How to build a portfolio somebody else can actually look at

A claim invites a question and a piece of work answers one. Somebody who says they are organised has handed the reader a job: decide whether to believe it. Somebody who shows a plan that other people followed has removed the decision entirely. That is the whole mechanism, and it holds even where the work is modest, because a small real thing is verifiable and an ambitious description is not.

The second effect is subtler and matters more. Showing work changes what gets discussed. Without evidence, a conversation stays at the level of whether you can do something; with it, the conversation moves to how you did it, what you would change, and what you would try with a similar problem. That second conversation is most of what a first interview consists of, and it is where somebody forms a view of you as a colleague rather than as a candidate.

The word carries the smell of design studios, and that association narrows it unhelpfully. A portfolio is any organised body of evidence about what somebody can do, and the form it takes follows from the field. Written analysis, plans, models, recordings, code, spreadsheets that do something useful, teaching materials, documentation of an event that ran, a translation, photographs of a physical object you built: all of these function identically once they have been selected and explained.

Some work genuinely cannot be shown. Anything confidential, anything involving other people's personal information, and much of what happens inside an organisation falls into that category. The portfolio then becomes a set of write-ups rather than artefacts: what the problem was, what approach was taken, what changed, with identifying detail removed. Check what you are permitted to describe before publishing anything, and where a placement is the source, the guide on internships covers asking that question while you are still there.

Fields with no tradition of portfolios are the ones where having any evidence at all stands out most. Where nobody in a discipline expects to be shown anything, a short, well-explained set of pieces does an unusual amount of work, because it answers questions the reader had quietly assumed would go unanswered. The absence of a convention also means there is no template to fall short of, so the only standard being applied is whether the material is clear and whether it holds up when somebody reads it properly.

Almost everybody in education is already sitting on usable material and has dismissed it, on the grounds that somebody else set the task. The problem with an assignment is not its quality; it is the packaging. It was written for a reader who set the brief, knows the constraints, and marks against criteria nobody outside the course has ever seen. Stripped of that context, it becomes unreadable, and the repair is presentation rather than production.

Three changes usually convert a piece. It needs a title that means something to somebody outside the course, a short statement of what the task actually was, and the removal of everything that exists only because of the assessment. A piece can also be re-cut. Rebuilding an assignment as the version you would have made without the constraint teaches you something on the way through and produces a stronger artefact than the original was.

A portfolio is judged by its weakest visible item more than by its strongest. Readers do not average things out; they form an impression quickly and then look for confirmation, so a weak piece sitting beside a strong one invites the conclusion that the strong one was luck. Cutting is therefore the main editorial act, and most collections improve the moment the pieces their owner was unsure about come out.

Completion carries more weight than ambition. A modest thing that was finished demonstrates that you can close something, which is rarer than starting well and difficult to prove any other way. Large unfinished projects demonstrate the opposite, however impressive the intention behind them. Where an unfinished piece is genuinely the strongest thing you have, label it as in progress and say what remains, so the reader is not left working it out alone.

Volume becomes a liability in another way too. A reader with limited attention who meets a long list will look at the opening items and leave. A reader who meets a short one will probably look at all of it. The selection therefore controls what actually gets seen far more than the ordering does, which is why the cut deserves more thought than the layout.

Work rarely speaks for itself, whatever the phrase suggests. An artefact shows the outcome and hides everything interesting about it: the constraint you were working inside, the option you rejected, the thing that went wrong and what you did in response. Somebody assessing you is generally more interested in the reasoning than the result, because reasoning is the part that transfers to a problem they actually have on their desk.

The explanation does not need to be long, and it works better short. What the problem was, what you decided, why you decided it that way, and what you would do differently is enough. Writing it has a side effect worth having as well: it rehearses the account you will give out loud when somebody asks about the piece, and the difference between a fluent account and one reconstructed on the spot is audible.

This is also the part that ages well. Tools and techniques date quickly, and a piece made with something now unfashionable can still persuade if the thinking around it is visible. Judgement is what is being demonstrated, and judgement does not go out of date at anything like the speed the artefact does. It follows that an older piece with a clear account attached often serves you better than a recent one presented bare, which is a useful thing to know when the newest work is not the strongest work.

The most common failure in this whole area is neither selection nor explanation. It is that the person who built the thing tested it on the device where it was made, signed into the account that owns the files, and never tried it any other way. Send the link to yourself and open it somewhere else, on a small screen, without signing in to anything, and the problems present themselves immediately.

Format matters less than access. A personal site, a shared folder, one well-made document, or a file you attach to a message all work, and each trades control against effort. What does not work is anything asking a busy reader to request permission, create an account, or install something first. Keep a version you can attach directly as well, since the address matters as much in ordinary conversations with people in your field as it does in formal applications.

Portfolios decay quietly. The work stays exactly where it was while you improve, so a set that represented you accurately at one point begins to undersell you without anything visibly changing. The maintenance habit that holds up is replacement rather than addition: when something new is stronger than the weakest item already there, it takes that slot and the weaker piece comes out. Reviewing at the end of a term, when work has been handed in anyway, gives the habit a natural anchor.

Set beside a CV, the two divide the labour between them. The CV makes checkable claims quickly and the portfolio supports them slowly, and the strongest version of each points at the other: a line referencing a piece, and a piece explaining what the line had to compress. How the CV side of that is assembled is covered in the guide on building a simple student CV.

One more thing is worth saying about audiences. A portfolio is useful to people who are not hiring you at all: a tutor writing a reference, somebody deciding whether to introduce you to a colleague, an employer offering paid work alongside study who wonders what else you could take on. Making the work easy to look at removes friction from all of those, and the wider picture of how early work accumulates is covered across the early career pillar.