A grade 6 teacher plans six weeks on Scratch. By the end of week two, she is teaching sequencing from the start. A third of the class has never dragged a block into a script. The rest are bored. She absorbs the loss, shortens the project, and still delivers something decent by week six.
Nobody reports this. It isn’t a failure. It’s Tuesday.
She is right to handle it that way. Reporting it would cost her more time than fixing it, and there is nobody obvious to report it to. The grade 5 course was fine. The grade 4 course was fine. Check any of them on their own terms and they meet their own goals.
The loss didn’t happen inside a year. It happened at the join between two years. And nothing in the school calendar is built to look there.
Why no review finds it
Two words get mixed up, and the difference matters.
A sequence is an order. Grade 3 does hardware. Grade 4 does block coding. Grade 5 does spreadsheets. It fits on one page and looks complete.
A progression is a set of promises. It says what grade 4 owes grade 5 not the topics, but the skills the next teacher is allowed to assume. You can only write it if someone has decided what each year hands over.
Most scope and sequence documents are sequences with a progression’s name on them. They list content. They don’t list promises.
That’s why they pass review so easily. A sequence gets checked one row at a time, and every row passes.
So the problem is invisible. Not because anyone hides it, but because the check isn’t built to look between the rows.
Five questions the sector doesn’t ask
Three of these have answers. Two don’t. The two without answers are the ones worth taking into a board meeting.
Whose job includes the word “between”?
Nobody’s.
Look at any school group’s org chart. Every role covers a year, a phase, a subject or a site. There is a head of primary. There is a head of secondary. The join between them belongs to neither.
Hiring a third person isn’t the fix. The join isn’t a job. It’s a document and someone has to own it, keep it current, and update it when a teacher or a tool changes.
In most organisations, nobody has ever been given that. That isn’t carelessness. It’s what happens when a structure is inherited instead of designed.
If the curriculum is an asset, where is it written down?
A school group knows what its buses are worth. It knows what its buildings are worth. It checks both every year.
The curriculum is worth more than either, and it appears on no list. So when it loses value, nothing shows it.
And it does lose value not through neglect, but through normal staff turnover. In most schools, the real logic of the progression lives in three or four people’s heads.
The hard version for an executive: if your three strongest computing teachers left this term, you would lose something you have never valued, and you would record the loss as a recruitment cost.
Why does the damage look like good teaching?
Reteaching is the main cost of a broken join. On an observation form, it looks exactly like a teacher meeting students where they are.
A leader watching that grade 6 lesson in week two sees good practice. That is a fair description of what is happening.
It is also a teacher spending three weeks of the year repairing a gap nobody wrote down.
So the cost isn’t just hidden. It gets praised. That is why it has never been treated as a problem.
Who in this industry is paid for continuity?
This question is about the people selling, and that includes us.
Publishers price by year. They ship by year. They renew by year. Schools budget by year. Procurement compares this year’s option with this year’s alternative. Inspectors sample a year.
A progression only proves its value in year four well after the decision that would have paid for it.
So continuity is nobody’s product. Not because the sector doesn’t believe in it. Because believing in it doesn’t pay.
Any provider who claims otherwise, including this one, should be asked a simple question: what in your business model would keep continuity alive during a bad quarter?
We’re leaving that one where it is.
Content is cheap now. So what are we selling?
A good teacher with an AI model can produce a year of lesson material in an afternoon. Not great material. But not clearly worse than plenty of what schools buy today, and the gap is closing fast.
If that’s true, the valuable thing has moved. It isn’t the content any more. It’s the set of decisions about what each year hands forward.
That part can’t come from a prompt. Someone had to decide what a nine-year-old should be able to do the following year, and be ready to defend it.
Which leaves an awkward question for anyone who sells content, and anyone who reviews it. What do you do when the cheap thing is the thing on your invoice, and the valuable thing is the PDF you give away?
We don’t know. It may be the biggest open question in education publishing right now, and almost nobody is discussing it.
The better question
“Is our curriculum aligned?” is the wrong question. You can answer it honestly and learn nothing, because alignment is checked one row at a time and every row passes.
Ask this instead:
what does each year promise the next, in skills a colleague could test?
No document can answer that. Only people can, out loud, in the same room. Which is why it rarely gets asked.
The four joins, and who owns them
Broken curricula don’t fail evenly. They fail in four places.
| The join | What it looks like | What it costs | Who owns it today |
| Grade to grade | Grade 6 assumes file skills nobody wrote down or taught | Reteaching, mistaken for differentiation | Nobody |
| Teacher change | The logic left with the person. The topic list stayed | A full rebuild every few years | Nobody |
| Student transfer | A grade 8 joiner has done computing, just not yours | Mistaken for a weak student | Nobody |
| Primary to secondary | Secondary assumes zero prior knowledge, because it can’t check | Years of primary work thrown away | Nobody |
That last column is the point of this article. Four different failures. Four different costs. One owner in common.
