Telefonica brand giffgaff have worked with Metageni to develop data-driven analytics which combines and offline and online marketing, and cross-device attribution.
The complete solution combined both offline media and online user journey data in a system of machine learning algorithms designed to predict sales.
“This is an awesome piece of work. The feedback internally has been fantastic….everyone is clear that this is it with regards to what we can do with the attribution data which is great, and the fact that we can assign CPA level metrics to each channel was something that went down very well internally.
Thank for all your help on this project”
Matt Roberts, Strategic Analytics Manager, giffgaff
These bespoke predictive models are used as the basis of an attribution modeling system. The algorithms distinguish the ROI of separate channels to help with marketing optimization, with spend from low-performing channels switched in to boost higher performing channels.
The outcome is increased sales growth and marketing efficiency.
Other features of the system included a cross-device inference model for improving device-based optimisation, and additional modelling for upper-funnel and offline to brand navigation channels such as brand and organic search, direct and re-targeting.
The complete system implemented by giffgaff included live backward integration with a big data cloud and custom live visualisations and query and tools.
The whole system is designed to create a data story for accessible insights to improve ROI through improved transparent attribution.
Predictions and attribution can be dynamically updated whenever new data is available in the data warehouse, and the machine learning models are re-trained to boost accuracy.
Find out more about giffgaff mobile here.
Can we help unlock the value of your analytics and marketing data?
Ask us about your marketing optimisation at: [email protected]
Can we help unlock the value of your analytics and marketing data?Ask us about your marketing optimisation at: [email protected]
