Fraser MacInnes is a mobile games industry professional who cut his teeth writing for Pocket Gamer. He's now working at iQU, a behavioural knowledge company working in the games sector.

You can read the first two parts of Fraser's data series on segmentation and targeting.

We hear the term 'gamer profile' banded about a lot in the industry - it's become interchangeable with sign-ups, registrations and subscribers.

Super awesome games company X has three squillion profiles in its servers.

Really?

What lies beneath?

Does it know all the platforms those profiles game on? How much they spend on DLC a year? How about something pretty simple like, how many titles they played last quarter? Or does super awesome game company X only know email addresses, locations and billing info?

Don't get me wrong, those three pieces of information are valuable, but in isolation, they don't really constitute a profile. A profile is made up of rich information that builds a picture of a gamer.

So what do profiles tell us about mobile gamers? Are they really the 'most promiscuous audience in the history of mankind' as Bing Gordon famously put it?

Drowning in games…

There's no sensible answer to that question, because it views so-called mobile gamers in a prism where they never touch or see any other platforms.

Games are much more disposable than they used to be. There are far more games and access to them is as easy as clicking a mouse, prodding a touchscreen or pressing a button.

Through an imperceptible process of osmosis, platform agnosticism has crept into every nook of gaming's fiefdoms, spreading what we play and what we play on across a much broader playing field.

Wheat from the chaff

There are two consequences of players churning through huge numbers of games.

The first is they generate a massive quantity of useless data that, while highly predictable, doesn't tell us much about those gamers' actual tastes. It's basically just noise.

The second is that over time, it creates a large number of cross reference points and opportunities to rule out certain elements of choice, creating a comprehensive profile of data that tells us a lot about a gamer's lifespan. This profile contains predictability factors across platforms, which can be used for wider pattern recognition among large groups.

What can gamer profiles do?

Imagine a big global games publisher - let's call them something fictional, like Electronic Parts. Like most large companies, Electronic Parts has a group of people responsible for high-level portfolio planning across all territories and platforms.

In the interests of efficiency and delivering what its customers want, Electronic Parts tries to understand correlations between titles and players by tracking all of the gamers that play its games on every platform.

Was gaining the extremely challenging Headshot Hero achievement in a blockbuster FPS on a home console a significant predictor of also winning the Horticulture Guru badge in a smartphone resource management game for Brazilians, Germans and Koreans?

Coincidence or causal?

If it was, does that not tell Electronic Parts something about how it should prioritise development of certain types of game in certain territories for the following year?

It really depends on who you are looking at and there are other co-dependant variables that complicate such a process. For example, gamers don't tend to stick to a single publisher, which would make Electronic Parts's data only as accurate as the correlates it can directly observe within its own pool of titles and platforms.

There are of course ways and means around this, but in a business as fast and unpredictable as the games market, it can be an expensive (to buy data), time consuming (to analyse it) and at the moment, somewhat inexact method of making decisions that carry sometimes terrifying financial ramifications.

Influence people

The last thing to consider here is social influence. Social influence is the big variable that makes gamer profiles as they are today, at times, unpredictable.

For example, you may think that your data set tells you enough to be able to target a Facebook gamer with a high degree of accuracy.

Then you discover that 80 percent of the growth of that title came from peer to peer recommendations, where individuals behaved in an unexpected way, because they were more interested in the game activity they see on their friends' wall than they were in a highly targeted ad.

Time after time

In time, social influence will shed light on some very interesting long term trends and patterns.

But in the games industry, where up until even five years ago, it was the biggest marketing budget that was the biggest factor in commanding the largest influence, we still don't have enough data (unlike, for example, the travel industry) to make the sorts of meaningful profile insights that we would like to.

The truth is the data industry will be much better at the profile part of the process when it has another five years of solid profile data to study. Until then, there are still plenty of quick wins to be had, where choosing the right profiles to focus on is crucial.

Much like segmenting and targeting, big numbers of profiles are useful, but less useful than a smaller number of more detailed profiles that help us to understand more. To use a worn out cliché - in data, whether it's segmentation, targeting or profiles, sometimes, less is more.

Part 4: User acquisition

You can follow Fraser's industry commentary on his blog, or else grab bite-size rants via Twitter.