Get a demo Do I subscribe? News sign-up
Print

Attribution a work in progress

News, 03 February 2016

LONDON: Digital and data ought to make more accurate attribution possible across the entire marketing mix, but many marketers struggle to cope with a complex environment and revert to measuring the ROI of individual channels.

Warc's Toolkit 2016 suggests that "connections planning" – where analytics, insights, marketing, media and creative teams work together from the outset – may be a way forward. (Non-subscribers can download a sample of this Toolkit chapter here.)

Major advertisers, such as Mondelez International, are moving toward this approach. As one executive pointed out, a market researcher tasked with delivering ROI needs to understand the reasons a particular media plan is put together.

That also means that researchers need to schedule their marketing-mix analysis so the results arrive at the most impactful moment – just before media-buying choices for the next year are made.

Another option involves a two-tier approach to modelling, with high-level "upstream" marketing mix modelling complemented with individual-level attribution that can help marketers adapt work once it's launched.

In the UK, Vodafone uses econometrics-based high-level modelling, while experimenting with different combinations of media for discrete groups of consumers.

By testing the different media weightings while the campaign is in-market, the mobile operator can get rapid feedback that helps its marketers optimise its campaigns.

Whatever approach one takes, marketers need to understand that "attribution is about more than measuring channel effectiveness", according to Lucy Campion, manager at Deloitte Digital.

"The real value is in its ability to drive a unified customer experience by providing a more transparent view of how customers engage with the brand," she said.

And that requires investment – Campion suggested that between 3% and 5% of marketing spend ought to go on attribution and understanding the drivers of customer behavioural change.

Data sourced from Warc