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 find the previous parts of Fraser's data series on segmentation, targeting, gamer profiles, user acquisition, and marketing acronyms here.

When it comes to game design, data is a bit of a dirty word. Designers see themselves as creators with a vision, not mathematicians with a formula.

It's true that there's no secret quantitative formula for making a great mobile game – or any game for that matter.

But there are certain things we can infer from data and analytics to avoid making obvious design mistakes.

Different by design

The hotspots where gamers give up, or instigate communication with another player (or fail to do so where it would improve the experience), or, indeed, spend money, are all measurable with analytics.

The right pattern recognition can help to provide a mind's eye view into the brain of a gamer.

Monetisation – there's another dirty word.

It's no secret that the games industry and gamers alike have monetisation fatigue. I've seen plenty of headlines over the last few months where developers have appealed to the industry to stop designing for monetisation models and start designing for gamers.

Guess what? They are completely right.

The numbers don't know

Data should not be used cynically.

It should not be used to exploit the wrinkles of human psychology to extract ever more cash from an audience while providing an experience that is addictive for the wrong reasons and of little entertainment value.

The big culprit here is resource management titles. You start the game and place some buildings on a luridly coloured cartoon map. You earn some in-game currency for placing those buildings. Yippee!

Your building attracts a certain type of non-player character (NPC), so you get some premium currency for attracting that NPC. Then you are asked to build something else to serve that NPC.

A completion timer is set, so you are asked to spend the in-game currency you earned not 01:30 minutes ago, on completing the building now. So you do.

Sparkly, shiny, flashy colours! Plinky plunky sounds! That feels nice.

Appointment to play

You're then asked to build something else as a result of now having that last building as a proud addition to your growing virtual world.

A completion timer is set again. Bam! You hit a brick wall and can't progress or do anything meaningful in the game until your building has finished.

Cleverly (or crassly, depending on how you look at it) the game has opened two possible options to the player.

It has created an appointment to play - which you can bet will have a timer-based notification attached to it, informed by data for its optimum effectiveness in getting the player to return to the game where they may end up spending some money after all -  while also giving you the option to pay now to keep playing.

So you - or at least a healthy percentage of you - opt to pay now to complete the building. You're motivated to do so for reasons you can't quite articulate – the endorphins that followed that sweet hit of audio/visuals that erupted after you completed your last building are starting to wear off.

MUST CLICK ON!

Lo and behold, you have almost enough in-game currency left to buy that thing that will move the timer on… but not quite. You can't buy that thing now. You have to wait, UNLESS you opt to buy some premium currency with some real cash.

Sounds familiar? Of course it does – that's the player progression for the first ten minutes of all too many so-called social games and it's no accident.

It is a quantitative process with provable, data-driven justifications for each step, each of which is leading the player inexorably towards buying in-game currency.

Rats in a monetisation maze

Without wanting to get into a debate about whether human pleasure can really be reduced to targeting and then stimulating the correct parts of the brain, it's surely important that developers treat their audience with a little more respect than they would a lab rat.

Using data for design, where that design largely consists of furtive manipulation, is just plain wrong.

Using big data to infer how gamers like to play games before optimizing for it, however, can be very useful.

A classic data-driven example of game design is that gamers in Asia seem to be unfazed with being able to buy the so-called 'win button' (that is, an item that gives a significant advantage) despite the huge imbalance this creates between paying and non-paying users of free-to-play titles.

Try the same in a western free-to-play game, and you will literally have people pitch tents outside your office to picket and protest.

This is not an exaggeration – it actually happened at one of the companies I worked at!

Data is A tool, not THE tool

Data, as a tool for design, is valid.

Necessary even. But that's only true when the data is about making the game better or more fun for the players (which, as illustrated by the example above, is a culturally sensitive issue) as opposed to simply more profitable for its publisher.

Data on its own is not a rulebook for games creation. At best, it's a guide.

Without the imaginative vim of a real human imagination driving the design forward, it's just an assembly manual for a cash-harvesting machine.

And if that's what you want to create, maybe you should be working in the gambling industry instead.
You can follow Fraser's industry commentary on his blog, or else grab bite-size rants via Twitter.