How does it work?
Our Adtriba optimization algorithm uses a smart trial-and-error methodology to vary the media budget allocations per day in order to achieve the optimal sum of the output KPI chosen by the optimization strategy. So, if you want to maximize conversions, it will allocate media spend in such a way, that the sum of conversions over the given timeframe is maximized, not spending more than the specified maximum budget.
When do I know that the recommendation is reliable?
When you look into the future, the recommendation becomes less reliable, since know can reasonably claim to know what exactly will happen in the future.
Can I do something to make the recommendation more reliable?
There are two main areas in which you can contribute to make our algorithms work better.
Firstly, you can help to improve the data that goes into SPHERE.
Those steps include:
Upload TV data (or other offline data)
Improve the campaign to source mapping in the funnel dimension Adtriba Sphere Source
Let us know of custom events and seasonalities that impact your business
Secondly, you can validate the results, monitor them and give us feedback!
Why can't I go further in the future than 4 weeks?
Predictions are hard - especially when they are about the future. The farther we go into the future, the less reliable our estimates can be. Since we don't want you to allocate your budget based on increasingly unreliable estimates, we decided to restrict the look-ahead window in such a way that the predictive accuracy of our algorithms is still deemed tolerable. Please be again aware that *no one* can know what exactly will happen in the future.
Are seasonalities covered in the model and does the model understand seasonality peaks?
Yes, by default we include weekly (day of week), yearly (month of year) seasonality, along with bank holidays, Black Friday and Cyber Monday. Please talk to the customer success team if you have "holidays" that are custom to your business and we will try to integrate those.