Districts score 83% on the thing they rate 5 out of 10
Two 2025 reports survey opposite ends of the same purchase order. One says districts are good at evaluating impact. The other shows what they actually decide on. Cost per user wins, because it is the only number that always exists.
Last month I wrote about the gap between what states tell districts to check and what districts actually check. This is the sequel, and it’s stranger, because this time part of the gap sits inside a single survey.
The two reports are Project Unicorn’s fifth annual State of the Sector, which scores hundreds of districts on interoperability and privacy maturity, and Clever’s Classroom of the Future 2025, which surveyed more than 2,500 teachers. One asks the people who sign the contract. The other asks the people who have to open the thing on Monday. They are describing the same purchase from opposite ends, and neither one can see the other end.
The contradiction inside one report
Project Unicorn’s overall district maturity score in 2025 was 47%. The single highest-scoring subscore anywhere in the survey — higher than infrastructure, higher than leadership, higher than anything in procurement — was Educational Impact Evaluation, at 83%. On paper, evaluating whether edtech is working is the thing districts are best at.
Forty pages earlier, in the procurement section, the same respondents were asked to rate their own capabilities on a ten-point scale. Their ability to monitor how edtech tools are actually used: 5 out of 10. Their ability to measure impact on student outcomes: 5 out of 10.
These aren’t the same instrument — one is points earned against a rubric, the other is a raw self-rating — so it isn’t a literal arithmetic contradiction. But it is the same respondents describing the same capability in two places, and the two answers cannot both be right. Districts are confident they evaluate impact. They are not confident they can measure it. That is the shape of a process that produces a verdict without producing evidence.
What fills the vacuum
Clever’s report has the answer, because it asked the other question: when you decide whether a licensed tool is worth keeping, what do you actually look at?
- Budget impact and cost per user — 60%
- Student performance outcomes — 45%
- Usage analytics and login data — 44%
- Direct feedback from users — 35%
- Teacher adoption and training completion — 20%
Cost per user wins, and I don’t think it wins because anyone believes it’s the best measure. It wins because it’s the only number that always exists. You do not need a data team, a roadmap, or an integration to divide the invoice by the headcount. Every other row on that list requires infrastructure the Project Unicorn survey says most districts don’t have: Enterprise Systems Integration scores 23%, only 9% of districts say most of their tools use a data standard, and the average district touches somewhere north of a thousand distinct applications a month.
When the evidence layer is missing, procurement doesn’t stop. It just falls back to the metric with the shortest supply chain.
The row that should bother everyone
Teacher adoption is last, at 20%. Clever’s own data says it should be first.
The report segments teachers into Super Users, Routine Users, and Reluctant Users, and the spread is not subtle. Super Users report 60–70% success reaching disengaged students; Reluctant Users report 15%. 87% of Super Users say edtech makes teaching easier — that drops to 34% among teachers who only touch it when required. Same tools. Same districts. The variable with the largest visible effect on whether a purchase works is the variable that one procurement process in five looks at.
And teachers aren’t in the room to raise it: 21% of educators report having any say in tool selection or privacy safeguards.
You can see the cost of this in one more pair of numbers. 60% of teachers say they want more edtech resources to engage students; 53% say the tools they already have are effective. A district reading only the first number buys more software. A district that could read both would know that a seven-point satisfaction gap, on top of a thousand-app catalog, is not a supply problem.
The obvious objection
Both of these reports are published by interested parties. Clever sells rostering and SSO, and its recommendation is that districts weight teacher adoption more heavily — which happens to be measured by, among other things, login data. Project Unicorn is a coalition whose parent runs an interoperability certification, and its recommendation is that districts demand interoperability. Neither is disinterested and both should be read with that in mind.
I think the finding survives anyway, and specifically because they’re interested in different things. Clever had no reason to go looking for a governance-maturity problem. Project Unicorn had no reason to go looking for a teacher-mindset problem. They measured from opposite ends, on different instruments, with different populations, and landed on the same hole: nobody has a working feedback loop between the tool being bought and the classroom it lands in.
What we’re doing about it
The tempting read, if you sell software to schools, is that this is good news — that a district which can’t measure impact also can’t measure yours. It’s the same mistake as being relieved that the privacy review was shallow.
Here is the version I actually believe. If cost per user wins by default, then the useful thing a vendor can do is make the other numbers as cheap to obtain as cost per user. Not a case study. Not a pilot report we wrote about ourselves. Usage and outcome data the district can pull without standing up a data team first, because State of the Sector is clear that most of them are not going to stand one up: interoperability initiatives are fully funded in 17% of districts, privacy initiatives in 24%.
For Lerad that means two commitments that were in the design before the first pilot. Every metric shows where its data came from and how fresh it is — if a platform doesn’t carry due dates, we say so rather than rendering a reassuring zero. And the teacher is the user, not the byproduct: the product has to survive contact with a Reluctant User in October, because Clever’s data says that’s the person who decides whether any of this worked.
We start piloting in one Brooklyn classroom this September. I’d rather find out there that we’re wrong than find out from a renewal spreadsheet that we were cheap.
Sources: Project Unicorn / InnovateEDU, State of the Sector 2025 (School System Data Survey, hundreds of LEAs and education service agencies); Clever, Classroom of the Future Report 2025 (2,500+ U.S. educators, surveyed Q2 2025 with Whiteboard Advisors).