First show what was wrong
I would not publish a case study that says “we increased leads by 200%” if the reader cannot see the starting point, the period, and the specific actions. Without that, a case study is a billboard, not proof.
The mistake I would remove first: Showing only a pretty result without a starting point.
What to prepare and what result to expect
- Outcome: The page answers “what you did in a similar situation” and leads to one next step.
- Choose the client, niche, original problem, timeline, and permission to publish. Decide in advance which numbers you can show.
- Keep source data and access rights separate from the output so you can verify what the client case study on the site actually did.
- After 28 days compare views, CTA clicks, requests, and lead quality — not only time on page.
Build the case study from facts, not praise
- Lock the before point
Collect facts: requests, conversion, cost per lead, timeline, team, and limits. For each metric note the period and source: CRM, analytics, or a report.
Проверьте: There is a starting point, not only the client’s impression.
Если не сработало: Use a range or a concrete operational result and state the limit honestly.
- Interview the process
Ask what hurt, why they chose the solution, what they tried before, which limits remained, and what changed in how the team works.
Проверьте: The text has a solution and context, not only a testimonial.
Если не сработало: Ask the client to name one thing they now do differently.
- Calculate without claiming extra credit
Metric growth = (after − before) / before × 100%. Lead conversion = leads / visits × 100%. Separately note if prices, ads, or the team also changed at the same time.
Проверьте: Comparison periods and conditions are visible to the reader.
Если не сработало: Remove the claim “this drove all revenue” and keep the verifiable part.
- Place one CTA
At the end, offer a process review, a calculation, or a discussion of a similar task. Do not add five identical buttons.
Проверьте: A request from the case study keeps URL/UTM and reaches the owner.
Если не сработало: Check the CTA on mobile and after form submit.
Evidence table before publication
Collect `case_id`, `client_type`, `period_before`, `period_after`, `metric_name`, `before_value`, `after_value`, `source`, `intervention`, `cost`, `owner`, `proof_url`, `client_approved`, and `limitations`. Every percentage should trace back to a chart, CRM export, screenshot, or approved document.
If before was zero, do not calculate percent growth: write “was 0, became 12.” Next to the result show limits: seasonality, sales changes, ad campaigns, and what was not part of the project.
Growth = (after - before) / before * 100
Conversion shift = CR_after - CR_before
Incremental profit = sales * gross profitA case study as a path, not a service list
First show the problem, then limits, then actions in order, and the result at the end. Events `case_study_view`, `case_study_proof_expand`, `case_study_cta_click`, and `generate_lead` help you see whether people read the proof and move to the next step.
How not to turn a percentage into manipulation
If before was 12,000 organic sessions, 42 requests, 10 qualified leads, and 2 sales, and after was 15,500, 68, 19, and 5, show all four rows. Lead growth is 61.9%, sales 150%, but next to them note the period, project cost, seasonality, and sales-team changes.
If the baseline is zero, write “was 0, became 12,” not “1,200% growth.” Before publishing, ask the client to confirm the numbers and the right to show the logo, screenshots, and quote.
Which numbers you can compare honestly
| Criterion | Question | Good sign |
|---|---|---|
| Input | What exactly enters the client case study on the site? | Choose the client, niche, original problem, timeline, and permission to publish. Decide in advance which numbers you can show. |
| Action | What is the system allowed to do on its own? | Only pre-listed actions, with no access to the whole account |
| Verification | How do you know the result can be accepted? | After 28 days compare views, CTA clicks, requests, and lead quality — not only time on page. |
| Failure | Where does an unclear case go? |
What should change after setup
The page answers “what you did in a similar situation” and leads to one next step.
Why “we improved everything” does not build trust
Showing only a pretty result without a starting point.
Crediting the project with all growth when other factors also changed.
Disclosing numbers without client permission.
Publishing a case study without explaining what should not be copied blindly.
When a case study needs a separate editor and analyst
Bring in a specialist if the case study involves confidential data, multiple metric sources, or should become part of an SEO cluster.
What to ask the client before publishing
Do you need exact money figures in a case study?
No. You can use a range, percentages, or lead counts if the client allowed publication.
Which CTA should you use?
One next step: request a review, get a calculation, or discuss a similar task.
What if you do not have exact analytics?
Do not invent them. Show an operational fact, the period, and measurement limits.
Do you need to name the client?
Only with permission. An anonymous case study is fine if context and result remain verifiable.
