Insights from CTV leaders at Dentsu, Horizon Media and more
Predictive analytics isn’t just for statisticians. Marketers can decode the cost-cutting properties of the analysis method to create more efficient marketing strategy. Assigning scores to customers according to their propensity towards certain behaviors is nothing new, but linking that behavior to specific marketing strategy decisions is an often forgotten method for conserving a marketing budget.
- Use churn modeling to identify loyal brand followers: Don’t waste retention efforts and resources on loyal brand followers when those resources would be better spent on neglected audience segments.
- Use uplift response modeling to identify consumers in frequent contact with the brand: Don’t waste efforts contacting consumers whom are already frequently engaged.
- Use response modeling to identify consumers that are dead-end clients: Don’t mount elaborate campaigns to win over consumer segments that attach high-negatives to your brand, if their numbers are relatively insignificant.
Common-sense rules apply when acting on insights; data from multiple sources always requires context as well as a logical connection between performance goals and the constant flow of real-time data. Learn more about predictive analytics here
More in Media
Digiday+ Research: Publishers pull back their dependence on digital revenue
After a year in which publishers shifted their revenue dependence away from traditional channels and toward digital channels, 2025 has seen a shift back toward more of a balance between traditional and digital revenue sources.
LinkedIn makes it easier for creators to track performance across platforms
Creator data is becoming more accessible to third-party vendors via a new API — another step in LinkedIn’s creator platform evolution.
Ad Tech Briefing: The ‘plumbers’ posing as the unlikely saviors of the internet
After several false dawns, can Cloudflare’s ‘anti-AI scraping tool’ finally offer publishers a road to commercial redemption?