How to build a design portfolio without clients or a degree

Hiring managers skim for thinking, not for logos of famous clients. Here is how a self-taught designer builds three to five case studies from scratch, without fake rebrands, and gets them seen.

A designer's hand sketching logo ideas with a pen on paper at a wooden desk, with a monitor and laptop showing colour swatches behind
Photo: Claireneon, CC BY-SA 4.0, via Wikimedia Commons

A design portfolio without clients works when it shows how you think, not who you worked for. Build three to five self-initiated case studies with a real problem, real constraints and a stated audience; show the alternatives you tried and why you chose one; label concept work as concept work; publish it on your own site; and get structured critique before you send it anywhere. Employers that hire on skills look at this evidence, and the data says skills-based hiring is growing slowly but pays off for those it reaches.

Key takeaways
  • 3 to 5 finished, self-initiated case studies beat twenty half-done pieces.
  • Constraints make concept work look real: sizes, print method, audience, budget.
  • Parallel prototyping won on every measure in a Stanford test of 33 designers.
  • 7%: how much students over-rated their own work versus staff, so get rubric critique.
  • 1 in 700 US hires came from dropped degree rules; the portfolio still has to do the work.

The two most common questions from self-taught designers on Graphic Design Stack Exchange are the ones in this post's title: what to put in a portfolio with no clients, and whether designs for imaginary companies count. The highest-voted answers agree on more than you would expect:

  • Fake clients are acceptable if the work is honest and strong, because, as one hiring designer put it, a demo project has no excuse for client compromises.
  • Real projects for yourself, your hobbies or a non-profit are better still.

This guide turns that advice into a method, with the research behind each step.

I write this as someone who learned design without a degree. I have done graphic and brand design since about 2013 and freelanced on more than 80 projects; my formal training is a BE in mechanical engineering, plus a car-design workshop through Launchpad. So the method below is the one I would hand a younger version of myself, checked against what hiring research actually says.

204UX hiring managers surveyed by Nielsen Norman Group on portfolios
3.5 ptsrise in non-degree hires where firms dropped degree requirements (2024)
10 ptshigher retention of non-degree hires at skills-based "leader" firms
65.5%of peer grades within 10% of expert grades in a Stanford design class

Can you get a design job without a degree or a client list?

The story

In July 2013, two New York designers launched a website nobody had commissioned. Jessica Walsh, a partner at the studio Sagmeister & Walsh, and Timothy Goodman, who ran his own studio, were single at the same time with opposite habits: he feared commitment, she fell for people too fast. They agreed to date each other for 40 days, fill in a daily questionnaire, see a couples therapist weekly, and publish it all, designed and hand-lettered, at fortydaysofdating.com.

By September 2013, Warner Bros. had bought the film rights. The site later passed ten million unique visitors, and Abrams published a 304-page book in January 2015. Walsh said afterwards that she spent 25 to 50% of her time on her own ideas. Both were already working designers, so this is no zero-to-hired tale, but it shows how far an unpaid, self-set brief can travel.

The Great Discontent: interview with Timothy Goodman and Jessica Walsh

Yes, but do not assume a dropped degree requirement means a changed hiring process. A February 2024 study by the Burning Glass Institute and Harvard Business School tracked 11,300 roles at large US firms that removed degree requirements and found the share of non-graduates hired rose by only about 3.5 percentage points on average.

11,300roles tracked that dropped degree requirements
3.5 ptsrise in non-graduate hires on average
3.6%of roles actually dropped the requirement
37%of firms were real skills-based leaders

Because only 3.6% of roles dropped the requirement in the period, the net effect was 0.14 percentage points across all hiring, about 97,000 extra jobs a year out of 77 million US hires, or fewer than 1 in 700 (Manufacturing Dive, reporting the study).

Where the change was real

Nearly all the change came from 37% of firms, which the authors call skills-based hiring leaders. At those firms:

  • non-degree hires stayed longer, with retention 10 percentage points higher than graduates;
  • they earned about 25% more than in their previous roles.

The practical meaning for a self-taught designer: your portfolio has to do the work a degree does for others, because most employers still lean on credentials by habit.

