Proving marketing return in higher education
Jonathan Sabarre · , updated
Every university marketing director has been asked the same question in a budget meeting: what did we get for that money? Most of us have answered with activity. Fewer of us have been able to answer with return, and fewer still have been clear about how much of that return is an estimate.
Why return is so hard to show
The difficulty is not a lack of data. Universities have web analytics, CRM records, application systems, event registrations and social reach. The difficulty is that the data sits in different places, owned by different teams, describing different stages of a long journey. A prospective postgraduate might read a news story, attend a webinar months later and then apply directly. No single system sees all of that.
There is also a cultural problem. Teams are often measured on outputs (campaigns launched, coverage secured) because outputs are easy to count. Outcomes are harder to connect to spend, so marketing looks like a cost rather than an investment.
Three views, three levels of evidence
The most useful thing a team can do is keep three questions apart. They are often presented as if they were the same thing. They are not.
| View | Question it answers | What it rests on |
|---|---|---|
| Observed performance | What did we spend, what happened and what did each stage cost? | Recorded spend and outcomes for a stated period and cohort, with overlaps and deduplication rules explained. |
| Estimated financial return | Under our stated assumptions, what value might marketing account for? | A chosen value basis, a declared influence or attribution share, an allocation rule and a sensitivity range. |
| Incremental impact | What would not have happened without the marketing? | An appropriate experiment or other defensible comparison, such as a holdout region. Only claim it when that evidence exists. |
Cost per enquiry is efficiency, not financial return. An estimated return depends on its assumptions. Neither shows that marketing caused the outcome.
Start with the "I" in ROI
Before you can talk about return, you need an honest figure for investment. That means everything it takes to deliver the work, not only media spend: paid advertising, agency fees, content production, events, technology licences and a fair share of staff time. Leaving staff costs out flatters every ratio that follows. If the people are part of how results are achieved, they are part of the investment.
Cost per stage, for every audience
Once you have a total, the most useful first view is cost per stage. How much does it cost to bring one visitor to the website, gain one newsletter subscriber, generate one enquiry, one application and one enrolment? Each figure answers a different question, and together they show where the funnel is efficient and where it leaks. Define each stage before you count it: a unique visitor is not an impression, and an enquiry is not an application.
Universities also speak to more than one audience. Students are the obvious one, but academic and employer audiences matter for research partnerships, staff recruitment and graduate outcomes. A good model tracks each audience through its own stages rather than forcing everything into a student funnel.
Blended and allocated views
There are two honest ways to divide cost. A blended view divides the total investment by the total at each stage, regardless of audience. It suits top-of-funnel measures such as visitors and subscribers, where audiences overlap. An allocated view assigns a share of investment to each audience, based on where effort went, and divides only that share by that audience's results. It is fairer for later stages, where student enrolments should not carry the cost of employer engagement work. Show both, and say which one you are using.
Be honest about attribution
Marketing rarely causes an enrolment on its own. Reputation, teaching quality, location, fees, friends and family all play their part. Claiming the full value of every enrolment for marketing is not credible, and finance colleagues know it.
State your assumption openly. A declared influence share says "we assume marketing accounts for this proportion of the outcome", and it is labelled as an assumption, not a finding. Multi-touch attribution assigns credit across channels and moments, which is still a model. Incrementality testing, for example holding back a campaign in one region and comparing results, is the evidence that comes closest to showing cause, when you can run it.
We take AMEC's Barcelona Principles 4.0 as a reference point here: be transparent about method, measure outcomes rather than outputs, and do not use advertising value equivalency (AVE). AVE describes what coverage might have cost as advertising. It says nothing about what the coverage achieved. Not every communications activity should be turned into a tuition figure; audience understanding and relationship quality matter in their own right.
A worked example
Illustrative figures These figures are fictional and are not drawn from any real institution. They show the method.
Suppose a university invests £750,000 in marketing and communications across the year. Using a blended view:
- 750,000 website visitors gives a cost of £1.00 per visitor.
- 75,000 newsletter subscribers gives a cost of £10.00 per subscriber.
Now take an allocated view for the student audience, which received half of the investment, £375,000:
- 12,000 enquiries gives a cost of £31.25 per enquiry.
- 6,000 applications gives a cost of £62.50 per application.
- 900 enrolments gives a cost of £416.67 per enrolment.
That is an enquiry to enrolment ratio of 7.5% for the period. Everything so far is observed performance, if the counts are real.
To estimate value, assume an annual fee of £15,000, a three-year programme and a 90% completion rate: £15,000 × 3 × 0.90 gives a lifetime value of £40,500 per student. Across 900 enrolments, the total is £36,450,000. This is gross tuition, not contribution after delivery costs, and a completion factor is not a substitute for a cohort cash-flow model.
With a declared influence share of 10%, the estimated value attributed to marketing is £3,645,000. Against the £375,000 student allocation, that gives an estimated return of 872%, calculated as (value minus investment) divided by investment. The same figures give 9.72×, which is attributed value divided by marketing spend. Both are estimates under the stated assumptions. Neither shows that marketing caused this value. Finance teams define return on marketing investment in different ways, so say which convention you are using.
Why the assumption matters
Keep the same £36.45m lifetime value and the same £375,000 investment, and change only the influence share:
| Influence share (assumption) | Estimated value | Estimated return | Attributed value divided by spend |
|---|---|---|---|
| 5% | £1,822,500 | 386% | 4.86× |
| 10% | £3,645,000 | 872% | 9.72× |
| 15% | £5,467,500 | 1,358% | 14.58× |
The estimated return moves from 386% to 1,358% without a single enrolment changing. That is the point of showing the working: the conversation moves from "trust us" to "here is our working, and here is the assumption you may want to challenge".
What changes when you can show your working
In my experience, the most valuable outcome is not the headline ratio but the change in the conversation. A team that can show cost per stage by audience, separate what it observed from what it estimated, and explain its assumptions earns a different kind of seat at the table. Budget discussions become about where the next pound is most likely to help, and what evidence would settle the question.
Sources and further reading
- AMEC, Barcelona Principles 4.0: amecorg.com