I came into the Middle East already several years ago, even before I started managing the probably biggest gaming companies activities in the Arab world. While I was working in a company that manages the billing of consumers of online games all around the world, I learned how the games companies applied a very scalable model of performance based marketing.

Understanding the isolated ROI of an additional unit of investment in online advertising should be the basis to identify which form or channel of advertising is actually the most efficient. Several parameters have influence on the efficiency of budget allocation. To list a few the could be

• Creatives: the imagery and composition of the visual “banner”.
• Copy text: a proper and convincing call to action.
• Channel: On which kind of digital platform or channel is the user confronted.
• Landing page: Where a user is taken when he clicks – the place to convert.

All these parameters are to be iterated in a way that for each of the parameters the best combination can be found. All of that requires the ability to track the user journey and consolidate the ROI for each individual combination:

Example: a gaming company might have 5 different creatives combined with 5 different “calls to action”.These 25 creative/copy iterations will then be inserted into 5  channels; let us say a sports page, fashion, news and so on. We need to consider that these 25 ads run now on 5 isolated channels and for each of the channels you will probably have different cost booking your ads in. Then you combine the 125 iterations of your message with various iterations of Landing pages. On these landing pages the actual conversion may take place. This could be a sale, a signup for a newsletter, a download maybe. the succesful conversion results in an “action”. If you iterate the 125 parameter combinations with 5 different landing pages a total of 625 versions will allow you to identify the most efficient cost per action (CPA).

Common mistake: An e-commerce executing this ongoing process of optimisation often thinks it is well of choosing the iteration that has the lowest CPA. I have also seen though that this can be deceiving to look only at the lowest CPA. Eventually you also need to analyse the tracked (future) revenues of the individual users, understanding that you shall not compare the CPA to the next purchase but to all future purchases from that particular (group of) clients. The sweet spot for a conversion based business model is where the “delta” of the CPA and the “Life Time Value” of a certain channel is maximised.

Finding this sweet spot is the challenge to master during the piloting phase, before actually budgets are to be scaled up. Still, do not forget that further parameters may also play a role, and that channel costs may change or that the audience on different channels may not be indefinite.

I would love to hear your thoughts on this. Please comment!