# The New Portfolio Structure > For the individual reader. A portfolio of screens now proves you had access to > a tool and a reviewer. Five sections, each proving something a hiring manager > cannot get from looking at a grid of mockups. Build it once, maintain it in > twenty minutes a month. --- ## What a portfolio is for now The reader has one question: **would this person's judgment make my product better?** Screens answer a weaker question, which is whether you can produce competent work, and that question got much easier to answer yes to. So structure the portfolio around decisions, and show the work as evidence rather than as the subject. --- ## The five sections ### 1. The decision (three of these) The core of the portfolio. Each one, one page, this shape: - The situation, in three sentences. What was at stake, what was unknown. - The options that were real. Show them. Two or three, with the appeal of each stated honestly. - The call, and the constraints that produced it. Rules with failure conditions, not principles. - What it cost. What you gave up, who disagreed, what you did about that. - What happened. Numbers if you have them, honest description if you do not. Pick three where the answer was not obvious. A decision with an obvious answer proves you can identify obvious answers. **Include one you got wrong.** Not performative humility, the actual mechanism: what you believed, what you saw, when you found out, what changed in how you work. Interviewers have been trained by a decade of polished portfolios and this section reads as real within two sentences. ### 2. The specification One artifact you wrote that let other people, or a system, produce work you would have made yourself. A constraint set. A rubric. A range spec. A refusal set. Show it as it was actually used, in the ticket or the brief, not reformatted into a case study. Add one line: what it was like before this existed. That contrast is the argument. ### 3. The thing you killed One feature, ritual, or piece of process you removed, and what replaced it. Killing well is a senior skill and almost nobody shows it. Include how you got agreement, because that is the hard part and everyone reading knows it. ### 4. The wrong path One failure state you designed properly: detection, disclosure, recovery, copy. This section is close to empty in most portfolios, which is exactly why it works. It is also the fastest way to signal that you have shipped something real, because people who have not shipped do not think about this. ### 5. The craft page One page. Your best-made work, shown without narrative. Do not skip it. Craft is necessary, and a portfolio that is all reasoning and no made work reads as someone who talks about design. One page is enough to answer the question. --- ## What to cut **Process diagrams.** Everybody's double diamond looks the same, and it describes a method rather than your judgment. **Every screen of a flow.** Show the two that carried the decision. **Team credit paragraphs.** One line naming your role honestly is enough. Long credit sections read as hedging about what you actually did. **Redesigns of famous products nobody asked for.** They demonstrate production capability, which is no longer the scarce thing, and they carry no real constraints. --- ## The one-page version Most readers will spend two minutes. Give them a top page with: - One line on what you are good at, specific enough to be wrong. - Three decision titles, each a sentence with the outcome in it. - One artifact, linked. - One thing you got wrong, in a sentence. If that page is good, they read on. If it is a hero image and a tagline, they do not. --- ## Maintaining it Twenty minutes a month, calendared. Add one decision while it is fresh, in the same shape you would write a decision record. The portfolio then assembles itself from work you were doing anyway, which is the only version of portfolio maintenance that survives a real job. The decision record and the portfolio entry are close to the same document. If you keep records, you already have a portfolio and just have not formatted it. --- ## For people early in their career The obvious problem: no shipped decisions yet. Two things that work, and one that does not. **Works:** analysis of a real product's decisions. Pick a product you use, find a decision they made, reconstruct the options they must have had, argue about the constraints, say what you would have done. This directly demonstrates the skill being hired for. **Works:** the specification section, built on personal or open-source work. A rubric for a small thing is still a rubric. **Does not work:** a redesign. It shows production and hides judgment, and it is what everybody else is submitting. --- ## The one thing to do this week Write one decision page for something you did in the last six months. One page, the five parts. It will take an hour and it will be the strongest thing in your portfolio, which tells you what the rest should look like. --- _From "The New Portfolio", chapter 21 of The Design Operating Model._ _falkster.com/design/the-new-portfolio_