Danone Waters owns Volvic, the number one bottled water brand in the UK*. Working with Ipsos MMA, they turned conventional measurement wisdom on its head by getting granular digital ROI from econometrics (marketing mix modelling or MMM). This leaves them well placed to know where to invest to grow in 2019, as they continue their mission of driving healthier hydration with consumers in the UK.
Goal:
Measure the return on advertising for Volvic water brands
Get detailed insights on the performance of online ads
Approach:
Marketing Mix Modelling (econometrics)
Granular data from Google
Independent modelling from Ipsos MMA
Results
Volvic got detailed results for online ads
YouTube delivered sales well beyond its share of budget
YouTube ROI was broken out by brand, device and targeting
In 2017, the growth in the bottled water drinks category was slowing and the competitive landscape was intensifying. Danone needed to work harder than ever to prove Volvic’s point of difference and continue to recruit people into the brand. With digital doubling its share of advertising budget from 2016, the brand and insights teams needed to understand how all channels were working together to drive sales alongside trade promotions and distribution. For example, how were their YouTube campaigns driving sales?
“Econometrics usually doesn’t provide deep insight on digital performance. By working with Google to get granular data, and modelling and insight from Ipsos MMA, we now have ROI for YouTube by brand, device and targeting type. We can see that it works, but more importantly, we can see where to focus next year to make it work even better.”
– Adrienne O’Brien, Volvic Brand Lead, Danone
Historically, Danone would have turned to econometrics, which looks at weekly data for all factors over 2-3 years and teases out the impact of each factor on sales. But until recently, such models have struggled to provide good insight into digital media performance. Granular, well-labelled data has been hard to obtain; and models have not been sensitive enough to find a statistically significant impact for particular channels and tactics.
Ipsos MMA addressed the first problem by working with Google to get YouTube data by ad format, targeting type, device and region. In 2017, Google launched a Global MMM program with third-party econometrics vendors to provide such data (on YouTube, display, and search) and advice on how its ad formats work. Where possible Ipsos MMA tried to obtain similar data for other media.
In order to improve the sensitivity of the models, Ipsos MMA ‘pooled’ three years of weekly data across nine UK sales regions and five sub-brands. This effectively boosted the data available for the model by 45 times, giving the model more statistical power.
The results gave Danone a very clear picture of the drivers of overall sales and year-on-year changes from 2016-17. But unlike some models, they also gave a clear picture of digital media performance. It was clear that digital channels were ‘punching above their weight’ and delivering more sales than one would expect. For example TV delivered 48% of sales due to media while consuming 63% of the budget (ratio of 0.77); while YouTube delivered 10% of sales due to media from just 4% of the budget (ratio of 2.28).
The more granular data allowed Ipsos MMA to drill down even further. For example, these breakdowns of YouTube ROI raise some great questions for planners.
What worked well about YouTube to exceed TV ROI? What worked so well for Touch of Fruit? Can focus be shifted from tablet to mobile?
These results have helped Danone and their media agency make evidence-based plans for 2019. Optimisation scenarios from the model are guiding strategic (advertising vs. other marketing) and tactical (channel and format) recommendations, whatever the budget and marketing conditions will be.
If you would like to learn more about better measurement of digital in MMM, reach out to your regular Google contact or one of the Google Measurement Partners for Marketing Mix Modelling.
