We Don't Need More School Data. We Need to Get Better at Using It.

Schools are not short of data.

We have percentages.

Grades.

Scaled scores.

Predicted grades.

Targets.

Assessment points.

Mock results.

Question scores.

Progress measures.

Spreadsheets with so many conditional formatting colours that sometimes I think we need another spreadsheet just to explain the first one.

But there is a question I have asked repeatedly throughout my time working with schools:

What did we actually do differently because we collected it?

That is the bit that matters.

And it feels particularly timely now.

In September 2026, the Department for Education announced its new Key Stage 3 Alliance as part of the RISE school improvement offer. Among its stated priorities are improving literacy and numeracy and, significantly, strengthening assessment and data coherence.

I think the word coherence is important.

Because the answer isn't necessarily another assessment.

It isn't another data drop.

And it certainly isn't another beautifully colour-coded spreadsheet that somebody spends hours producing before it is filed away.

The real challenge is connecting assessment to what happens next.

What don't pupils know? Why don't they know it? Who needs something different? And what are we going to do about it?

That is where assessment becomes useful.

The problem isn't that schools don't collect enough data

Having worked across schools and with teachers at different stages of their careers, I have rarely encountered a school suffering from a complete absence of assessment information.

Usually, the opposite is true.

Teachers are surrounded by it.

The problem comes when we confuse having data with having insight.

Imagine I tell a Year 8 English teacher:

Class average: 62%.

What should she teach tomorrow?

We don't know.

Imagine I tell her:

The class performed securely on retrieval and vocabulary questions, but struggled when they had to select evidence and explain its effect. Eleven pupils lost marks on essentially the same analytical skill.

Now we have something we can work with.

That distinction sounds obvious.

Yet schools can spend an extraordinary amount of time producing the first type of information.

A score tells you what happened. It doesn't always tell you why.

This is one of the limitations of headline assessment data.

Suppose two pupils both score 58%.

On a tracker, they look identical.

But look underneath the score and we might find something completely different.

Pupil A knows the curriculum content extremely well but repeatedly misinterprets multi-step questions.

Pupil B understands what questions are asking but has significant gaps in the underlying knowledge.

Same score.

Different problem.

Different response.

If I only know the percentage, I might put both pupils into the same intervention group.

If I understand the questions behind that percentage, I might make two entirely different teaching decisions.

That is the difference between recording attainment and diagnosing learning.

Why Key Stage 3 deserves particular attention

Key Stage 3 can sometimes become the awkward middle child of school assessment.

Primary has established statutory assessment points.

Key Stage 4 has GCSE specifications, mocks, predicted grades and the gravitational pull of examinations.

Key Stage 3 sits between them.

And that creates a risk.

Assessment can become either too loose to tell us very much or too GCSE-shaped, too early.

Neither is particularly helpful.

The new RISE Key Stage 3 Alliance is therefore interesting because the Department for Education has specifically identified assessment and data coherence as an area for development.

For me, coherent KS3 assessment means being able to answer:

What was taught?

What was learned?

What wasn't secure?

Is the problem isolated or recurring?

Which pupils are affected?

What will we do next?

And crucially:

Did that action work?

That last question is often missing.

Assessment shouldn't end when the spreadsheet is completed

I have sat in plenty of data meetings during my career.

There is a familiar pattern.

We identify that a pupil is behind.

We discuss why.

We create an intervention.

We write the intervention down.

Then everybody moves on to the next assessment cycle.

But the important question isn't:

Did we create an intervention?

It is:

Did the pupil learn what they couldn't do before?

Those are not the same thing.

An intervention can take place perfectly and achieve very little.

A topic can be re-taught without the misconception actually disappearing.

A pupil can attend six weeks of additional support without closing the specific gap that prompted it.

Assessment therefore needs to become a cycle:

Assess → identify → respond → reassess.

Without the final step, we have measured activity rather than impact.

Start with the question, not the pupil label

This is one of the reasons I am such an advocate of looking beneath overall marks.

Let's say 34% of a class answered Question 7 correctly.

My first question isn't:

Which pupils failed the test?

It is:

What was Question 7 actually testing?

Perhaps it tested fractions.

Perhaps it required inference.

Perhaps pupils had to identify a scientific variable.

Perhaps it required them to apply knowledge in an unfamiliar context.

Perhaps the underlying knowledge was secure, but the wording caused difficulty.

Perhaps the whole class struggled.

Perhaps disadvantaged pupils performed differently from the rest of the cohort.

Perhaps one teaching group struggled while another didn't.

Each of those tells us something different.

This is where question-level analysis becomes much more useful than simply looking at total marks.

Not because schools need more analysis.

Because they need more precise analysis.

But there is a warning here: don't re-teach everything

One of the easiest mistakes after an assessment is to produce a list of everything pupils got wrong.

Then we look at 14 gaps and think:

Right. We need to re-teach all of those.

We probably don't.

Some incorrect answers are noise.

