Client Use Cases: 1-800 CONTACTS Flags Customers with Look-Alike Data
Updated: Sep 30, 2019
The following is the fourth post in a five-part series on how 1-800 CONTACTS uses Compellon20|20 to drive change within their organization. Thank you to Neil Wieloch, PhD, Director, Marketing Strategy and Insights, for writing these posts and for sharing 1-800 CONTACTS’ successes with us. In case you missed them, the first post in this series was about identifying the drivers of churn, the second one was about creating predictive variables, and the third one was about unifying Marketing silos.
The Key to Activating Insight: Flagging All of Your Customers
Conducting traditional marketing research can be a great way to paint a “big picture” of what is happening in the marketplace in order to develop high-level strategy. But the challenge is then in the activation: how do you decide where to go? Specifically, how can you identify clear and measurable objectives, track strong indicators of progress, and link marketing activities to actual individual behaviors (in order to measure actual value)? As we all know, unless an initiative can be clearly measured, it doesn’t exist.
The Insights Team at 1-800 CONTACTS recently found themselves in this position: using questions from their ongoing post-transactional satisfaction survey, they identified a few key consumer segments that acted in different and meaningful ways. But since the only customers that were segmented were the ones that completed the survey, only about 1.5% of the entire active customer database could be segmented and, therefore, targeted, tested, and tracked.
Were they to implement these insights from a traditional marketing approach, the Insights Team would have had to make assumptions about all their customers and run a number of A/B tests to see which tactics would best work in aggregate – rather than test and learn among each of the segments individually.
Instead, the team decided to use the Compellon engine to run look-alike data and flag the entire customer database according to each customer’s likelihood to be a member of one of the segments. Specifically, the Insights Team appended all their active customers with third-party demographic and psychographic data, then ran it (along with their transactional behavior data) through Compellon20|20, and generated predictive models for each of the segments. Once the models were built, the team then set up an API to update and flag all new customers on a weekly basis.
Now 1-800 CONTACTS can test, treat, and track each segment differently – and link their actions directly to financial results based on individual customer purchase behavior. Additionally, the same models that Compellon20|20 generated to determine segmentation identified the drivers that 1-800 CONTACTS uses to target and message different prospects for acquisition.