They also stack. A student who crosses two joins loses more than twice the ground, because the second gap blocks the recovery that would have fixed the first.
And the primary-to-secondary join proves itself right. Secondary assumes nothing, so primary work goes unused, so the assumption looks correct.
If your grade 8 computing teacher had started from zero this year, would anyone in the school have noticed?
The test, and it takes one meeting
Get three teachers from three different stages in a room. Ask five questions, in this order.
- What does your year assume students can already do? Ask for skills, not topics. A topic list means nobody ever set the prerequisite.
- Who taught those things, and when? If nobody can name the year and the unit, nobody owns that assumption.
- What do you hand to the next year? Compare it with question one from the teacher above. The gap between the two answers is your problem.
- How much of your first half-term goes on reteaching? Teachers know this to within a week. Multiply it across twelve years.
- How would you place a student who joined halfway through? If the honest answer is “watch them for a few weeks,” you don’t have a progression. You have separate courses.
The gaps show up fast, because teachers already know where they are. What’s missing isn’t information. It’s a room where those five answers sit side by side, and someone whose job it is to write down what they show.
It runs both ways
Forward: the promises are written down, so the next teacher builds instead of repairs, so each year gets further, so the results get stronger, so the joins get looked after.
Backward, and this is the version already running: a join breaks quietly, so a teacher covers for it, so that year gets less far, so the year after lowers its assumptions to be safe, so the join gets wider.
Every step in that second chain is sensible. Every person in it is doing the right thing locally.
Both take about three years. The failing direction is cheaper at every single step, which is why it wins by default.
None of this is a story about people getting it wrong. Every decision in it is defensible, and that’s the hard part. The structural version is this:
we organised everything in this system year by year the jobs, the budgets, the inspections, the purchase orders, the product plans, even the words we use and then wondered why so little survives the join.
One more thing, about the words
This question is for the people who write and speak about this sector rather than run schools in it.
We have a huge vocabulary for what happens inside a year. Pedagogy. Engagement. Differentiation. Mastery. Rigour. Scope. Sequence.
We have almost nothing for what passes between years. “Transition” describes a student moving, not a skill being handed over. “Progression” has been used so loosely that it no longer separates anything.
Things without names don’t get owned. They don’t get measured. They don’t get budgeted.
If this sector produces one good word for what a year owes the next, it will do more for curriculum coherence than another decade of framework updates.
How much of what we call a twelve-year education is really twelve one-year educations placed end to end? And if someone asked us to prove otherwise, which document would we reach for?
Where KODEIT sits in this
KODEIT ICT is built as one twelve-stage progression rather than twelve separate courses. It runs from Digital Foundations through communication and safety, hardware and networking, coding, multimedia and ethics, design and programming, web development, emerging intelligent technologies, robotics and on to analytical insights. Each strand is meant to hand forward what a student meets early is there to make the later work possible, not just to come before it.
That’s worth nothing unless it holds at the joins. So take question three from the test above and point it at us, and at every provider you meet this year: for this grade, what exactly does your material promise the teacher who takes the class next year?
A programme that answers in skills is a progression. A programme that answers with a topic list is twelve courses in a box, however good each course is.
It’s a fair test. It takes four minutes. We’d rather be judged on it than on a feature comparison.
Frequently asked questions
We have a scope and sequence document. Isn’t that the same thing? Usually not. Check whether it says what each year assumes and what it promises, in skills a colleague could test. If it lists topics by year, it’s a sequence. Useful for planning, but it won’t show you where you’re losing time.
Who should own this in a school group? Owning the document matters more than the job title. Someone has to keep the promises between years written down and update them when staff or tools change. In most groups this is part of an existing role, not a new hire.
How do we put a number on it for the board? Question four of the test gets you most of the way. Reteaching weeks per teacher per year, multiplied across stages, converts into teaching days and then into money. It’s rough and it will be on the low side. It’s usually big enough to end the debate about whether this matters.
Is this a computing problem, or does every subject have it? Every subject has it. Computing has it worse. The skills depend on each other more tightly, the tools keep changing, the subject is often taught by non-specialists, and there’s less shared memory to fall back on. Several groups run this test in computing first because the answers come fastest there.
We’re a publisher. Doesn’t this argue against how our whole list is built? Partly, yes. That’s worth sitting with rather than answering quickly. The practical version: the chain of promises between your years is something you probably already own and have never written down, priced or defended. It’s also the part of your list an AI model can’t copy from a sample chapter.
Doesn’t AI make all of this cheap enough to ignore? It makes content cheap. It doesn’t make the decisions for you. Someone still has to say what each year hands over. If anything, cheap content makes that chain more valuable, because it’s the one part that’s hard to copy.