Eloquii is a high-growth internet retailer that specializes in fast fashion for women who fall outside of the 0 to 12 size range. The company was launched as an independent entity in 2014 with the vision of designing clothing and accessories exclusively to fit a fuller figure. Eloquii provides the plus-size market with fun, edgy apparel based on the latest runway trends.

As a high-growth, venture-backed internet startup, Eloquii is focused on attracting new customers to the brand. But Marketing Vice President Kelly Goldston realized that the key to maximizing value is not just acquiring new customers, but also managing the cost of acquiring new customers in relation to the lifetime value of the customers they acquire. In order to optimize CAC:CLV (the ratio of customer acquisition costs to the lifetime value of the customers acquired), Eloquii realized they needed to look beyond the traditional metrics like ROAS (Return on Ad Spend) to understand the longer term impact of their performance marketing investments, and they had to work pro-actively to grow the lifetime value of their customer base.

The Solution

The Eloquii team envisioned an approach that seemed revolutionary at the time. They wanted to utilize predictive analytics and advanced customer segmentation to optimize the lifetime value of their customers. But doing this in-house would require significant data science resources and lots of time to build and test algorithms.

The team turned to Custora’s Customer Analytics platform for the solution. Custora builds a profile of each user’s attributes and behaviors and then looks across millions of data points for hidden relationships within different purchasing patterns —using machine learning to anticipate the future behavior of each customer. It was exactly the type of advanced customer analytics approach the team had been looking for to predict the lifetime value of each customer and optimize the long term return of their performance marketing programs.

The Eloquii team focused on three areas. One–understanding which performance marketing investments attracted customers with the highest customer lifetime value. Two—using predictive models to identify high-value customers very early in their relationship with the brand. And three—predicting and preventing churn.

In the first case, Eloquii used the predicted lifetime value of each customer to determine which advertising channels were delivering the highest value customers. They evaluated the investment in each channel based not on just on immediate conversion, but on the predicted purchases that those customers would make in their lifetime with the brand.

In the second case, Eloquii used Custora to predict which customers would be high value “VIPs” long before they passed the VIP purchase threshold. The Custora algorithms identified these “early VIPs” and Eloquii enticed them to increase their purchase frequency by acknowledging how important they are and by providing early access to designer collections and sales.

In the third case, Eloquii used Custora to predict when individual customers were beginning to drift from the brand. Actual behavior was compared to predicted shopping patterns. Rather than waiting to win back customers once they were already deemed lost, Eloquii was determined to prevent churn by intervening early, when customers showed the first signs of “cooling down” or “at-risk” behavior. Customers were sent “we miss you” emails, and high-value customers were sent surveys to try and understand why great customers were slipping away.


Eloquii saw an immediate impact after implementing the Custora solution. To determine actual lift of the customer retention program, Eloquii ran a rules-based win-back program (emails triggered after a set number of days post-purchase) and tested it head to head with Custora’s customer-specific churn predictions. Custora’s predictive churn algorithm delivered 27% more revenue than the rule-based method. Compared to a control group, the early VIP program generated a 12 to 15% revenue increase. And Eloquii’s overall performance marketing budget was tuned and optimized for lifetime value maximization.

“Rules-based segmentation helps us deliver the right message to the right person at the right time. Custora’s predictive segmentation helps us deliver the right message to the right person at the right-er time”, said Kelly Goldston. “Our predictive churn program with Custora outperforms a traditional rules-based approach by 27%”.