Most advice about building a graphic design portfolio tells you to show your best work. That is technically true and practically useless. Every designer thinks they are already doing that. The real problem is different: portfolios fail because they show too much work with too little context, arranged in an order nobody chose deliberately.
At Cantonax we sit on both sides of this. We review portfolios when we hire, and we watch clients review portfolios when they pick a partner. The pattern is consistent and slightly uncomfortable for designers: the person deciding rarely looks at more than six projects, and they almost never scroll to the bottom. They decide fast, then look for reasons to confirm the decision.
So this article is not a list of inspiration links. It is a breakdown of the exact pieces, case study details, and presentation choices that move a reviewer from browsing to emailing you.
How clients and art directors actually review a portfolio
Before the list, understand the review behaviour you are designing for. In practice, a portfolio review happens in three passes:
- The scan (10 to 40 seconds). Homepage or first three PDF spreads. The reviewer is asking one question: does this person work at the level we need? Nobody reads a word yet.
- The relevance check (2 to 4 minutes). They hunt for something that resembles their own problem. A restaurant owner looks for food and hospitality. A SaaS founder looks for interfaces and decks. If they cannot find an analogue, they leave.
- The proof read (5 to 15 minutes, only if you survived). Now they read. They want to know what you did versus what the team did, what the constraint was, and whether the work moved anything.
Almost all portfolio advice optimises for pass one. The work that actually gets hired is optimised for passes two and three, because that is where the money conversation begins.
The contrarian core: fewer projects, more context
A portfolio with 25 projects signals volume. A portfolio with 7 projects and real narrative signals judgement. Judgement is what clients are buying, because they cannot evaluate your kerning but they can absolutely evaluate whether you understood a business problem.
Cutting is the hardest edit you will do. Use this rule: if you would not be delighted to receive a brief exactly like that project again, remove it. Your portfolio is a magnet, not an archive. Whatever you show, you get asked to do again. framer.com has covered this at length.
The 10 pieces clients actually want to see in a graphic design portfolio
You do not need all ten. You need six to eight of them, chosen to match the clients you want. Each one answers a specific unspoken question in the reviewer’s head.
1. A full brand identity system, not a logo
Logos in a grid prove nothing anymore. What proves capability is the system: primary and secondary marks, type hierarchy, colour with usage rules, iconography, photography direction, and at least three real applications.
Show one spread of the logic (why this direction fits the audience), then the applications. The unspoken question you are answering: can this person hand our team something we can actually run with after they leave? A comparable breakdown sits on wix.com.
2. Packaging or print, photographed in the real world
Flat mockups from a template pack are transparent to anyone who has seen 200 portfolios. A single photograph of your work printed, folded, on a shelf, or in someone’s hands outperforms ten glossy renders.
- Show the dieline or flat artwork next to the physical result
- Mention the substrate, finish, and print method if you specified them
- Include one imperfect production shot on press or at the printer
This is one of the fastest trust signals in a portfolio because it proves the file left your computer and survived contact with reality.
3. A multi-page layout project
Editorial design, an annual report, a brand guidelines book, a pitch deck, a catalogue. Anything that demands hierarchy across many pages. Clients with content problems, which is most of them, need to see you can hold a grid together over 40 pages without it collapsing.
Show the master grid and a spread that breaks it deliberately. That single slide separates people who use templates from people who build systems.
4. One campaign shown across formats
Take one concept and show it living in five places: a social carousel, a story format, a display banner set, an email header, and something physical or out of home. The point is not the concept. The point is consistency under different constraints, which is what a marketing manager pays for every month.
5. A digital or web design piece with hand-off evidence
Even if you are not a UI designer, most clients will ask you to touch something digital. Include a landing page or a small website with:
- Desktop and mobile side by side
- A component or style snippet (buttons, forms, states)
- A short line on how you handed off: Figma file, spacing tokens, exported assets
Hand-off evidence quietly tells a business owner that working with you will not create extra work for their developer.
