all started with a Nokia 6110 and Snake. A generation of teenagers trying to beat their own high score, buttons 2,4,6 & 8 considerably more worn than the rest. 14 years later and mobile gaming is a multi-million pound industry in its own right, tipped to end console gaming as we know it.
In an age where there is a new YouTube video going viral every week, the same can be said for those mobile games which capture a global audience seemingly overnight. It is well documented that in just 5 weeks, the popular game Draw Something had gained over 12 million active users and amassed more than 20 million downloads. The initial interest was via a small marketing campaign but the largest increases in downloads (and to an extent the reason for the games longevity) was the result of to a huge ‘word of mouth’ surge emphasized by unofficial Facebook and Instagram ‘advertising’ from users posting their own drawings to the social networking sites.
Given the potential audience, it’s no wonder mobile games are seen as prime real estate for advertisers and a potential quick source of income for developers. At RealityMine, we like to look a little deeper into the way people use their phones, and analysing gamer patterns on mobile devices can produce some fascinating insights.
The treemap above shows the average session duration of the top games played in May by a sample of consumers. The larger the rectangle, the higher the average duration. The colour of each rectangle is indicative of the number of unique users of each game; the darker the colour, the more unique users.
Typically you would expect the average duration to be somewhat dependant on the type of game, so you would presume a Sudoku session would take longer than one turn on Word Feud (a 2 player ‘Scrabble-type’ game). This is demonstrated in the treemap as Word Feud has the lowest average duration at 3 minutes. The most popular games such as Angry Birds™ and Candy Crush Saga have average session durations which are a lot greater than the time it takes for ‘one go’ on the game and break the expected trend. For example, one level on Candy Crush Sage may only take 2-3 minutes to play, yet the average duration is 15 minutes. This is a good example of an addictive game and an accurate way of measuring the games appeal to consumers.
It is interesting to note that Word Feud has the highest number of unique gamers indicating that is has the biggest fan base (50% of the panel). For this reason it is important to remember that these measures should be used in conjunction when assessing popularity. From an advertiser’s point of view, the treemap tells us that Candy Crush Saga and Subway Surf would be the best apps to advertise through as they have two of the three highest audiences, while still occupying a prolonged session of screen time for the consumer.
The very nature of mobile games makes them more appealing than console games to an advertiser as most mobile games are relatively inexpensive compared to their console counterparts. This has led to an acceptance by the mobile gamer to expect adverts as part of the bargain. The consumer reasons, “I get a free/cheap game and in return you get to serve me some ads”. This would be less acceptable if you had paid £50 for a game and were constantly interrupted by adverts and is part of the reason why console in-game advertising is usually more subtle. FIFA 13™ uses the pitch-side advertising boards for product placement in exactly the same way that its real world cousin does, to avoid annoying the gamer.
At RealityMine we have looked into daily gaming patterns on mobile devices to demonstrate the key times throughout a typical day when games are being played most.
The top graph above shows the split in the duration of game use for each hour of the day.
The graph demonstrates that the majority of gaming is done between 6pm and 10pm, with a substantial amount of gaming between 7am and 10am. People therefore (and somewhat expectedly) spend the most time gaming outside of work hours, when they are relaxing at home and to a lesser extent on their morning commute.
There is a small peak at 1pm when people are on lunch at work, but this is noticeably less than the other peak time in the day. This could possibly show that people are more likely to play games when they have nothing else to do as a lunch break is a more social event than a morning commute.
The middle graph shows the total unique gamers at each hour of the day. We see a similar pattern to the percentage of game duration but there are some key conclusions we can draw from this by comparing the two sets of data. First of all, there is less of a drop of between 10am and 5pm in the bottom chart. This demonstrates that people still play games at this time, but their time spent playing is less than at the previously defined ‘peak times’.
The bottom chart shows the number of game sessions per hour throughout the day. This gives a view of how many times the games are launched. This is a particularly useful view if adverts are served as part of the app launch. On this basis, 7am would be just as valuable a time for advertisers as 9am, even though the overall game duration is larger at 9am.
There is a strange interdependency between developers and advertisers when it comes to mobile games. Developers may create two versions of their game, one free, one low cost. The free version will usually contain adverts, with the paid app promising the removal of those adverts. The advertiser uses the free app as guaranteed front of screen time with the consumer, while the developer uses those same adverts as an incentive for the user to upgrade to the paid version. For the developer it’s a win-win situation, as they are either being paid by the advertiser or the gamer. But think of the sacrifice the advertiser is making just to get adverts in that free game version. Can you imagine an advertiser agreeing to a similar relationship on any other media?
A different view of the data highlights differences between weekdays and weekends in mobile gaming.
All averages have been calculated as an average per day to compensate for having more weekday records than weekend records.
Here we see the same three measures (Total Duration, Distinct Gamers, Average Number of Sessions) split by weekend and weekdays. It is immediately apparent that there are differences in peak gaming times between the two types of day. Weekday total gaming time peaks earlier in the morning on a weekday, as people tend to wake up earlier for work. Interestingly, the bottom chart shows that the average number of game sessions per hour is almost always higher on a weekday. This could be because of interrupted use during the week as people make their commute to work. The average number of distinct gamers is also a lot higher on weekdays, meaning advertisers have a much wider audience at these times.
As we’ve tried to demonstrate, looking at patterns in this gaming data can give developers and advertisers a real insight into consumer behaviour and using it effectively can ensure that both get the best return on investment possible, either through targeted advertising or variable charging for ad space.