What firms did after dropping degree requirements
See the data as a table
Share of firms (%)
Skills-based hiring leaders (real change)37
In name only (no change in hiring)45
Backsliders (gains that faded)18

Source: Burning Glass Institute and Harvard Business School, Skills-Based Hiring: The Long Road from Pronouncements to Practice (Feb 2024) · Backsliders are given as approximately one-fifth in the report; 18% is the remainder. US data, all occupations, not design alone.

India's job market is not covered by that study, and I found no equivalent Indian dataset for design roles, so treat it as a direction rather than a local number. The direction matches what Indian studios and in-house teams tend to do in practice: shortlist from the portfolio link first, then check the CV. For the longer path from zero to paid work, see how to become a self-taught graphic designer in India.

What do hiring managers look for in a design portfolio?

They look for evidence of process and judgement: the problem, your role, the constraints, the options you explored, and what changed because of your work. Nielsen Norman Group surveyed 204 UX professionals who hire and found they valued how a candidate went from an opportunity to a result far more than a gallery of finished screens.

UX Design Portfolios: Best PracticesRachel Krause · Nielsen Norman Group · 2019

A survey of 204 UX professionals in charge of hiring, run as part of NN/g's research on UX careers. Respondents repeatedly asked to see the problem, the candidate's specific role, constraints and timelines, how research shaped decisions, how designs changed between iterations, and what was deliberately left out. The common mistake they reported was portfolios made of final screens only. Limit: an industry survey of UX hiring, not graphic design; the article reports themes, not percentages.

Read the report (free)

Two consequences follow:

  1. The first screen decides. A reviewer reads a portfolio the way you read a newspaper: headline, picture, maybe the first paragraph. Each case study must make its point in its first screen.
  2. Constraints stand in for the client. When there is no client, the constraints are the only thing that makes the work look real.

A logo for "a café" proves little; a logo for a 12-seat filter-coffee counter near a Bengaluru metro station that must work on a paper cup, a steel tumbler and a 1 cm Swiggy icon proves a lot.

1. HookFinal work +one-line result2. BriefProblem, user,constraints3. Options3+ directions,side by side4. DecisionWhy this one,what you cut5. ProofMock-ups in use,test or feedbackFirst screen:must stand aloneMost reviewers skim; few read every word
The five blocks of a case study. Put the strongest image and a one-line result in block 1, because many reviewers never scroll further.

Step by step: building the portfolio from zero

Plan on three to five projects, each finished to a professional standard, rather than twenty half-done ones. The steps below take a self-taught designer from an empty folder to a portfolio that can be sent.

  1. Choose the job you want, then the projects. A brand designer, a UI designer and a packaging designer need different evidence. Pick one target role and make at least three projects point at it.
  2. Write a real brief for each project. One page: who it is for, what problem it solves, the constraints (budget, sizes, print method, languages, deadline) and how you will judge success. Self-initiated briefs are allowed; vague ones are not.
  3. Pick honest sources of work. Your own projects, a local business or non-profit that agrees to it, a public-interest redesign clearly labelled as unsolicited, or a set brief from a competition. Avoid passing off a redesign of a real company as client work.
  4. Explore in parallel, not in sequence. Make three or more distinct directions before choosing, and keep them. The research below shows this produces better work, and it gives your case study its "options" block.
  5. Test the work in use. Put the logo on a cup, the poster on a wall photo, the app screen in front of five users. Record what changed.
  6. Write the case study. Use the five blocks above. Keep text short and specific; numbers beat adjectives.
  7. Get structured critique. Use a rubric so feedback is comparable, revise, and only then publish.
  8. Publish on your own domain, then mirror the best two projects on Behance or Dribbble for discovery, and send a direct link to the most relevant case study with each application.
A designer seen from behind sketching rows of geometric letter M logo variations on paper beside a laptop
Rows of alternative marks for one brief. Keeping these sketches is cheap, and they become the options block of a case study.Photo: Claireneon, CC BY-SA 4.0, via Wikimedia Commons

Is it okay to design for fake or real companies you never worked with?

Yes, if you are transparent. Invented companies and unsolicited redesigns are a normal way to learn, and hiring designers on Stack Exchange accept them; what damages trust is presenting concept work as if a real client commissioned it, or using a real brand's trademarks as though you had their approval.

