Schools measure EdTech return on investment by comparing the full annual cost per active student with two returns measured against a baseline. One is teacher time saved; the other is learning impact, such as mock results in the subjects using the tool. The most common mistake is dividing cost by licences bought rather than students who actually use the platform.
Why is EdTech ROI hard to measure?
Because cost is easy to count and impact is not. A licence price arrives on an invoice. Learning gains and time saved have to be measured deliberately, against a baseline taken before the tool arrives. Without that baseline, schools fall back on usage figures, which show activity, not return. The UK DfE asks AI suppliers to support claims “with robust and transparent evidence” (DfE). A school’s own ROI calculation is how it holds any vendor to that.
How do you calculate true cost per student?
Add every cost, then divide by students who actually use it.
Annual cost per active student = (licence + set-up + training time + integration + support) ÷ students actively using the platform
An illustrative example: a school pays 20,000 a year in licences for 500 students, plus 4,000 in set-up and staff training in year one. If only 300 students use the platform regularly, the real cost is (20,000 + 4,000) ÷ 300 = 80 per active student, not the 40 per student the licence suggests.
Published prices vary widely, so compare like with like. CENTURY lists a school licence from £2,860 a year. exam-mate prices topical papers per account, from USD 25 to 50. Seneca’s core product is free, with paid school tiers, and several others quote on request. See the 2026 platform comparison.
How do you value teacher time saved?
Measure hours, then decide what the hours are for. A before-and-after log is the simplest method: two weeks of marking and preparation time before the tool, and two weeks after, on the same tasks. See how much marking time AI actually saves.
Annual time saved = hours saved per teacher per week × teachers × teaching weeks
An illustrative example: 1 hour a week saved for 10 teachers over 36 weeks is 360 hours a year. The school can put a cost on those hours. The more important question is where they go: planning, feedback conversations or intervention. Time saved only becomes a return if it is used for something that matters. For context, OECD’s TALIS 2024 found lower-secondary teachers spend 4.6 hours a week on marking and correcting on average.
How do you measure learning impact?
Against a comparison, not just before and after. Options, from strongest to weakest:
- Comparison groups. Similar classes with and without the tool, sitting the same mock.
- Year-on-year cohort comparison. The same subject and mock, this cohort against last year’s, using the same boundaries.
- Topic-level change. Re-test results on the topics targeted through the tool.
- Usage correlation. Whether heavier users improved more. This is the weakest measure, because motivated students both use tools more and improve more.
Use the evidence to set expectations. The Education Endowment Foundation’s Toolkit rates feedback as “high impact for very low cost based on extensive evidence” (+6 months). It rates homework as “moderate impact for very low cost based on very limited evidence” (+5 months) (EEF). A tool that improves feedback or homework should be judged against gains of that kind, not against vendor headlines.
How should a school decide whether to renew?
Put the three numbers side by side each year:
| Measure | Question it answers |
|---|---|
| Cost per active student | Are we paying for students who don’t use it? |
| Teacher hours saved, and where they went | Did it give teachers time, and was the time used well? |
| Learning impact against a comparison | Did results improve in the subjects using it? |
Renew if the tool shows impact at an acceptable cost per active student. Renegotiate if usage is low but impact is good among users, and fix adoption before the next term. Stop if neither time nor learning has moved after a fair trial. See why students stop using learning platforms and the EdTech procurement checklist.
How schools do this with AI Buddy
AI Buddy shows active-student usage, practice completion and results by class and topic. Schools can therefore calculate cost per active student and topic-level impact from the platform’s own data, alongside their mock results. We encourage schools to set a baseline before starting AI Buddy and to run the comparison above at the end of the first term. That is the fairest way to judge any platform, including ours.
Frequently asked questions
How do schools measure EdTech return on investment?
By comparing full annual cost per active student with measured teacher time saved and learning impact against a baseline or comparison group.
How do you calculate cost per student for EdTech?
Add licence, set-up, training, integration and support costs, then divide by the number of students actively using the platform, not the number of licences.
Is there an EdTech ROI calculator?
The formulas above work as one: cost per active student, annual hours saved, and impact against a comparison. Tutopiya’s ROI Calculator applies them.
What is a good return on EdTech?
Measurable learning or time gains at an acceptable cost per active student. Evidence-rated strategies such as feedback show what realistic gains look like.
Discover how AI Buddy helps schools strengthen teaching, learning and evidence-informed school improvement. Or start a short consultation with our schools team using the form below — we will get back to you directly.
Sources
- Education Endowment Foundation, Teaching and Learning Toolkit
- Department for Education, Generative AI: product safety standards (updated 19 January 2026)
- OECD, Results from TALIS 2024: The demands of teaching
- Vendor pricing: CENTURY, exam-mate, Seneca