Insurance firm Aviva has been part of a ground-breaking experiment in automation of digital planning by media agency Zenith.

A Zenith taskforce has spent the past six months examining how machine learning technology could be used to efficiently process large amounts of data and to automate the most complex and time-consuming aspects of digital planning.

Using live Aviva campaigns, the taskforce collected advertising cookie data from the tech stack of a DSP and matched it with corresponding first-party sales data. By applying algorithms developed for this exercise, the taskforce was able to precisely attribute sales conversions to specific digital interactions. Then, in what it claims is an industry first, Zenith pushed the algorithm output back into the DSP’s stack to automatically optimise Aviva’s digital planning.

It believes it has found a way of not only solving the long-standing problem of accurate attribution, but closing the automation loop across multiple digital touchpoints. This process saw a six per cent improvement in cost per quote for Aviva on car search, and a 10 per cent improvement in CPQ through automation.

“We’re delighted that as part of our commitment to digital and media transformation at Aviva we are breaking new ground with this pilot automation of our search and display,” said Aviva head of trading James Turner. “The benefits of attribution modelling will be realised in terms of ROI improvement as well as through operational efficiency.”

Zenith is continuing the tests by adding first party drivers-of-demand data data – such as how price affects sales or the success of creative assets – into the algorithm, in order to enhance the effectiveness of the automated planning changes.