Most people planning a career change open with the wrong question. They ask what they need to learn, and the answer sends them off for twelve months of study that nobody in their target market ever asked for. Before you pick a course, you need a spec, and that spec is sitting in public on job boards right now.
The gap you imagine is not the gap on the job ad
From the outside, a new field looks like a wall of technical unknowns, because the unknowns are the only part you can see. The parts of the job that would be familiar to you are invisible, precisely because they are familiar. So the operations manager moving toward product management signs up for a SQL course and spends the winter on window functions, while the ten postings she is actually targeting ask for discovery interviews, roadmap writing, stakeholder pushback and forecasting. She has been doing all four under a different job title since 2019.
Over-correction is the default failure mode of a career change, and it is not caused by laziness. It is caused by the opposite. A syllabus is a comfortable place to hide. It has a start date, a progress bar, an end. Applying with a profile you are not sure about feels like standing in traffic. Studying feels like progress in a way that auditing never does, and that is exactly what makes it dangerous this early.
Run the audit first: ten postings, one tally
This takes an evening. Not a weekend, an evening.
- Fix a target that is a job title and a seniority, not a field. "Product manager, mid-level, B2B SaaS" is a target. "Tech" is not.
- Collect ten live postings you would genuinely apply to this month. Paste them into one document. Ten is the minimum that makes the tally mean anything, and it is also the point where the repetition starts to feel absurd, which is the useful part.
- Delete the company blurb and the values paragraph. Keep only the requirements and the day-to-day responsibilities.
- Tally every requirement by how many of the ten postings mention it.
- Anything that shows up in seven or more is your real spec. Anything in two or three is one hiring manager's preference and you can ignore it.
Ignore the nice-to-have list entirely on the first pass. It is written by people who are describing their fantasy hire, and treating it as a checklist is how career changers talk themselves out of applying. If you want the wider view of where this audit sits in the whole move, the step-by-step career change playbook covers the sequencing around it.
Now sort your tally into three buckets. The sizes will surprise you.
Bucket one: things you already do under a different name
This is the biggest bucket, usually by a wide margin, and it is a translation problem rather than a learning problem. "Stakeholder management" is the eighteen months you spent chasing three department heads for sign-off on a budget nobody wanted to own. "Data-informed decisions" is the retention spreadsheet you rebuilt because the reporting tool was lying. "Enablement" is training the new hires. "Resource planning" is the shift rota.
The test is blunt: if you have done it more than twice, with a real consequence attached when it went wrong, it belongs to you. You just call it something else. Write the posting's phrase in one column and your version in the next, and keep that mapping, because you will use it on your CV, in your cover letter and out loud in an interview. Postulit, which turns a LinkedIn profile into a CV, is useful at this stage mainly because it makes you look at your history as a list of discrete outputs instead of the story you have been telling yourself about it. The career change cover letter is where that mapping has to be most explicit, since a recruiter will not do the translation for you.
Most of the gap disappears here. Not all of it. But when people say they need two years of retraining, the honest number after this bucket is often a few weeks.
Bucket two: things you can evidence in a few weeks
These are narrow: a specific tool, a reporting format, a framework, a procedure. Measure them in what you can produce, not in hours studied. A course on a tool is bucket two only if you can practise on the real thing and ship something small with it by the end of the month.
The output is the point. One dashboard someone else uses, one written teardown, one small piece of work with a link. Building a portfolio during a career change is not a side quest here, it is how bucket two gets closed at all.
Bucket three: the genuine blockers
There are usually one or two. Sometimes zero. A licence. A regulated qualification you cannot sit without supervised hours. A system you cannot touch without employer access, which is why clinical software, trading platforms and some industrial instrumentation stay closed to outsiders no matter how motivated they are.
These are the only items that deserve a year and real money. Once you know which one you are facing, the format question becomes answerable and bootcamp versus degree stops being an abstract debate. If the blocker carries a fee and lost income, sort out how you will fund the switch before you enrol, not after.
A skill is not closed until someone outside your head can verify it
This is the half that people skip. You can be genuinely competent and still be unhireable in the new field, because the recruiter is not assessing your skill. They are pricing risk, and they have twenty minutes.
What counts as evidence, ranked honestly:
- Paid work in the new field, however small. One invoice from one client for one weekend of work outranks everything below it, because somebody who had a budget already took the risk on you and did not lose. Rate, scale and prestige barely matter. The transaction is the proof.
- A real artefact with a link. Something a stranger can open and judge in ninety seconds without you narrating it. Published, dated, yours.
- A volunteer or internal project with a named stakeholder. Named is doing the work in that sentence. "I helped the marketing team" is nothing. "I rebuilt the reporting for Sarah in marketing, and she will take a call about it" is real.
- Course certificates. A long way behind the first three.
What a certificate actually buys you
A certificate alone closes nothing, and I would rather say that plainly than be diplomatic about it. It proves you attended and completed an exercise where the problem had already been shaped for you, the data was already clean, and there was no stakeholder changing the requirements in week three. Every hiring manager in the field knows this, because they have hired the person who had four certificates and could not start a piece of work without being told how.
It does two useful things. It gets a keyword past an automated filter, and it gives you a structure that produces artefacts, if you keep the artefacts. So finish the course, then put the thing you built at the top of your CV and the certificate in a line near the bottom. You will also have to narrate the whole sequence out loud, which is its own skill, and explaining the pivot in an interview is easier when the audit is what you are describing rather than your feelings about the change.
The sequencing rule
Audit, translate, then learn only what survived. In that order, every time. The translation step comes before any spending decision because it usually removes most of the reason to spend.
Then re-run the tally every four to six weeks against fresh postings. Requirement lists move, and they move fastest in exactly the fields people are changing into. A skill that appeared in nine of ten postings in spring can drop to four by autumn, and a tool nobody mentioned becomes standard in a quarter. If you built a twelve-month plan against a snapshot, you are studying for an exam that has been rewritten.
Start tonight: ten postings, one tally, three buckets. Then send two applications on Friday against the translated version of what you already do, before you have learned a single new thing. The replies are data too, and they are cheaper than a course.