Proving the ROI of Top-of-Funnel Higher Ed Campaigns

Last-click attribution is broken for higher ed. To defend awareness spend to a board, universities have to trade it for models that connect early campaigns directly to enrolled students , and the numbers say those campaigns are worth defending.

Ask most enrollment marketers to justify their OTT, streaming audio, or display budget and you’ll get some version of the same uneasy answer: it builds the brand. True, and useless in a budget meeting. When the CFO asks what a dollar of awareness spend actually returned, “brand lift” doesn’t survive the conversation. Search and social, sitting at the bottom of the funnel where the last click lands, take the credit. Everything upstream gets cut first.

The problem isn’t that awareness doesn’t work. It’s that the way most of us still measure it guarantees we’ll never see it work.

Why last-click fails higher ed specifically

Last-click attribution assigns 100% of the credit to the final touch before a conversion. In e-commerce, where the journey from ad to purchase can take minutes, that’s a defensible shortcut. In higher ed, it’s a category error.

A prospective student might encounter your brand on connected TV in October, hear a streaming audio spot in December, and finally search your name and request information in March. Last-click hands the entire enrollment to that March search query. The OTT and audio that built the awareness making the search happen at all? Zero credit. Do that quarter after quarter and the data will “prove” that awareness channels don’t drive enrollments, right up until you cut them and watch your search volume dry up.

You can’t fix a measurement problem with this fundamental with better tagging. You need a different model.

What multi-touch measurement actually shows

Marketing Mix Modeling (MMM) takes the opposite approach. Instead of crediting a single touch, it looks at everything you spent, week by week, across every channel — alongside enrollments, seasonality, and economic conditions — and statistically separates the enrollments your media caused from the ones that would have happened anyway. It answers one question last-click can’t: how many enrollments would we have gotten with no ads at all?

Everything above that baseline is incremental — real students your media brought in. And because MMM doesn’t track individuals, it’s privacy-safe by design, which matters more every year.

We ran exactly this analysis on eight months of always-on programmatic media for a university with both online and campus-based programs: roughly $19.9M in spend across five channels and five program areas, modeled in Google’s open-source Meridian framework. The results reframed the entire debate about upper-funnel spend.

The finding: awareness pays its way — about 5x over

Across the window, media drove 2,933 incremental enrollments — about 46% of all enrollments — at a blended 7.95x return in student value for every dollar of media spend.

It’s worth pausing on what “student value” means, because every return figure in this piece rests on it. MMM measures results in whatever unit you give it, and here that unit is enrollments, not dollars. Left in its raw form, the model’s output reads as enrollments per dollar of spend, a number too small to mean anything in a budget meeting. To translate it into terms a CFO recognizes, we assign each incremental enrollment an average student value of $54,000, which represents the tuition revenue a typical enrolled student brings to the institution over the course of their program.

So a 7.95x return means every $1 of media generated roughly $7.95 in tuition revenue from students who would not have enrolled without it. Put another way, media brought in each incremental student at about $6,800 in cost per enrollment, against a student worth $54,000. Two things keep that number honest: it counts only incremental enrollments, net of the students who would have come anyway, and it’s a revenue figure, not margin. Every channel return in the table below is calculated the same way.

The headline everyone expected: search led on efficiency, returning about $9.63 per dollar, with paid social behind it at $5.54. Those two carried 90% of the budget and converted intent that already existed. No surprise there.

The headline nobody expected: the awareness channels held their own.

OTT returned 5.36x — within a rounding error of social’s 5.54x. Read that again. The channel most likely to be first on the chopping block performed on par with the channel we treat as a workhorse. Audio and display weren’t far behind. These aren’t the returns of overhead. They’re the returns of demand generation that lower-funnel channels later get paid to convert.

The multiplier last-click can never see

Here’s the part that makes the strongest case of all — and the part a single-touch model is structurally incapable of detecting.

When two channels run in the same week, each makes the other more effective. We measured this synergy and ranked it, and one channel stood out as connective tissue for the entire plan: audio. Small in raw volume — around 80 enrollments — but when it ran alongside its partners, it lifted them dramatically:

• Audio made display 79% more effective
• Audio made search 75% more effective
• Audio made social 71% more effective

An audio budget that looks marginal on its own line item is quietly making your highest-volume channels work harder. Cut it to save money and you don’t just lose 80 enrollments — you lose the lift it was handing to everything else. Last-click will never show you this, because it credits one touch and blinds you to every interaction between channels. Multi-touch modeling is the only way to see the pairs worth protecting.

What this means for the board conversation

The takeaway isn’t “spend more on awareness.” It’s stop judging awareness channels by the wrong yardstick.

1. Don’t hold OTT, audio, and display to search’s cost per enrollment. They play a different funnel role. Measured by that role, they earn their place at roughly 5x return.
2. Keep upper-funnel channels always-on for reach, while concentrating incremental performance dollars where intent converts.
3. Protect high-synergy pairs. Cut one channel and the shared lift collapses with it.
4. Adopt a measurement model that can actually see all of this. The reason awareness spend is perpetually vulnerable isn’t performance — it’s attribution. Fix the model and the budget defends itself.

For years, top-of-funnel media in higher ed has been funded on faith and cut on math. Multi-touch modeling finally puts the math on the other side of the table. When you can show a board that awareness drove nearly half your enrollments and multiplied the return on everything downstream, “it builds the brand” stops being a defense — and starts being a number.

This analysis was produced by EducationDynamics using Google’s Meridian MMM framework. Figures reflect a single eight-month, $19.9M programmatic engagement and are specific to that campaign; results vary by institution, spend level, and market.