Most marketing portfolios never get past the first ten seconds. A reader opens the page, scrolls once, and closes the tab, not because the work behind it was bad, but because nothing on the page proved it was good. The gap between a portfolio that gets read and one that gets closed is not talent or experience. It is a method: a fixed way of building a case that turns a list of past projects into evidence someone can evaluate in under a minute. This guide walks through that method step by step, with separate detail for SEO, PPC and multi-channel work, and a section on the one problem almost every marketer runs into: how to show results you are not fully allowed to disclose.
Why Most Portfolios Get Closed Before the First Case Loads
A reader scanning a portfolio is running a three-question filter, whether they realize it or not, and they run it in a fixed order.
- Does this person work on my specific problem? Not "are they good at marketing" but "do they do the narrow thing I need." A portfolio that tries to cover SEO, PPC, social and email equally usually reads as less credible on any single one of them than a narrow SEO portfolio or PPC portfolio would.
- Is there evidence, or just a claim? A skills list is a claim. A client logo, a number, a named case is evidence. Readers scan for evidence first and read claims only if the evidence holds up.
- Does this person think, or just execute? This is the question almost nobody answers on purpose, and it is the one that separates a portfolio that gets a reply from one that does not.
Four things answer these questions before a single case gets opened.
Positioning: what changes when you get specific
Take the same person and write their headline two ways.
- Generic: "Marketing specialist with a background in SEO, PPC and content."
- Generic, SEO-specific version: "SEO specialist focused on improving search rankings."
- Specific: "I move mid-market e-commerce catalogs into the top 5 for their highest-value category pages within two quarters."
The generic versions describe a job function, something already visible from a resume or a LinkedIn title. The specific version names a result, an audience and a rough timeline, and it answers the first filter question on its own: does this person work on my problem. Write your own version of the specific line before you write a single case study. Everything else on the page should support that one sentence.
- A row of proof, not a paragraph of description. Client logos, even small or mid-size ones, read as evidence in about three seconds. A paragraph describing your experience reads as a claim, no matter how well it is written.
- A compact number line. Years active, projects completed, an average rating or repeat-client rate. Three or four figures a reader absorbs in one glance, not a biography they have to read to find them.
- One fully built case above the fold. The first case is what convinces a reader to open the second one. It needs a headline shaped like a story, not a label like "SEO Case Study 1."
A page that opens with a skills list, a stock photo and a line about being "passionate about marketing" gives a reader nothing to evaluate, so they leave before reaching the actual work.
The Six-Part Shape of a Case That Holds Up
Most case studies describe activity: what was audited, what was published, what was launched. The ones that actually convince describe judgment, using the same six-part shape every time: the situation, the goal, your specific role, the constraints you worked inside, the decision you made, and a measurable result.
Role and constraints are the two steps almost everyone skips, and they are exactly where a real case separates from a padded one. Compare "I managed the SEO for a client" with "I was the only strategist on a three-person team, working inside a fixed content budget, and chose to focus two clusters instead of five because the other three had a traffic ceiling too low to justify the spend." The first tells a reader nothing. The second shows exactly how you make decisions under pressure, which is the thing a client or employer is actually trying to evaluate when they read a case study.
Building an SEO Case That Holds Up to Scrutiny
An SEO case earns trust through a level of detail that only someone who actually did the work can produce: a specific crawl finding, a named clustering method, a real indexation number.
At minimum, name:
- What the diagnosis actually showed, not just what you fixed. A log file, a crawl, or Search Console data, read before any work started.
- How you clustered, by intent or by volume, and the reasoning behind that specific choice for that specific site.
- The order you shipped in. Canonical work, content, internal linking, in the sequence it actually happened, not a bundled list.
- Traffic and position together. A position jump without a traffic number can be a vanity metric on its own; both together prove the ranking mattered commercially.
Skip phrases like "conducted a technical audit" or "optimized on-page content." They appear on nearly every SEO resume and, on their own, prove nothing about what you actually found or decided.
Building a PPC Case That Proves It Was Your Work
A PPC case has a built-in advantage over SEO: results are faster to measure and easier to attribute directly to your own decisions.
Include, at minimum:
- The specific bidding change, named precisely (manual bids to target CPA, last-click to data-driven attribution), not "optimized the account."
- Account hygiene work. Negative keyword counts, search term audits, the unglamorous maintenance that actually moves CPA over time.
- How you tested creative. Number of variants, what the test measured, and which version won and why.
- CPA, ROAS and CTR shown together, not just whichever metric looks best on its own. Anyone who has run PPC will notice if only the flattering number appears.
If you managed a budget, state the size and the time period. A vague "delivered significant improvement" reads as a warning sign to anyone who has actually managed paid media.
Building a Case for Work Across Multiple Channels
A multi-channel case is the hardest one to write well, because it is tempting to simply list every channel that was touched. The version that actually works shows how the channels were connected, not which ones existed.
The strongest version shows:
- One shared goal that every channel was measured against, instead of five separate campaigns that happened to run at the same time.
- The attribution model used to decide which channel gets credit, since without one, "which channel worked" has no real answer.
- One specific targeting or sequencing decision, such as retargeting only warm CRM segments instead of a broad cold audience, that shows strategy rather than channel management.
- A business outcome, not a vanity metric. Lead-to-customer conversion or acquisition cost says more to a hiring manager than impressions or reach ever will.
Showing Results Without Breaking Confidentiality
Almost every client engagement carries some expectation of confidentiality, and portfolios tend to handle it in one of two bad ways: breaking it outright, or hiding behind vague language that proves nothing.
A blurred logo next to an exact dollar figure protects very little, because the figure itself, combined with the industry and timeframe, is often enough to identify the client on its own. The safer pattern keeps the client anonymous by industry and size, and states the result as a relative change or a range: "organic revenue up 2.4x in eight months" instead of the underlying dollar figures. If a prospective client asks for the exact numbers later, in conversation, that is a case-by-case decision, ideally made with the former client's agreement, not a default for a public page.
If You Do Not Have Paid Client Results Yet
An empty portfolio is worse than a practice one, and a fabricated one is worse than either. Three honest options work well:
- A self-directed project, on your own site or for a nonprofit you volunteer with, built using the same six-part case shape with real, checkable numbers.
- A public teardown. Choose a real site with a visible SEO or PPC problem, explain what you would change and why, and go deep enough that a reader can tell you actually reviewed the account or the code, not just the homepage.
- A documented practice project, clearly labeled as practice, showing the same decision-under-constraint thinking as paid work would.
Label these honestly. An employer respects a clearly marked practice project far more than a small real case inflated to look bigger than it was.
Where a Case Study Should Live
A personal site, even one page, outperforms a PDF or a LinkedIn-only presence for a simple reason: it is the only format fully under your control, and the only one search engines can index, which means it can bring in interest on its own over time. A PDF works well as something you send after first contact, and a LinkedIn featured section works as a secondary pointer, but neither should be the only place your case studies live. If a full site is not realistic yet, one well-built page is enough to start.
What Makes a Case More Believable Beyond the Case Itself
Even a well-built case carries less weight on its own, without three supporting signals a reader takes in at the same time.
- A testimonial with a real name attached. "Great to work with" proves nothing, because there is no way to check who said it. A quote with a name, role and company acts as independent confirmation, and one specific sentence about a result outweighs a paragraph of generic praise.
- Any outside mention. An article, an interview, a conference talk, even one, signals third-party validation faster than another paragraph describing your own skills.
- Services described as the client's own question, not your process. Not "I offer SEO audits" but "I want to know exactly where I am losing traffic and what to fix first." Mirroring the client's actual internal question reads faster than a description of your workflow.
None of these three replace a real case with a result, but without them, a portfolio reads as self-description instead of something independently confirmed.
Common Mistakes That Undo Otherwise Solid Work
- Listing tools and skills instead of outcomes, which reads as a checklist rather than proof of judgment.
- Writing in general theory instead of naming a specific situation, decision and number.
- Publishing five thin cases instead of two or three fully built ones.
- Taking credit for a team result without stating your own specific role in it.
- Letting cases sit unrefreshed for a year or more, which quietly signals the work has gone stale.
- Ignoring layout entirely. A strong case loses credibility fast inside a cluttered, hard-to-scan page.
FAQ
What goes in a marketing portfolio with no paid client work yet? A self-directed project or a documented public teardown, clearly labeled as practice, built with the same situation-goal-role-result shape used for a real case.
How many cases does a strong portfolio need? Two to four fully built ones beat ten thin ones. Depth is what proves judgment; volume does not.
Is it safe to publish exact client numbers? Usually not without the client's sign-off. Use a relative change or a range, and keep the client anonymous by industry and size instead of blurring a logo next to a precise figure.
Should an SEO case look different from a PPC or multi-channel case? The six-part shape stays the same. What changes is the evidence: traffic and position data for SEO, CPA, ROAS and CTR for PPC, and an attribution-based result for multi-channel work.
Summary
A portfolio that gets read is built on evidence, not volume: a specific positioning line, visible proof of past clients, and two or three cases built on situation, goal, role, constraints, decision and a measurable result. Whether the work is SEO, PPC or spread across channels, the details a template cannot supply are exactly what make a case believable. If you want help turning results like these into steady organic visibility instead of a one-time portfolio page, see our SEO marketing and content optimization services, or read how we evaluate AEO tools and get pages cited in AI Overviews.
