Back to Blog

Case Study: People-Based Measurement Enables Top Retailer to Discover 350% Increase in Return on Ad Spend

  • - Randy Antin
  • 2 min read

Reallocating marketing budgets based on analytics can be complex, but it’s certainly possible. Case study in point: we worked with Ovative/group on a top ten retailer’s people-based measurement strategy, helping them discover more than three times the return on ad spend (ROAS).

The retailer came to us with this measurement problem: According to its online attribution models, display media contributed significantly to total online revenue, but it had no visibility into whether this was reflected in total enterprise sales (i.e., digital and offline) or if the revenue attributed was in fact incremental.

Discovering ROAS Optimization Opportunities

Ovative/group worked with LiveRamp to conduct intelligent, people-based A/B testing for one of the retailer’s major display campaigns. IdentityLinkTM’s identity resolution capabilities enabled the test analysis by connecting tokenized customer data, online and offline transaction files, and display impression logs, and resolving these to the people level. A custom proprietary reporting environment from Ovative/group enabled data analysis across an array of customer-level attributes.

The Power of People-Based Marketing and Measurement

Testing revealed a massive enterprise ROAS optimization opportunity: Display, when combined with direct mail, resulted in 4x greater sales lift than other channel combinations. Based on the test insights, the retailer implemented a best practice to deliver display impressions exclusively to customers who were also direct mail recipients.

If it had delivered this optimal media mix to its audience, the retailer could have:

  • Increased ROAS by 350%
  • Reduced spend by 30%
  • Doubled ROAS by suppressing display impressions served to the lowest performing audience
  • Discovered that display was driving six times as much revenue when considering total enterprise sales as opposed to solely online transactions

Pretty powerful hindsight.