Challenge: Sitting on a goldmine, holding a blunt shovel

Pod sales, not coffee machines, are the profit centre for the Nespresso business. So, one important job for advertising is to encourage customers who have bought machines to buy coffee pods through their Nespresso account, rather than buying supermarket-bought compatible products. ​

As you can imagine, Nespresso has a large CRM database of customers. Every person who has bought a machine can (in theory) be messaged about how Nespresso pods make the best coffee. ​

Historically, to do this, we have matched Nespresso’s CRM database against the identity graphs of media owners’, to create pools of existing customers we can advertise to in display, VoD and social.

But our confidence in the accuracy of this method had dwindled.

People are messy, with multiple devices, emails, and IP addresses. They share them with others, and they don’t browse the web consistently logged in. This means the data media owners use to create IDs is unreliable, and so targeting is ineffective.  This issue is getting worse as our web usage fragments across devices, places and accounts.​

Worse, recent changes to how browsers and operating systems deal with tracking have made it even harder to unpick the tangled knot of identity online.​

This was a big problem. The rise in WFH had created a boom in new customers, but the profit opportunity would go to waste if they couldn’t be encouraged to buy pods from Nespresso.  We needed a new approach.

Solution: A whole new way of making IDs

Enter Epsilon, our Publicis-owned platform with a whole different approach to building identity online.

Epsilon know that stitching together IP addresses, emails and devices likely lead to having several IDs that are in reality the same person. This means low levels of genuine reach and high frequency: essentially chasing a small group of customers around the web, pissing them off and generating little new revenue.

So Epsilon start from the one data point that most people in this country have just one of: name and address.  ​

Using the highly reliable information people give when buying from online retailers, Epsilon have built a 29m strong identity graph of consumers in the UK, that is demonstrably more accurate than anything else of equivalent scale.

Result: Forget shovels, we’d built a jackhammer

When we matched Nespresso’s database with Epsilon’s, we immediately increased our match rate by nearly 2x. This meant we could reach a far greater number of Nespresso members every week. ​

Now, with a better grasp on identity, we could reduce frequency to a more sensible level, whilst also reaching more members.  ​

After 90 days, we had generated significant incremental revenue, and increased incremental revenue ROI.

Customers being messaged via Epsilon targeting also had a 5% higher average order value than our site-wide average.​

In addition, by delivering fewer impressions to make a sale, we could hit revenue targets whilst deploying less advertising.

This case shows how CRM can translate into media in a way that genuinely finds customers in the wild and leverages the opportunity for profit.