A Performance Marketing Model Distribution Algorithm Explained by Kimia

Partner Post

- Kimia Premium Online Advertising

At Kimia, it’s our main goal to maximize the revenue of our publishers.

Take for example a pure CPM model, even CPC: with a simple algorithm that distributes the traffic between the advertisers that have the best bids, we would have covered this “goal”.

We all wish it were that “simple”, but we will see that there are more variables to consider.

In a performance model (whether its CPA, CPS, CPL and CPI models) where the revenue is known after the actions of the end user on the product of our advertiser, the task of finding the best ad fit for the traffic is not that simple, because in reality we do not know which advertiser is really the one generating the most for that traffic until after having sent the clicks.

In terms of the behavior of the end user, there are a multitude of variables to consider when redirecting them, such as the subject of the source page, the nature of the product they will be landing on, the type of user (age, sex, tastes, etc.) and even location, day of the week and hour.

In all models, other practical factors come into play, especially in terms of the advertisers and their campaign needs, like frequency capping, retargeting, verticals, categories …

It’s also our responsibility, and interest, that our advertisers receive quality traffic, that really fits with their products and services, to achieve their goals (this will benefit both publisher and advertisers, as source that generate more revenue will obtain a better payout from advertisers) Therefore again we need to look closely at the source of the traffic, its vertical and even the rate of cancellation of subscriptions (unsubscription rate).

Considering all the above factors and variables, the distribution of traffic from our publishers to our advertisers needs an algorithm that maximizes the revenue for both parties, and this is again, not simple, it requires a well thought out strategy backed up by strong technology, algorithms and techniques.

To begin with we will need a very large data hosting capacity, “big data”, predictive algorithms based on machine learning techniques, data analysis, regression … That’s to begin with!. These “terms” are nowadays being thrown around forums and blogs, they aren’t just “trending” topics, or concepts to continue to sell services, they are actual techniques that are starting to work now due to them being based on a need of experience, in Kimia´s case, on our experience from being part of the industry for over seven years.

If you are daunted by these concepts, don’t be, if you wish to learn more do not hesitate to contact us.

Kimia’s aim now is to use this technology combined with our premium human resources to achieve the revenue goals of all the players involved in this market. Technology, top performing offers in all markets, and the best quality traffic, positions Kimia as one of the leading companies in terms of sales/traffic volume and quality.