Global Affiliate Marketing Network Solves Sales Recovery Conundrum

The Challenge

The affiliate network noticed that as time goes on, people use more and more devices on a daily basis to shop: phones, tablets, laptops, smart TVs and more. The issue they faced was seeing all of the sales across all of these devices belonging to each user. The client had its own cross device tracking solution based on deterministic data, but the deterministic approach had limitations as it relied on hashed email addresses which only some users wanted to supply. Unidentifiable transactions were costing the client money and they required a solution to fully identify them.

Our Approach

Create a graph that shows all of the different devices that belong to a given user. Follow the corresponding clicks on the different devices for each user. Attribute the sales that otherwise could not be connected with a user, thus recovering sales that would have been lost.

How We Did It

A phased engagement combining machine learning, close product collaboration and iterative delivery cadence.

1. 1

   The client supplied ROQAD with sales transaction data, which was used to train machine learning algorithms.

2. 2

   ROQAD worked closely with the client's product and account management teams to adjust the solution to fit their needs.

3. 3

   A more frequent product delivery was agreed: starting from once per month, then every 2 weeks, and then weekly.