Source of workCredibility with reviewersRiskHow to label it
Your own product, event or channelHigh: real users and real constraintsLow"Self-initiated, in use since…"
Non-profit or local shop, free or paidHigh: a real decision-maker said yesScope creep; get a short written briefClient name, with permission
Competition or set briefMedium: comparable with othersCopied briefs look generic"Brief from…"
Invented companyMedium if constraints are specificLooks like a student exercise if vague"Concept project"
Unsolicited redesign of a real brandMedium to high if the reasoning is sharpTrademark use; implying endorsement"Unsolicited concept, not affiliated with…"
Recreated tutorialsLowReviewers recognise themLeave out

Borrow constraints from real objects

The best self-initiated work borrows its constraints from the physical world. One rule I keep coming back to from my own work on motorcycle graphics is that a graphic has to respect the object: on a fuel tank or side panel, the shape of the panel comes first and the design follows its contours.

A case study that shows that kind of constraint, with photos of the real object, reads as professional even without a client. The same logic runs through the logo design process and how to write brand guidelines, which make good portfolio pieces in their own right.

Why should a portfolio show several directions, not just the final design?

33novice designers in the experiment
1,180,320ad impressions over 15 days
501clicks in total

Because exploring several options in parallel produces measurably better work, and showing them proves you can judge. A Stanford experiment found that novice designers who made several prototypes before getting feedback beat those who iterated one at a time on every measure.

Parallel Prototyping Leads to Better Design Results, More Divergence, and Increased Self-EfficacyDow, Glassco, Kass, Schwarz, Schwartz and Klemmer · ACM Transactions on Computer-Human Interaction · 2010

33 participants (average age 22, none professional designers) each designed a banner ad for a real client; half made prototypes in parallel and got feedback on several at once, half worked serially with critique after each one.

The ads ran on MySpace for 15 days: 1,180,320 impressions and 501 clicks in total. Parallel ads won on click-through, on time visitors spent on the client's site, and on ratings by the client and ad professionals; the two best ads were parallel (735 and 578 clicks per million impressions) and the two worst, with zero clicks, were serial. Parallel designers' work was also more diverse, and their self-rated design confidence rose. Limit: small sample, one task type.

Read the paper (free)
SerialParalleldot = feedbackfeedback on several at oncefinalfinal
The two conditions in Dow et al. (2010). Parallel designers compared feedback across several prototypes before committing, which led to better and more varied ads.

For a portfolio, this means two things: do the parallel exploration for real, and show it.

A row of three directions, with one sentence each on why two were dropped, is the strongest evidence of judgement you can give a reviewer who has never met you.

How do you get useful critique without a teacher?

65.5%of peer grades within 10% of staff grade
7%average self-inflation of own work
12.4% to 9.9%grading error with clearer rubric wording

Use a short rubric and several reviewers. Research on peer assessment in a large online design class at Stanford found that peers using a clear rubric graded close to expert staff, and that clearer rubric wording cut grading error further.

Peer and Self Assessment in Massive Online ClassesKulkarni, Wei, Le, Chia, Papadopoulos, Cheng, Koller and Klemmer · ACM Transactions on Computer-Human Interaction · 2013

Two runs of an online HCI design course, the first large class to use peer and self assessment. In the second run, 42.9% of peer grades fell within 5% of the staff grade and 65.5% within 10%. Students rated their own work about 7% higher than staff did. Rewriting rubric items with parallel structure and unambiguous wording cut median grading error from 12.4% to 9.9%, and telling students about their own grading bias improved later accuracy. Limit: a university course with set assignments, not portfolio review for hiring.

Read the paper (free)

The 7% self-inflation is the useful warning: you will rate your own case study higher than a stranger would. Build a five-line rubric and score each project 1 to 5:

  • clarity of the problem;
  • strength of the constraints;
  • range of options explored;
  • quality of execution;
  • evidence of outcome.

Ask three designers you respect to score it too. Revise anything where their average sits a point below yours.

Use this to plan your portfolio's timeline

Most portfolios stall because the time cost is never estimated. Enter how many case studies you will build and the hours each stage takes you; the planner gives total hours and the weeks it takes at the hours you can spare.

Portfolio build planner

With the defaults, four case studies take about 154 hours, or roughly 19 weeks at eight hours a week. That is why three strong projects sent in two months beat six promised for next year.

Website, Behance or PDF: where should the portfolio live?

