How to define the most effective advertising mix?

Leveraging Amazon Marketing Cloud data for Philips Hue

The challenge

Our client Philips Hue is running digital campaigns on Amazon via different levers:

  • Sponsored ads, for visibility on the market place
  • Display and video campaigns via Amazon DSP, in order to target people who have interacted with the brand or to set up audience extensions by activating Amazon’s retail data.

These parallel activations were previously analysed separately. But to continue to grow the brand on the platform and develop sales, we needed to understand :

  • How do our different activations connect?
  • What are the synergies caused by these different activations?
  • What is the contribution of each of the activated levers?

The solution

Amazon’s new Clean Room, Amazon Marketing Cloud, removes the barriers to measurement by bringing multiple sources of activation data together for the first time within a single study scope. To do this, we have developed customised studies using our strong expertise in data science.

Our teams of data scientists and traders have developed a suite of customised studies to better understand the results and impacts of different campaigns in order to define activatable insights for our client Philips Hue:

  • A study on reach and overlap of exposures and conversions between our on-site visibility activations and our display and video activations
  • A multi-levers study to determine the impact of exposure to one or more levers in the conversion rate
  • A study to understand the impact of the position of each lever in the conversion journey to determine the relevance of the chosen levers and the best combination to lead to conversion

We therefore analysed the overlap of our different activation levers across the funnel for a month.


The first study shows that our display and video activations considerably extend the overall reach of campaigns and that they do not target the audience reached by on-site visibility campaigns. It also shows us that display activation is involved in almost all conversions.

The second study shows that 90% of the targeted audience is single-levers, whereas more than 50% of purchases come from people who have been reached by at least 2 levers. This study also shows that the distribution of purchases by lever is roughly the same for old and new buyers.

This latest study shows that in the top 3 conversion paths, every conversion journey starts with display exposure.
The best path to conversion is therefore made up of display exposure first, followed by video exposure and then on-site exposure to sponsored products, probably as a result of a search on the store.

The combination of these three studies allows us to clearly establish the impact of the different levers on conversion.

This study enabled us to determine the best conversion paths in order to refine the budgetary management for the coming year and for the end-of-year holidays.

Key features

10 %

of people exposed to Sponsored Ads were also exposed to Display


purchase rate with exposure to complementary advertising formats

+ de 50 %

of purchases are generated by people who have been reached by at least 2 levers