1. A narrower pitch gets you more interviews
Positioning came up across the reviews, and Nate’s short version is that it’s your answer to competition. Far more designers are applying than there are jobs. Name a specialty and people assume you’re weaker at everything else and especially good at that one thing, so you can apply to fewer jobs but your odds of an interview for each one go way up. It doesn’t have to be an industry, either. One junior resume put a technology focus right in the title line, and Nate called it the strongest thing on the page.
A grad student had two different industries in their headline and asked whether to pick one. Nate said yes: pitch both and you’re up against people who specialize in just one of them, and a hiring manager will interview them first. Then the rest of the portfolio has to back it up, with case studies from that space and its vocabulary on the page. Our slider shows the one-liner narrowing in layers, from “I’m a designer” to “I’m a product designer who builds design systems for healthcare teams” as you add a discipline, a craft and an industry.
You can overdo it. One value prop named a single kind of fix for a single kind of client, and Tim worried it would alienate the very audience it was written for. On the slider that’s the last column, where so few jobs fit that the narrowing starts working against you. Some details, like where you live, cut the number of jobs without changing how anyone sees you at all. Nate’s take is that overshooting and then pulling back a notch is the right way round. Write the narrow version, then ask whether most of the people hiring in that space would call it their problem. If only some would, widen it a step.
2. Unshipped or under NDA, the work still counts
The same junior resume was built mostly on self-initiated projects, one still in progress, and the question was whether to label them as concepts. Don’t. “Conceptual” marks you as new to the industry before anyone reads about the work. If you did real work with real people, it’s experience, even if it never launched or earned ten dollars. The further you took it (talking to real people, putting it out as a side project), the easier it is to list. Give the role a title that matches the work and the direction you want to go, and if it’s ongoing, end the dates in “present”, the way you would for a job you still have.
Honesty is the other half. A resume isn’t a background check, but people may look you up, so it helps if your first entry is something they can verify. Anything that looks off once someone checks can cost you their trust, which is one reason we steer people away from concept projects for big, famous brands. And you need to be able to talk about every line with confidence if someone asks.
Another designer was working around an NDA and a non-compete. Nate pointed out those are different contracts: a non-compete limits which industry you can work in next and doesn’t necessarily stop you showing what you did. Tim’s rule of thumb, and he was clear it isn’t legal advice, is that work that’s already public, say in a press release or a public changelog, is usually work you can show. Use screenshots of the live product, talk about the project at a high level, and you can probably still give rough estimates in place of exact numbers.
3. No numbers yet? Show your math
Numbers came up from two directions. One designer’s projects had been sitting in user acceptance testing on long, red-tape-heavy timelines, so there was nothing to measure yet. Another had resume bullets built around output, like how many screens they designed. Tim’s answer to both: a business doesn’t care whether it took one screen or a hundred. It cares whether users behave differently now, and how close that change sits to money.
The value scale we showed twice runs from awards, testimonials and output at the low end, through adoption and use over time, up to behavior change. The closer the behavior change gets to financial impact, the better: error reduction, retention, conversion. (Tim thinks “time saved” is bogus. Nate’s fonder of it.)
When nothing’s shipped, tie each decision back to the problem that caused it, then estimate the benefit. Tim’s example: say you changed one thing that’s meant to cut an error by maybe 25%. On a site with 100,000 users a week, that’s around 25,000 people who won’t make that mistake. Be honest that it’s a projection, say the work is in testing, and leave it there. The same shape works for a resume bullet: something the business cares about, how you measured it, and what you did to drive it. That’s the XYZ formula.
4. Write your headings for people who skim
Nobody reads case studies. Recruiters and hiring managers scroll, read the headings, look at the images, and maybe stop at the one section that’s relevant to them. If the story isn’t in your headings, most of them won’t see it.
A trick we use a lot: delete any heading that only names a step in your process, like “Testing focus”, and promote the first sentence of the paragraph underneath. Someone who reads nothing but headings should still learn what you did and what you found. If storytelling feels hard, narrate what happened and put yourself in it. “I observed consistent friction points across all interviews” works as a heading on its own, and it puts you in the owner’s seat. A case study is the pitch for hiring you. The product is the vehicle.
Then look at your image-to-text ratio. If you’ve written the draft, you’ve done the hard part, and what’s left is making it easy to take in. Nate pointed to another portfolio that day as the model: lots of text, laid out visually, still easy to skim. Lists are the easy place to start, so ask whether each one could become a visual. And give a strong image room of its own instead of squeezing it into a row of three.
5. Don’t ship the AI’s defaults
We’re fans of AI, and no portfolio needs to hide that it was used. The trouble is output nobody chose. Tim recognized one portfolio’s case study copy as model-written because he sees that phrasing in his terminal all the time. Another site had a handful of defaults: a stock-looking search field, a command-K hint, small arrows, all-caps eyebrow labels. Lots of visitors won’t notice. The ones who do may start wondering how much of the work is yours.
We push for what we call intent coding. Vibe coding is letting the tool think for you and taking the first thing it gives you. Intent coding means starting with a plan and a shape in mind, pushing back, and editing the output the way you’d edit a frame you just dropped on a canvas. These tools were trained on everything. They don’t know your problem.
AI can still help with writing you’re stuck on. Nate fed one case study into NotebookLM and asked for a slideshow. The images were invented, but the section titles it wrote were the story-carrying kind from the last section. Use a draft like that to find the shape, then rework the wording before it goes out, since Tim finds Claude’s prose full of patterns that people read as unpleasant or suspicious. And go back through your UI looking for anything you didn’t pick on purpose.
We also covered smaller resume fixes: lead each role with the job title unless the company is a name people know, and add a phone number where that’s the norm, because it signals you’re a real person and recruiters sometimes just call. For big companies that interview in one long block, the portfolio mostly gets you to the recruiter, and the rest usually rides on a presentation of two outcome-heavy case studies. A personal about page matters when you’re hoping for referrals. And in the Q&A: hiring usually slows in November and December, so it’s a good stretch to get your materials ready. Apply on the company’s own site rather than through Easy Apply, and use LinkedIn for networking and referrals.