How to Build a Portfolio When You Have No Job Experience

The hardest version of the job search is the one where every posting asks for two years of experience in the thing you are trying to get two years of experience in. Courses do not break the loop, because a certificate proves attendance, not capability. Work does — and you can do work before anyone hires you.

What a portfolio is for

It is not a gallery. Its job is to let a hiring manager skip a step in their risk assessment: instead of inferring whether you can do the work, they can look at you doing it. That means three good pieces beat eleven mediocre ones, and each piece needs the reasoning visible, not just the output.

Four sources of real work

1. The problem you actually have

Build the thing you personally need. A budget tracker, a scraper for the flats you were searching, a spreadsheet that priced out a trip. Genuine problems produce genuine constraints, and constraints are what make a project interesting to talk about.

2. Public data, specific question

Government statistics offices, open transit feeds, national health datasets, sports records. The trick is a narrow question: not "an analysis of the census", but "which of this city's neighbourhoods lost the most under-35s in ten years, and what happened to rents there".

3. Small organisations that cannot afford anyone

A local charity, a sports club, a school, a family business. They have real needs and no budget. Do one bounded piece of work — a website, a donor spreadsheet, a stock system — with a fixed end date agreed up front, and get a testimonial in writing when it is done.

4. Rebuild something badly done

Take a public report, dashboard or interface that is genuinely poor and redo it. Publish both versions side by side with the reasoning. This works especially well for design, data visualisation and technical writing, because the comparison does the arguing for you.

Portfolios outside technical fields

Portfolios are treated as a software and design convention, which leaves everyone else assuming the idea does not apply. It does; what changes is the artefact.

Operations and administration. The artefact is a process. Take something you have run — a rota, an inventory, an onboarding sequence, an event — and publish the before and after: what it was, where it broke, what you changed, what the new version costs to run. A one-page process map and a short write-up is enough.

Marketing. Do the work for a real organisation, however small, and show the reasoning as well as the output: who you were writing for, what you tested, what you would do next. Where results cannot be shared, a teardown of a public campaign with your rewritten version alongside it does the same job.

Care and support work. The portfolio here is evidence rather than product: completed mandatory training, a reflective account of a difficult situation and how it was handled, and written references from anyone who has supervised you, including voluntary work.

Teaching and training. A session plan, the materials that go with it, a short recording where the setting allows it, and a paragraph on what you would change after running it. Training a colleague counts.

Two cautions. Where you handle people's data — health, care, education, finance, HR — strip identifying detail and say that you have. And where regulated or public sector employers recruit against a published person specification, as in the UK and Ireland, it is evidence mapped criterion by criterion that gets scored; there the portfolio supports the application rather than replacing it.

Write it up properly

For each project, answer six questions in under 600 words: what was the problem, who had it, what did you try first, what went wrong, what did you end up doing, and what would you change now. That last question is the one interviewers remember, because it is the only part a copied project cannot answer.

Include the mess. A project write-up that says "the first approach failed because the data had duplicate IDs and I had assumed it was clean" is more convincing than a flawless narrative. Everyone experienced has broken something.

What that looks like, compressed

Here is the shape those six questions produce for a project of the third kind above — a payment tracker for a volunteer-run sports club.

The problem. The treasurer tracked members' subscriptions in a notebook and chased late payers from memory. Every season a few people paid twice and a few never paid at all.

Who had it. One volunteer, with no software budget and no time to learn a new tool.

What I tried first. A shared spreadsheet, one row per member and one column per month. Correct and unusable: three dozen columns, and the treasurer worked from a phone.

What went wrong. The design assumed a desk. It also assumed everyone paid monthly, when a third of members paid once a year at the start of the season.

What I did instead. One row per payment rather than one per member, entered through a short form on a phone, with a summary sheet showing who was behind and by how much. Entry took seconds, and the arithmetic stopped being anyone's job.

What I would change. I would have watched the task being done once before designing anything. I built for the data structure rather than the moment of use, and the redesign was avoidable.

Under two hundred words, and it evidences a real user, a failed first attempt, a diagnosis and a lesson that transfers.

Where to put it

Anywhere stable and public: a simple site, a repository with a decent README, a public document, a PDF you attach. Avoid platforms that require a login to view. Put the link in your CV header, not buried at the bottom, and name it plainly — "Portfolio: three projects" beats a bare URL.

How long this takes

Realistically, three projects at a reasonable standard takes eight to twelve weeks alongside other commitments. That is faster than most qualifications and considerably more persuasive. Start with the smallest one, publish it before it is perfect, and let the second and third be better.

Nobody has ever been rejected for a portfolio that was too honest about its limitations. Plenty have been caught out by one that was not theirs.