Case studies are the most persuasive material an agency produces and the easiest to make misleading. A chart that goes up and to the right proves little by itself. This guide shows how to read a case study as evidence rather than advertising.
What a useful case study contains
Look for these elements. Missing ones are not automatically disqualifying, but they tell you what to ask.
- Client context. Industry, business model, size and market, even if anonymized.
- Starting position. What was wrong or missing at the start, and how that was diagnosed.
- Actions taken. Specific work, not just a category such as "technical optimization."
- Time frame. Start date, end date and when each action shipped.
- Metrics and sources. Where numbers came from and how they were defined.
- Business outcome. Leads, orders or revenue, not only traffic.
- Limits. Other factors that may have influenced the result.
The key questions
What was the baseline?
An increase from a very small starting number looks dramatic and may be modest in absolute terms. A percentage without the underlying numbers is hard to interpret. Ask for the before and after in absolute values.
Which metric, and is it the right one?
Traffic can rise while enquiries fall. Rankings for obscure terms matter little. Check whether the metric connects to the business goal and whether branded and non-branded traffic are separated. Branded growth often reflects offline marketing, not SEO.
What else changed?
Seasonality, a product launch, a campaign, a price change, a competitor leaving the market, a site redesign or an algorithm update can all move organic numbers. A credible case study acknowledges such factors or controls for them, for example by comparing the same period year over year.
How do we know the actions caused the result?
Attribution in SEO is imperfect. Be cautious when a study claims a single tactic produced a result in a complex program. Stronger evidence includes staged rollouts, comparisons against sections of the site that were not changed, or clear timing between a change and a response.
How long was the measurement window?
A short spike may fade. Ask whether results were sustained and what the trend looks like beyond the highlighted period.
Is the client real and reachable?
Anonymous case studies are common because of confidentiality. Ask whether the agency can arrange a reference conversation or share more under NDA.
Was the starting problem comparable to mine?
A turnaround for a penalized site, a launch for a new brand and incremental growth for an established one are different jobs. Check that the example resembles your situation.
Warning signs
- Only percentages and no absolute figures.
- Screenshots with axes cropped or labels removed.
- Cherry-picked keywords or date ranges.
- Outcome language that implies guarantees.
- No description of actions beyond buzzwords.
- A single standout example presented as typical.
- Credit claimed for work that was done by the client's in-house team or other vendors.
- Results from a client where the agency's role is unclear.
Questions to put to the agency
- What was the client's situation at the start, and what did you identify as the main constraints?
- Which of the actions did you consider most important, and why?
- What other changes were happening at the same time?
- Can I see the data export or a screenshot with date ranges visible?
- What did not work?
- Was the client's internal team involved and what did they contribute?
- Can I speak to the client?
Good agencies answer these without discomfort, and often volunteer failures and learning points.
Source labels on this site
When you browse the directory, pay attention to where each claim originates. Listings show whether a claim comes from the agency's own site, a published case study or independent corroboration. A figure that appears only on the agency's own site is a claim, not a verified fact. Sponsorship never buys verification, and promoted listings are labeled the same way as others.
Build your own comparison
Put each agency's best case study into a table with the elements above. Gaps become your follow-up list. You will quickly see which agencies can describe their work precisely and which rely on presentation. Then validate with reporting samples, see Understanding SEO Reporting, and the checklist in Questions to Ask Before Signing.