Who actually starts a coding bootcamp.

If you are considering a coding programme, the question underneath all your other questions is usually some version of am I the kind of person who does this. It is rarely asked out loud, because it sounds like a confidence problem rather than an information problem. It is an information problem. Nobody publishes the answer, so people imagine one, and the imagined answer is almost always a twenty-three-year-old who is already quite good at computers.

We had the actual answer sitting in our own database, so we counted it. 553 students who enrolled and paid, 456 of them with a payment on record. Not applicants, not leads, not people who filled in a form and drifted away. People who made the decision. Here is what they look like.

One note on method before the numbers, because it matters for how much weight to put on each one. The denominators are different for different figures, and we have carried each one with its figure throughout. Transactional facts like format and payment are recorded for nearly everyone. The onboarding survey is voluntary, so the background questions were answered by 113 people and the confidence questions by 62. Where a number rests on 62 responses we say so, and you should treat it as directional rather than precise.

Almost nobody arrives knowing how to code

This is the finding people find hardest to believe about themselves and easiest to believe about others. 56% had written no code at all, and 87% were at or near zero. Exactly one person in the whole sample described themselves as a working developer, and we suspect they were being modest about something.

They are credentialled, just not in this

The lazy assumption about people who retrain is that they are escaping a lack of qualifications. The records say the opposite. 61% already hold a degree and a further 15% hold a diploma, with only 15% at high school or earlier (n=113). What they do not hold is a computing qualification: 73% studied something else.

And then there is the figure we did not expect and have thought about more than any other in this dataset. 27% of these students had already studied computer science or IT, and enrolled anyway.

Sit with that for a second, because it is not a comfortable finding for anyone in education, us included. More than a quarter of the people paying for a practical software programme have already completed a formal computing education. Whatever a CS degree delivers, a meaningful proportion of its graduates do not feel it delivered the ability to walk into a job and build things. That reframes the usual bootcamp-versus-degree argument, which we set out at length in our comparison of the two paths.

There is no pattern in what they do now

68% came from a job outside technology entirely, 18% were already in a technical role of some kind, and 15% were still studying (n=62, multi-select). We went looking for a dominant profession in the occupation field, some cluster we could name in a marketing page. There is not one. The list below is a sample of what is actually recorded, and its lack of shape is the finding.

The bit that surprised us: they are confident and afraid at once

At onboarding we ask people to rate, from 1 to 10, how confident they are of completing the programme. The answers are almost comically positive. Half say 10 out of 10. Nearly nine in ten say 8 or above. Not one person in the sample rates themselves below 7 (n=62). On a separate question, 87% rate their commitment to the goal at 5 out of 5 (n=61).

On the same form, a few questions later, there is an open text box asking what they are most nervous or uncertain about. We read all of them and coded each into a single primary category. The same people who had just rated themselves 10 out of 10 said this.

39% named their own ability to learn as the single thing they were most afraid of. Not the job market, which came in at 13%. Not whether the money would be wasted, also 13%. Themselves.

And these are people who had already paid. This is not the hesitation of somebody browsing, which could be dismissed as tyre-kicking. It is what remains after the decision has been made and the money has left the account. Self-doubt is not spent at checkout. It comes with you into week one, which is a genuinely useful thing for a school to know about its own students and a genuinely useful thing for you to know about the people who will be sitting next to you.

If that is the thing standing between you and starting, the honest practical answer is not encouragement, it is evidence. Spend a week actually doing it and find out. That is most of why our free trial exists in the shape it does: it is the cheapest available experiment on the question 39% of our students said they were most worried about.

They buy the format they said would not work for them

Asked how they learn best, the answers are emphatic. 97% say hands-on projects, 65% say live discussion, and only 32% say reading. Asked what keeps them going, 87% name mentor feedback, ahead of accountability check-ins at 65% and solo progress tracking at 47% (n=62, multi-select).

Then look at what the same population actually bought. 55% chose part-time or self-paced formats. Only 9% chose full-time in person (n=520 of 553), which is the format that most reliably delivers the live discussion and mentor feedback they just told us they need.

There is no contradiction to scold anyone about here. People have jobs, families and rent, and our application data shows 75% of applicants have under 25 hours a week available. The constraint is real and the choice is rational. But it does mean a large share of learners are buying the format that gives them the least of what they themselves identified as the thing they need, which is worth knowing when you choose yours. If you go part-time or self-paced, the feedback loop is the thing you have to construct deliberately, because it will not arrive on its own.

The money filter nobody talks about

77% of enrolled students paid the entire tuition in one payment, with 23% splitting it across two or more (n=456). Held next to another number from our applications research, that becomes uncomfortable: among applicants, 75% said they could not pay a lump sum, and 47% were only looking for free resources (n=262).

Put those side by side and the shape of the industry becomes visible. The people who enrol are disproportionately the ones who could pay at once. Ability to pay upfront is quietly acting as a filter on who gets to retrain at all, and it is filtering on something that has no relationship to aptitude. We do not have a tidy answer to that. It is a large part of why we publish the entire curriculum in public and run a free crash course, and it is the honest context behind our scholarships and payment plans. It does not solve the filter. It moves it a little.

Not everyone is here for a job

The bootcamp conversation is framed almost entirely around employment, and the goals data says that framing is too narrow. A stable full-time job is the most common answer at 62%, but half name remote work, a third want freelance or client work, and a quarter intend to build their own thing (n=113, multi-select). 73% name AI as the area they want to work in (n=62).

What this does not tell you

An honest read of any dataset includes what it cannot support, so here is ours, plainly. There is nothing in this analysis about outcomes. No placement rates, no time to hire, no salaries, no employer names. That is deliberate and permanent. Outcome claims require a standard of evidence we cannot meet from enrolment records, and the industry is full of outcome numbers that quietly cannot either. These records describe who enrols and what they believe when they do. That is the entire scope.

Second, this is one school’s students, most of them Malaysia based, and it is therefore a picture of who chooses us rather than a census of everybody retraining. Third, several figures rest on 62 or 113 voluntary responses, which is why every number on this page carries its denominator and why we would rather you treated the small-sample ones as a strong hint than as a measurement.

The full report, with the methodology written out properly, is at The Confidence Gap, and the companion study of what stops people before they enrol is at What Stops People Retraining for Tech. Both are published under CC BY 4.0, so you can quote or republish the figures with attribution. We only ask that the sample size travels with the number.

If you are trying to work out whether you belong here

The composite person in this dataset is around 28, employed in something unrelated to technology, holds a degree in something else, has never written a line of code, learns by doing, needs feedback from a human, has under 25 hours a week, and is privately worried they are not clever enough for this. That is not the exception in the room. That is the room.

Which does not mean it will work out for everyone, and we are not going to pretend the data says that. What it does mean is that the specific fear that you are unusually unprepared is not supported by the record. Whatever you decide, decide it on the real distribution rather than the imagined one.

And whichever direction you go afterwards, employment, freelance, your own product, or applying it inside the job you already have, the base is the same base. Coding fundamentals with AI in the workflow pay off in all four directions, which is why we would rather you tested the thing than agonised over it. A week of the free trial costs one signup and no card, and it answers the question this whole article is really about.