Taking the Guesswork Out of Personalization: How CSU Global Identified What Worked Before Spending on Media
In short:
CSU Global worked with Delve Deeper to pressure-test messaging before launch using Virtual AI Panels, so they could focus digital advertising investments only where creative would genuinely influence decisions, rather than learning what worked or didn’t, after the fact.
The challenge:
| CSU Global wasn’t short on personalization ideas, but they lacked confidence in which ideas were worth funding. Like many organizations, they were wary of relying on post-launch testing as the primary way to learn, knowing that traditional personalization often requires significant media spend just to determine what works. That approach not only creates waste, it also leaves teams explaining results after the fact rather than making informed decisions upfront.
At the same time, leadership wanted clearer answers to a familiar question: why certain messages resonate, which audiences are truly persuadable, and where personalization would actually change behavior versus simply add complexity. CSU Global needed a way to understand audience motivation and message effectiveness before committing budget, so media dollars were spent learning less and performing more. |
The approach:
| Delve Deeper helped CSU Global move personalization earlier in the decision process, shifting it from a media optimization tactic to a strategic filter for investment decisions. Instead of starting with channels or demographic segments, the work began by understanding how prospective students actually make enrollment decisions.
Using the Jobs to Be Done framework, Delve Deeper conducted qualitative research to capture the real motivations, anxieties, and tradeoffs driving enrollment, grounded in the language prospects themselves use. These insights focused on the underlying “job” students were trying to solve, rather than surface-level attributes such as age, geography, or degree interest. Those human insights were then scaled through Virtual AI Panels, allowing CSU Global to simulate how different audience segments would respond to specific creative concepts before campaigns went live. This made it possible to evaluate which messages had the potential to influence behavior, which messages were unlikely to matter, and where personalization would meaningfully change outcomes. AI was used not as a replacement for research, but as a way to extend it and reduce the cost of learning in-market. |
Summary:
Delve Deeper Results:
The Virtual AI Panel analysis gave CSU Global a clearer picture of how different audience segments would respond to messaging before launch. The work revealed that some prospects were likely to convert regardless of creative, while others were unlikely to convert under any circumstances. Most importantly, it identified a critical persuadable segment whose decision depended on receiving the right message at the right moment.
With this clarity, CSU Global was able to focus media spend and creative effort on audiences where personalization could actually influence enrollment decisions. This reduced unnecessary creative and media waste, improved overall efficiency, and gave the team greater confidence that personalization was driving real impact rather than adding complexity. The result was a more disciplined approach to personalization that prioritized effectiveness over experimentation.