A pupil misread something.

A question was unusually difficult.

One obscure piece of knowledge was forgotten.

A careless calculation happened.

The skill is identifying the patterns.

If a significant proportion of pupils have misunderstood the same concept, that deserves attention.

If one pupil has a very different profile from everybody else, that deserves attention too — but probably not a whole-class lesson.

If a weakness appears repeatedly across several assessments, that becomes even more interesting.

Good assessment analysis helps us distinguish between:

“They got this question wrong.”

and:

“There is a learning gap here that requires action.”

Those are not automatically the same thing.

Five questions I would ask after every meaningful assessment

If I were working with a department on assessment tomorrow, I would start here.

1. What did pupils do well?

We are often very good at hunting deficits.

But knowing what is secure matters too.

It prevents unnecessary re-teaching and tells us where teaching has been effective.

2. Which errors occurred across the cohort?

This is where patterns become useful.

If half the year group struggled with the same curriculum objective, that may indicate a curriculum, sequencing or teaching issue rather than an individual pupil issue.

3. Which pupils don't fit the overall pattern?

This is incredibly important.

The class may perform well overall while four pupils have significant gaps.

Averages can hide children.

4. What are the two most important things to do next?

I deliberately say two.

Not twelve.

Assessment needs to lead to manageable action.

What will make the greatest difference to subsequent learning?

5. When will we check whether it worked?

This should be decided at the same time as the response.

Not six months later.

Not at the next formal data drop if that is too far away.

How will pupils demonstrate that the gap has closed?

That might be an exit question.

A retrieval quiz next week.

A similar problem in a later lesson.

A short assessment.

Another extended response.

The method matters less than completing the loop.

Data coherence doesn't mean everyone needs the same spreadsheet

This is another distinction I think school leaders need to make.

Consistency is useful.

Uniformity isn't always.

A history department and a maths department do not necessarily need assessment to look identical.

Their curricula are different.

The knowledge is different.

The disciplinary thinking is different.

The assessments themselves are different.

What should be consistent is the purpose.

Across a school, teachers should understand:

Why are we assessing this?

What will the information tell us?

What decision will we make from it?

How will we know whether our response worked?

That is coherence.

It is far more meaningful than forcing every subject into an identical data structure.

Don't turn formative assessment into another data collection exercise

This is where we also need to be careful.

Not every question needs recording.

Not every retrieval quiz needs adding to a central tracker.

Not every exit ticket needs analysing by SLT.

Teachers gather assessment information constantly.

A pupil's answer.

A mini-whiteboard.

A misconception heard during discussion.

A paragraph written in class.

A quiz.

A question that half the room cannot answer.

Much of that should simply change what the teacher does next.

There is an important distinction between day-to-day formative assessment, which informs ongoing teaching, and in-school summative assessment, which helps schools understand performance after a period of teaching.

If we attempt to formally record every piece of formative assessment, we risk destroying the very thing that makes it useful: its immediacy.

Sometimes the correct response to assessment is simply:

They haven't understood this. I'm teaching it differently tomorrow.

No spreadsheet required.

So where does technology actually help?

This was one of the reasons we built LoftTrack.

Not because I thought teachers needed another dashboard.

They absolutely don't.

I wanted to remove the part of assessment analysis that technology can do much faster than a human — collating marks, identifying patterns and organising question-level information — while leaving the important decisions with the teacher.

LoftTrack takes assessment information and allows teachers and leaders to examine performance at question, pupil and cohort level, including comparisons between pupil groups. It helps identify stronger and weaker areas so that schools can move from a set of results to meaningful next steps.

But the software isn't the interesting part.

The interesting part is what the teacher can now ask.

Instead of spending an evening building a spreadsheet, I want that teacher thinking:

Why did they struggle here?

Is this knowledge or application?

Is this a whole-class issue?

Who needs targeted support?

What am I changing next lesson?

That is professional thinking.

Technology should create more time for it, not replace it.

Group data needs careful handling too

There is another valuable layer to assessment analysis.

School leaders quite rightly want to know whether particular groups of pupils are experiencing different outcomes.

Disadvantaged pupils.

Pupils with SEND.

EAL pupils.

Boys.

Girls.

Different teaching groups.

But again, the headline number is the beginning of the conversation.

Not the end.

Suppose disadvantaged pupils have a lower average mark.

That tells me there is a difference.

It doesn't tell me why.

The more useful questions are:

Where is the difference appearing?

Is it across the curriculum or concentrated in particular areas?

Is it consistent across classes?

Is it knowledge, vocabulary, reading demand, attendance, question interpretation or something else?

Is the pattern actually shared by the whole group, or driven by a small number of pupils?

Data can help schools identify priorities.

But data alone cannot tell a school which approach will solve the problem.

Professional knowledge of the pupils, curriculum and context still matters enormously.

Data can identify the problem. It cannot automatically prescribe the solution.

Beware the beautiful dashboard

Educational technology has made it possible to visualise almost anything.