6. A before and after redesign
The single most persuasive slide type for non-designers. Clients cannot evaluate a beautiful thing in a vacuum, but everyone can see improvement. Use it for a rebrand, a packaging refresh, a document overhaul, or a self-initiated redesign of something broken.
Rules that keep it honest: show the old version at the same size and quality, do not mock the previous designer, and explain what constraint forced the old version to look that way.
7. A typography-led piece
One project where type does the heavy lifting and there is nowhere to hide. A poster series, a wordmark, a book cover, a lettering piece. Art directors use this as a craft check. It takes them four seconds and it is the difference between a shortlist and a polite no.
8. A project with an ugly constraint
Include one project where the budget was tiny, the deadline was 48 hours, the brand colours were fixed and awful, or legal removed half your copy. Then explain how you solved it anyway.
This is counterintuitive, but real clients have constraints and they are terrified of designers who only perform under perfect conditions. Constraint stories are the most quoted slides in internal hiring conversations.
9. A personal passion project, framed correctly
Conventional advice says cut all personal work because it is not real. We disagree, with one condition: frame it as a brief, not as a mood.
Unframed personal work looks like a hobby. Framed personal work looks like initiative. Compare:
| Weak framing | Strong framing |
|---|---|
| “A poster series I made for fun.” | “Self-initiated: 12 posters in 12 weeks to build a repeatable system for fast-turnaround event promo. Format and grid reused later for a paying client.” |
| “Concept rebrand of a famous brand.” | “Unsolicited concept: the current identity fails on small mobile placements. Here is the diagnosis, the fix, and the test at 24px.” |
| “Personal illustration.” | “Built an illustration language I now offer as an add-on service. Applied here to three sample use cases.” |
Passion work also does something client work cannot: it shows what you would choose if nobody was paying you. Clients who want a distinctive point of view are specifically looking for that signal. This overview fills in the gaps.
10. One project with numbers attached
You only need one. A sales lift, a conversion improvement, a crowdfunding total, an event attendance figure, a reduction in production time, a decrease in support tickets after you redesigned a form.
If you have no numbers, ask a past client for one sentence by email. If they cannot give a metric, use a qualitative outcome with a name attached: “The new deck was used in 40 investor meetings and the founder credited it in the seed announcement.” Specific beats vague every time.
What each case study must contain
This is where most portfolios lose the job. The images are fine. The context is missing. Use this structure for your top three projects, and a shortened version for the rest.
| Element | Length | Why it matters to the client |
|---|---|---|
| Client and sector | 1 line | Instant relevance check |
| The problem | 2 to 3 lines | Proves you design for a reason, not for a look |
| Your role | 1 line | Removes doubt about team credit |
| Constraint | 1 line | Shows you work in the real world |
| Key decision | 2 lines + 1 visual | Demonstrates judgement, the thing being bought |
| Final work | 5 to 9 images | The craft evidence |
| Outcome | 1 to 2 lines | Turns a nice project into a business case |
| Timeline and deliverables | 1 line | Helps the client imagine scoping you |
How much process to show
Show one decision, not the whole journey. Nobody wants 14 sketch scans and a mood board. They want the fork in the road: two directions, the one you chose, and the one sentence explaining why. That is the difference between a process section that builds confidence and one that gets skipped.
Presentation choices that quietly decide the outcome
Order matters more than quantity
Put your strongest project first, your second strongest last, and your most relevant to that specific client in position two. If you are pitching a hospitality brand, the hospitality work is not buried in slot nine.
Yes, this means maintaining variants. A tailored 8-project PDF sent to a specific prospect converts dramatically better than a link to everything you have ever made.