Put the master copy on your own domain, mirror selected projects on a platform for discovery, and keep a short PDF for applications that require attachments. Each format does a different job.

FormatBest forWeakness
Own websiteControl of order, story and URL; shows you can shipNeeds upkeep; no built-in audience
Behance or DribbbleDiscovery and search inside the design communitySame template as everyone; favours visuals over reasoning
PDF, 8 to 12 pagesJob portals and email attachmentsFiles go stale and get forwarded without context
Printed bookIn-person reviews for print, packaging and identity rolesCostly to update
A measured ink drawing of an ornamental iron gate between two stone pillars, with handwritten notes and dimensions, from an eighteenth-century architect's sketchbook
A page from the sketchbooks of the Maastricht architect Mathias Soiron, with dimensions and notes beside the design. Showing the working is an old tradition, not a new portfolio trend.Photo: Kleon3, Public domain, via Wikimedia Commons

How is AI changing what a portfolio needs to show?

Because image generators make polished visuals cheap, reviewers increasingly look past the final image to the decisions behind it. Autodesk's 2026 AI Jobs Report, built on about 4 million job listings, found AI UX Designer was the fastest-growing AI title in design and make industries, up 145% year on year.

The same report says 82% of students feel confident with everyday AI tools but only 36% feel ready to use the AI specific to their field. For a portfolio, that gap is an opportunity: a case study that shows where you used AI (research, variants, mock-ups), where you did not (type, layout, final judgement), and how you checked the rights to every asset says more than any single render.

Figma's March 2026 analysis of its State of the Designer survey adds context: 82% of hiring managers said their company's need for designers had grown or held steady, yet only 20% thought the industry was improving.

The full method is in how to use AI in a design workflow, and the wider argument in what changes when you design with AI. For how I apply "human-directed, AI-built" to my own marketing work, see Applied AI and Work.

Questions people ask

What should I put in my portfolio if I have no clients yet?

Three to five self-initiated case studies aimed at one type of job: your own projects, work for a non-profit or local business, set briefs, or clearly labelled concepts. Each should show the problem, constraints, options and the reasoning behind the final choice.

Should I create designs for non-existing companies?

Yes, as practice and as portfolio pieces, provided you label them as concepts and give them specific constraints. Reviewers accept invented clients; they reject vague briefs and work presented as commissioned when it was not.

How many projects should a design portfolio have?

Quality beats count. Three to five finished case studies is a common recommendation, and Nielsen Norman Group's survey of hiring managers stresses depth of process over volume. Cut any project you would not want to be asked about in an interview.

Do I need a design degree to get hired?

Not always, but most employers still use degrees as a filter. In the 2024 Burning Glass and Harvard study, only about 1 in 700 US hires was a non-graduate who got in because a requirement was dropped. A strong portfolio and a direct link to the most relevant case study are how you get past that filter.

Sources

  1. HuffPost: Warner Bros. acquires 40 Days of Dating film rights (Sep 2013)
  2. The Great Discontent: Timothy Goodman and Jessica Walsh interview (40 Days of Dating: launch, ten million visitors, book)
  3. Burning Glass Institute and Harvard Business School: Skills-Based Hiring, The Long Road from Pronouncements to Practice (Feb 2024; 11,300 roles, 3.5-point rise, 37% leaders, retention and pay figures)
  4. Nielsen Norman Group: UX Design Portfolios, survey of 204 hiring professionals (2019)
  5. Dow et al.: Parallel Prototyping Leads to Better Design Results, More Divergence, and Increased Self-Efficacy, ACM TOCHI (2010)
  6. Kulkarni et al.: Peer and Self Assessment in Massive Online Classes, ACM TOCHI (2013)
  7. Autodesk: AI Jobs Report 2026 (2026)
  8. Figma: Design's influence is expanding, and here's why that feels hard (Mar 2026)
  9. Graphic Design Stack Exchange: What to put in my portfolio if I have not had any clients yet?
  10. Graphic Design Stack Exchange: Should I create designs for non-existing companies for my portfolio?
  11. Graphic Design Stack Exchange: Tips and resources for beginning designers
design portfoliocase studyself-taught designerskills-based hiringparallel prototypingpeer critiqueBehanceconcept work
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Artblog.

Keep reading in Artblog

From the other blogs