The trade-off between the exploration of new possibilities and the exploitation of old certainties is one of the most basic dilemmas just as in any business so in mobile advertising.
We all face the question of whether to make a certain decision based on the knowledge we currently possess (“exploitation”), or rather wait and invest more time and effort in gathering further information for a broader perspective that could lead to a better decision in the future (“exploration”).
At a restaurant, do you order the dish you know and love, or go for something new? There is even a mathematical problem regarding this dilemma called Feynman's restaurant problem. The same is in mobile advertising. Mobile marketers often are in conflict between trying different things to explore the optimal way to achieve their goals, or exploiting their current situation and knowledge to make the best out of it.
The Aarki Way to the Dilemma
The key is to find the right balance between these two strategies since neither of them is uniformly superior to the other. Exploitation strategy maybe a right decision for achieving short-term goals while for a long-term success exploration strategy could be undertaken.
Working in an industry of increasing personalization, we often see our clients faced with these two contradictory needs of maximizing their revenues on one hand, and learning what makes their customers tick on the other. Using transfer learning from non-attributed events can lead to reducing the amount of exploration needed by giving a better starting point.
To learn more how Aarki deals with exploration-exploitation trade-offs through bayesian bandits, look out for the second part of this article.