Website, PDF, or Behance: use all three, differently
| Format | Best for | Practical notes |
|---|---|---|
| Own website | Inbound clients, search visibility, credibility | Own the domain. Fast load, images around 2000px wide, WebP, real project pages so Google can index them |
| Tailored PDF | Pitches, applications, email follow-ups | 16:9 landscape, under 15MB, hyperlinked contact page, file named FirstnameLastname-Portfolio-ClientName.pdf |
| Behance / social | Discovery and reach | Treat as a feeder. Always link back to your own site |
The technical details reviewers notice without noticing
- Load speed. If your homepage takes six seconds, a third of reviewers are gone before the scan starts
- Mobile. A large share of first views happen on a phone, often forwarded in a message thread. Test every case study vertically
- Image consistency. Same background, same shadow treatment, same crop logic across projects. Mixed mockup styles read as chaos
- Contact on every page. Email, not just a form. Some clients will not fill a form
- A named human. A real photo and two sentences about how you work outperforms a stylised avatar and a poem
Your homepage above the fold
Three things, nothing else: what you do, who you do it for, and three thumbnails. Example: “Brand identity and packaging for food and drink companies.” Specific positioning loses you the wrong enquiries and wins you the right ones at a higher rate.
What to leave out
- Student exercises, unless the craft is genuinely exceptional and clearly labelled
- Spec work grids with 40 unused logo concepts
- Work you cannot explain because a director made every decision
- Anything under NDA that you have not cleared, which is a live risk to your reputation
- Long “my design philosophy” intros. Nobody has ever been hired because of a philosophy paragraph
- Autoplay audio, heavy intro animations, custom cursors that hide the work behind a performance
- Undated dumps. If your newest project is four years old, clients assume you stopped working
A practical 14-day plan to rebuild your portfolio
- Days 1 to 2: List every project. Score each on craft, relevance to target clients, and whether you want more of it. Keep the top eight
- Days 3 to 4: Email three past clients asking for one outcome sentence or number
- Days 5 to 7: Write the eight case studies against the structure table above. Text first, images later. This is the step everyone skips and it is the one that converts
- Days 8 to 10: Reshoot or rebuild imagery for consistency. Photograph anything physical
- Days 11 to 12: Build or rebuild the site. Set the homepage positioning line
- Day 13: Export the tailored PDF master and make one client-specific variant
- Day 14: Send it to three people who will be honest, then to five real prospects
The short version
A strong graphic design portfolio in 2026 is not a gallery, it is an argument. Six to eight projects, each answering a question a nervous buyer is already asking, with enough context to prove judgement and enough craft to prove capability. Cut the rest. The work you remove is what keeps the right client from finding the work that would have hired you.
FAQ
What should a graphic design portfolio contain?
Six to ten projects covering brand identity, print or packaging, multi-page layout, a multi-format campaign, one digital piece, and at least one project with a measurable outcome. Each project needs the problem, your role, one key decision, and the result. A well-framed personal project is a legitimate inclusion.
How many projects should a graphic design portfolio have?
Six to ten for freelancers and studios, four to eight for juniors. Beyond ten, reviewers stop reading and your weakest piece sets your perceived level. Depth beats volume in every review we have observed.
Should I use a website or a PDF?
Both. The website is your inbound asset and your searchable credibility. The tailored PDF is your outbound weapon for pitches and applications, ideally reordered for each prospect. Keep the PDF under 15MB in 16:9 landscape so it opens cleanly on a laptop.
Can I include personal or concept work?
Yes, if you frame it as a brief with an objective and a constraint rather than as a hobby piece. Label it clearly as self-initiated. One or two well-framed passion projects show initiative and point of view. Five unframed ones suggest you have no clients.
What size should portfolio images be?
For web, roughly 2000px on the long edge, compressed as WebP, ideally under 300KB each. For PDF, 1920×1080 pages at 150dpi image resolution keeps quality high and file size sendable by email.
What if I have no client work yet?
Create three self-directed briefs that mirror the work you want to be paid for, write them with real constraints and a defined audience, and add one small pro bono project for a local organisation so you have a real stakeholder and real feedback in your process. That combination reads far better than a grid of unbriefed exercises.
How often should I update it?
Review it every quarter and swap in one new project each time, removing the weakest. A portfolio whose latest work is more than 18 months old signals inactivity, regardless of how good the work is.