Discovering where and how conversions are happening is essential for measuring the performance of your media channels. Predictive modelling is a data analytics solution that allows a business to assess their marketing budget over various channels. It explains how investments in multiple components contribute to revenue.
With the appropriate media mix model, a business can use its past marketing performance to increase future ROI by optimising the channel’s media budget allocation.
We use a dataset, including the following elements in our tool quaartz:
In the fast-evolving digital media spectrum, the conventional techniques applied to investigate marketing’s results are no longer the best. We step on the cusp of a media metamorphosis. Media will become more targeted and personalised. Marketers need to adjust to this fact: to grow their business, they need tools to find the perfect media mix.
This challenge raises the following marketing question:
Can the marketing media mix be fine-tuned to improve ROI?
Based on the dataset described earlier, we asked Quaartz to predict the sales figures depending on the media marketing mix. Quaartz predictive models result in a 99% R2 and a 3% MAPE. It means that the predictive models are very accurate.
Quaartz has selected the following four main criteria out of the eleven available in the dataset as being relevant and essential for the prediction, with various relative weights:
For this specific campaign, Quaartz shows that investing in TV advertising is worth it, up to a plateau. It also shows that, in this case, the date chosen for the advertisement does not have a significant impact on the revenue generated.
In three days, the advertising team achieved to: