Most agency case studies are sales collateral that never gets read. A well built one is something else entirely: a page that ranks for commercial keywords, answers the exact objection a prospect has at the bottom of the funnel, and closes deals while you sleep.
This guide gives you a reusable SEO case study template you can apply to any client project, plus the part nearly every other template skips: how to structure the page so it also ranks for your service keywords. We cover the problem-process-results format, the metrics worth publishing, how to present before and after screenshots credibly, how to get written client approval to publish numbers, and how to wire title tags, URLs and internal links together. tripledart.com makes the same point with more data.
Why your case studies should double as ranking pages
Agencies typically bury case studies in a gated PDF or a portfolio grid with 40 words of text. That is a waste of an asset you already paid for.
A published, indexable case study page does four jobs at once:
- Bottom-funnel search capture. Queries like “saas seo agency results”, “local seo case study dentist” or “ecommerce seo agency case study” are typed by people with budget.
- Topical proof for your service pages. Internal links from a case study to a service page pass relevance signals and context that generic blog posts cannot.
- Sales enablement. Your team sends one URL instead of a 12MB attachment.
- Trust signals. Named clients, real screenshots and verifiable dates do more for conversion rate than any “award winning” badge.
Treat each case study as a landing page with a job, not as a blog post.
The SEO case study template at a glance
Here is the skeleton. Every section has a purpose and a rough word count so the page stays readable.
| Section | What goes in it | Target length |
|---|---|---|
| Headline + result snapshot | Client type, headline metric, timeframe | 40 to 60 words |
| Client profile | Industry, size, market, starting point | 100 to 150 words |
| The problem | Business pain, not just SEO symptoms | 200 to 300 words |
| Audit findings | Diagnostic data, baseline numbers | 200 to 350 words |
| The process | Phased actions with dates | 500 to 900 words |
| The results | Before/after table, screenshots, revenue | 300 to 450 words |
| Client quote | Named person, role, photo if possible | 40 to 80 words |
| What we would do differently | Honest reflection, builds credibility | 100 to 150 words |
| Call to action | Link to the matching service page | 50 words |
Total: roughly 1,500 to 2,500 words. Long enough to be substantive, short enough that a prospect scrolls to the CTA.
Step 1: Write the problem section like a business problem
The most common mistake is opening with “the client had poor technical SEO”. Nobody buys a solution to a crawl budget issue. They buy a solution to “we are spending 14,000 a month on Google Ads and our pipeline dies the moment we pause”.
Structure the problem in three layers
- The business pain. Revenue plateau, over-reliance on paid, seasonal collapse, expansion into a new market, a site migration that went wrong.
- The search symptom. Flat organic sessions, no rankings for money terms, cannibalisation, thin category pages, a manual action.
- The constraint. Budget, dev resource, legal review, a locked-down CMS. Constraints make the story believable and let prospects recognise themselves.
Example opening that works
“A 38-person B2B logistics software company was generating 91% of demo requests from paid search. Organic traffic had been flat at around 4,200 monthly sessions for 18 months. With a funding round closing and a mandate to cut customer acquisition cost by 30%, they needed organic to carry a third of pipeline within a year. They had no in-house developer and a Webflow build they could not fully restructure.”
Four sentences, and a prospect in logistics software already feels understood.
Step 2: Document the process so it is credible but not copyable
Clients want evidence of a repeatable method. Competitors want your playbook. The balance: be specific about sequence and reasoning, general about proprietary assets.
Break the work into phases with dates
Phased storytelling is what separates a case study from a brag list. Use a simple format:
| Phase | Timeline | Key actions | Early signal |
|---|---|---|---|
| 1. Diagnosis | Weeks 1 to 3 | Crawl, log file review, keyword gap vs 4 competitors | Baseline set |
| 2. Technical fixes | Weeks 4 to 8 | Indexation cleanup, internal link rebuild, template speed work | Indexed pages up 22% |
| 3. Content build | Months 3 to 7 | 18 bottom-funnel pages, 2 pillar guides, entity coverage | First page 3 rankings |
| 4. Authority | Months 5 to 11 | Digital PR, 31 referring domains, partner placements | Money terms into top 5 |
Always include at least one thing that did not work
“We tested a programmatic location template and pulled it after six weeks because impressions grew but qualified sessions did not.” That single sentence makes every other claim on the page more believable. It also signals real operator experience to an AI summary engine or a human skimming for substance. A comparable breakdown sits on huble.com.
Step 3: Choose the right metrics (and skip the vanity ones)
A results section stuffed with impressions looks like padding. Pick three to six metrics, present them as before and after with a clearly stated period.
Metrics ranked by persuasive power
| Tier | Metric | Why it matters |
|---|---|---|
| Tier 1 | Revenue from organic, closed deals, cost per acquisition | Speaks the language of the person signing the contract |
| Tier 1 | Qualified leads, demo requests, calls, quote forms | Closest proxy to revenue when revenue is confidential |
| Tier 2 | Non-brand organic sessions, conversion rate by landing page | Proves you grew demand rather than harvesting existing brand search |
| Tier 2 | Keywords in positions 1 to 3 for commercial intent terms | Specific, checkable, hard to fake |
| Tier 3 | Impressions, total keyword count, domain authority scores | Use as supporting context only, never as the headline |
Rules for presenting numbers honestly
- Always state the comparison window. “Jan to Jun 2026 vs Jan to Jun 2025” beats “after our campaign”.
- Separate brand from non-brand. If brand search grew because the client raised a round, say so.
- Give absolute numbers alongside percentages. A 700% increase from 40 sessions is noise. From 40,000 it is a business.
- Name the attribution model. Last non-direct click in GA4, or CRM-sourced, or call tracking. Prospects with in-house analysts will ask.
- Mention external factors. A core update, a seasonal spike, a competitor exiting the market. Admitting tailwinds costs you nothing and buys trust.
The before and after table
This is the single most-screenshotted element of any case study. Keep it clean:
| Metric | Baseline (12 months prior) | After 12 months | Change |
|---|---|---|---|
| Non-brand organic sessions / month | 3,140 | 18,970 | +504% |
| Demo requests from organic / month | 11 | 74 | +573% |
| Commercial keywords in top 3 | 2 | 41 | +39 |
| Blended cost per demo | 218 | 96 | -56% |
Step 4: Present before and after screenshots that survive scrutiny
Screenshots are evidence. Treated carelessly, they become a liability. Experienced buyers zoom in and check the axis.
What to capture
- Google Search Console performance chart with a 16-month date range visible and both clicks and impressions toggled on.
- GA4 landing page report filtered to organic, comparing two equal periods.
- Rank tracker history for three to five named money keywords, not a vague “visibility index”.
- CRM or call tracking dashboard if the client permits it. This is the most convincing screenshot you can possibly publish.
- Visual before and after of a redesigned template, side by side at the same width.
Presentation rules
- Never crop the y-axis to exaggerate. Leave the zero baseline visible.
- Keep the date range in the frame. A chart with no dates is worthless.
- Blur only what you must. Redact the property name if required, but do not blur the numbers you are claiming.
- Annotate deployment dates. A thin vertical line labelled “new category templates live, 14 Mar” ties cause to effect.
- Export at 2x resolution and serve as WebP with descriptive alt text, for example alt=”Search Console clicks growing from 3,100 to 19,000 over 12 months”.
- Use lazy loading below the fold so the page does not fail Core Web Vitals under the weight of eight dashboard images.
If you use a before and after slider, make sure it still renders as two static images for crawlers and for anyone with JavaScript blocked.
Step 5: Get written client approval to publish the numbers
This is where most case studies die. The project finished nine months ago, the contact has moved on, and nobody ever asked for permission. Build approval into the engagement instead of chasing it later.
Ask at the right moment
- In the contract. Include a logo and results usage clause with an opt-out. Make publishing the default, not the exception.
- At the first big win. The week they hit a milestone, not at the end of the retainer. Enthusiasm has a short half-life.
- At the quarterly review. Present the slide, then say: “We would like to turn this into a public case study. Can we get sign-off?”
Offer a tiered disclosure menu
Give the client options instead of a yes/no question. Approval rates climb dramatically.
| Level | What is published | Typical use |
|---|---|---|
| Full | Brand name, logo, revenue figures, named quote | Happy clients, non-competitive niche |
| Partial | Brand name, percentages only, no absolute revenue | Most common outcome |
| Anonymised | “A UK-based industrial equipment distributor, 60 staff” | Competitive sectors, listed companies |
| Indexed only | Method and screenshots, figures as index (100 to 604) | Finance, legal, healthcare clients |
A permission request you can copy
Hi [Name],
The Q3 numbers are strong enough that we would like to write them up as a public case study on our site. Two things in it for you: a dofollow link back to [domain] from a page we will actively promote, and a credible third-party account of the work your team shipped internally.We would publish: [list the exact metrics and screenshots]. We would not publish: [pricing, margins, customer names, anything under NDA].
You get full approval over the draft and the quote before anything goes live, and we will remove it on request at any point in the future. Happy to send the draft first if that is easier. Can I go ahead?
Follow verbal approval with a short written confirmation email and store it. Then send the final draft as a staging link, not a Word document, so they approve exactly what visitors will see.
If they say no to everything
Publish the methodology anyway, anonymised, with the metrics expressed as index values and screenshots cropped to remove identifiers. An anonymous case study with a real method still outperforms no case study, and it still ranks. The team at digitalagencynetwork.com reached a similar conclusion.
Step 6: Build the page so it ranks for service keywords
Here is the part the generic case study templates leave out. Your case study is a commercial page. Optimise it like one.
URL structure
Keep case studies in a single crawlable directory and include the vertical or service in the slug:
- Good:
/case-studies/b2b-saas-seo-logistics-software/ - Good:
/case-studies/local-seo-dental-clinic-manchester/ - Avoid:
/portfolio/project-17/or/blog/2026/10/our-latest-win/
Do not put the year in the URL. You want to refresh the page with new data in 18 months without a redirect. Put the date in the body copy and in the dateModified field instead.
Title tag formulas that earn clicks
Combine the vertical, the service, the outcome and the timeframe. Keep it under about 60 characters where you can.
- SaaS SEO Case Study: 74 Demos a Month From Organic
- Dental SEO Case Study: 3x Calls in 9 Months (Manchester)
- Ecommerce SEO Case Study: +504% Non-Brand Traffic
The H1 can be longer and more human than the title tag. The meta description should state the headline metric and the timeframe, because that is the snippet that wins the click.
Internal linking: the bit that actually moves rankings
A case study hub with no internal links is a dead end. Wire it up in both directions:
- Service page to case study. Each service page should link to two or three relevant case studies with descriptive anchors like “see how we grew non-brand traffic for a logistics SaaS”.
- Case study to service page. Link once in the process section and once in the CTA, using the service keyword as anchor text, for example “technical SEO audit”.
- Case study to case study. A “more results in this industry” block at the foot of each page.
- Hub page. A
/case-studies/index filterable by industry and service, linked from the main navigation, not the footer alone. - Blog to case study. Whenever a blog post makes a claim, link to the case study that proves it. That is exactly what this paragraph would do on a live agency site.
Schema markup
Mark the page up as Article with an about property naming the service, and add FAQPage schema if you include a question block. If the client provided a quote, use Review or Quotation markup where appropriate. Do not fake AggregateRating on a case study page.
On-page essentials checklist
- Headline metric visible above the fold, no scrolling required
- H2s written as questions or outcomes so they can be lifted into AI summaries and featured snippets
- A summary box of three to four bullets at the top for skimmers and for extraction
- Publish date and last updated date shown on the page
- Author byline linking to a real consultant bio with credentials
- Client logo with alt text, hosted locally
- Two CTAs: one mid-page, one at the end
- Page weight under 1.5MB after image compression
The fill-in-the-blanks SEO case study template
Copy this into your CMS and replace the brackets.
- H1: [Industry] SEO Case Study: [Headline Result] in [Timeframe]
- Snapshot box: Client type, service delivered, engagement length, three headline metrics
- H2: Who the client is. Sector, size, markets, what they sell, who they sell to
- H2: The problem they came to us with. Business pain, search symptoms, constraints
- H2: What the audit uncovered. Three to five findings with baseline data
- H2: The strategy. One paragraph explaining the central bet and why
- H3: Phase 1 to 4. Actions, dates, one screenshot or data point per phase
- H2: What did not work. One honest paragraph
- H2: The results. Before and after table, two or three screenshots, period stated
- H2: What the client said. Named quote, role, photo
- H2: What we would do differently next time. Short reflection
- H2: Could this work for your [industry] site? CTA linking to the service page
- FAQ block: three to five questions prospects in that vertical always ask
Common mistakes that kill a case study page
| Mistake | Fix |
|---|---|
| Gating it behind a form | Publish it openly, gate a deeper appendix or template if you must capture emails |
| Publishing as a PDF only | HTML page first, PDF as an optional download |
| Percentages with no base numbers | Always show the starting figure |
| No dates anywhere | Date every chart, phase and claim |
| Identical template text across ten case studies | Vary the narrative; near-duplicate pages compete with each other |
| Never updated | Add a 24-month follow-up section showing whether results held |
How many case studies do you need?
Aim for one per service and vertical combination you actively sell. An agency selling technical SEO, local SEO and digital PR across three verticals has nine target pages. That is a realistic 12-month publishing plan at roughly one per six weeks, and it builds a cluster that supports every service page you own.
Quality beats volume. Three deeply documented case studies with real screenshots and named clients will outperform twenty paragraph-length testimonials every time.
FAQ
What is an SEO case study template?
It is a reusable page structure for documenting a client project: who the client was, what the problem was, what you did in phases, what the measured results were, and proof in the form of screenshots and a client quote. The strongest versions are also optimised as landing pages so they rank for commercial search terms.
How long should an SEO case study be?
Between 1,500 and 2,500 words for a flagship case study. Shorter than 800 words and there is not enough substance to be persuasive or to rank. Longer than 3,000 and prospects stop scrolling before the call to action.
Can I publish client results without permission?
No. Publishing identifiable client data without consent risks breaching your contract and your NDA, and it damages the relationship. Either secure written approval, or anonymise the client and express figures as percentages or index values. Check your master services agreement for an existing publicity clause before you ask.
Should case studies live on the blog or on their own section?
Their own section, at /case-studies/. They are commercial pages with a different intent from editorial content, they should sit in the main navigation, and keeping them separate lets you link cleanly from service pages without the URL implying the content is dated.
What if the results were modest?
Publish them with context. A 35% lift in qualified leads for a business in a brutally competitive market, achieved on a small budget, can be more persuasive than a 700% traffic increase from a near-zero baseline. Explain the difficulty, and prospects in similar situations will self-identify.
Do case studies help with AI search and LLM citations?
Yes. Pages with original data, clear question-shaped headings, dated claims and named sources are the kind of content that gets extracted and cited. A case study with a proprietary before and after dataset is original content by definition, which is exactly what generic blog posts are not.
How often should I update a published case study?
Review every six months. Add a short follow-up block if results held, grew or declined, refresh the screenshots with a longer date range, and update the last-modified date. A case study showing sustained results over 24 or 36 months is far more convincing than a 12-month snapshot.
Next step
Pick your single best client result from the past 18 months, request approval this week while the numbers are fresh, and build the page using the structure above. One properly optimised case study, internally linked from the matching service page, will do more for your pipeline than another six months of general blog posts.