Graphs move.

Charts filter.

Cells change colour.

Dashboards look impressive on a screen in a leadership meeting.

But I would judge an assessment system by a much less glamorous test:

Did it improve the next lesson?

If the answer is no, I don't really care how beautiful the dashboard is.

This applies to Learning Loft too.

I don't want a school using LoftTrack because the reports look impressive.

I want them using it because a teacher spots that Question 9 exposed a misconception, changes Wednesday's lesson and later checks whether pupils can now answer an equivalent question successfully.

That is the point.

What should leaders look for in an assessment system?

I would ask six things.

Does it reduce duplication?

Teachers shouldn't be entering the same information into multiple systems simply because different people want different reports.

Does it reveal more than the total score?

If all we have done is digitise a mark book, we haven't necessarily improved assessment.

Can teachers get from data to action quickly?

If analysis takes longer than the intervention it creates, something has gone wrong.

Can we see individual pupils as well as cohorts?

Whole-school averages matter.

Children matter more.

Can we look for patterns over time?

Assessment becomes more powerful when we can see whether a weakness persists, improves or reappears.

Does it preserve teacher judgement?

Software can surface a pattern.

A professional decides what it means.

That distinction should remain.

A simple model for better assessment conversations

For schools reviewing assessment this year, I would make data meetings shorter and the questions better.

Instead of:

“Why is Year 8 below target?”

try:

“What have we learned from this assessment that we didn't know before?”

Then:

“Which two findings matter most?”

Then:

“What will we do differently?”

And finally:

“When will we know whether it worked?”

Four questions.

That's enough to turn a data meeting into a teaching and learning conversation.

Because ultimately that is what assessment should be.

We don't need to know everything about every pupil all the time

I think sometimes our desire to be evidence-informed pushes schools towards believing more information is automatically better.

It isn't.

Useful information is better.

There is a huge difference.

A teacher does not need 47 metrics about a child to know what to teach next.

A head of department does not need every possible graph.

SLT does not need teachers spending Sunday afternoon formatting data that nobody will use on Monday morning.

What we need is enough reliable information to make a better decision.

Then we need to act.

Then we need to check.

What I think “data coherence” should mean for schools

The phrase is likely to become increasingly familiar as work around Key Stage 3 develops.

My hope is that schools interpret it carefully.

Coherent assessment should not mean more assessment.

It should mean a clearer thread between:

Curriculum → teaching → assessment → diagnosis → response → reassessment.

Everybody should understand why the assessment exists.

Teachers should be able to see something useful in the results.

Leaders should be able to identify meaningful patterns without generating unnecessary workload.

And pupils should ultimately experience something different because of what we learned.

That is the standard I would use.

Final thought

I've worked in education long enough to remember plenty of different approaches to school data.

More data.

Less data.

Data walls.

Flight paths.

Progress grids.

Trackers.

Data drops.

Dashboards.

Systems change.

The fundamental question doesn't.

What does this tell me about what this child needs next?

If our assessment system cannot answer that, collecting another number probably won't solve the problem.

But if we can move from a percentage to a gap, from a gap to an action, and from an action back to evidence of learning, assessment becomes something very different.

It stops being something we do to produce data.

It becomes part of teaching.

And that, for me, is exactly where it belongs.

Frequently Asked Questions
What is assessment data in schools?

Assessment data is information gathered about what pupils know, understand and can do. It can come from formal tests and examinations, but teachers also gather formative assessment information through questioning, quizzes, written work, discussion and classroom tasks.

How should teachers use assessment data?

Assessment data should help teachers identify secure learning, misconceptions and curriculum gaps, decide what needs to be re-taught or revisited, identify pupils who need targeted support and then check whether that response improved subsequent learning.

What is question-level analysis?

Question-level analysis, often shortened to QLA, looks beneath a pupil's overall test score to examine performance on individual questions. Those questions can then be linked to topics, skills or assessment objectives to help identify specific areas of strength and difficulty.

What is data coherence in schools?

In practical terms, data coherence means assessment information works together rather than existing as disconnected data points. Curriculum, teaching, assessment and subsequent intervention should form a clear cycle.

Does assessment data always need to be recorded?

No. Day-to-day formative assessment exists primarily to inform teaching. Formal recording is more appropriate where information needs to be retained, compared or reported. Trying to formally record every classroom interaction can create unnecessary workload without improving teaching.

How can schools identify learning gaps after an assessment?

Looking at individual question performance rather than only overall percentages can reveal recurring weaknesses in particular topics, skills or assessment objectives. Schools can then compare these patterns across pupils, classes and groups before deciding whether the appropriate response is whole-class re-teaching or targeted support.

What is LoftTrack?

LoftTrack is Learning Loft's assessment tracking and question-level analysis tool. It helps turn assessment information into question-level, pupil-level and cohort insights so teachers and leaders can identify strengths, weaknesses and meaningful patterns without manually building complex spreadsheets.