Skip to main content
Media Optimizer Overview

Overview of the Sellforte Media Optimizer

O
Written by Oskari Raunio
Updated over a month ago

Media Optimizer helps plan marketing budgets based on the Marketing Mix Modelling results​

Media Optimizer is a planning tool that enables users to ​

  • Find the optimal media mix for the given budget​

  • Find out the required budget to meet a specific sales target.​

​Media Optimizer builds on the Marketing Mix Modelling results:​

  • It uses the historical investments to find out reference values for the budgets and sales targets.​

  • It derives the marketing response curves from the Marketing Mix Model.

Media Optimizer uses historical values as a reference by default​

  • Media Optimizer uses the diminishing returns of the marketing response curves to find the optimal budget allocation. The optimization is limited by budget and number of week constraints.​

  • By default, the budget and the number of active weeks per marketing activity are set to historical reference values.​

  • User can give the Optimizer more freedom to find the optimal budget by adjusting the marketing activity specific budget constraints.​

  • User can also specify the active weeks per marketing activity either by setting the number of weeks or by selecting specific weeks from a calendar.​

  • The recommended approach is to make gradual adjustments instead of drastic changes.​

Baseline forecasts are included in the optimization based on a set of pre-computed scenarios​

  • Sellforte’s analytics pipelines compute a customizable set of baseline scenarios that users can select from when running the Optimizer.​

  • Users can choose different baseline scenarios for different Optimizer scenarios to find out how budget allocations would change if e.g. business growth is slower or faster than expected.​

  • Initially the baseline scenarios are limited to varying year-on-year growth scenarios but additional scneario types are under development.​

  • For historical date ranges, the historical baseline values are used.​

Media Optimizer infers the seasonality of marketing uplifts from the baseline seasonality​

  • Optimizer uses the seasonality and trends in the baseline to adjust predictions: higher uplifts and better ROIs are predicted for times when baseline is larger.​

  • The time variation of the predictions is utilized to find the optimal allocation over weeks. ​

  • Time variation in the predictions is also used to find the optimal weeks if the number of active weeks is specified.​

Media Optimizer needs sometimes guidance from the user to understand real world limitations​

  • MO does not understand media planning and media-specific limitations and human intervention/adjustments might be required​

  • MO might sometimes suggest too high/too low budgets (eg. too high investment relative to target group size) or ​

  • MO might recommend investing in channels which are unfit to planned campaigns (eg. recommendation to invest a significant amount to search for brand-campaign)​

  • Media Optimizer is biased towards past investments and our own MMM results  optimizer will not recommend budget for media that has not been used or tested befor​

  • Media Optimizer forecasts average performance. The predictions can differ even drastically from individual historical MMM result data points.

Did this answer your question?