Insight · Working with agenciesBOFU
How to judge a marketing agency's case studies
The short answer
Case studies are marketing too. A "312% increase in ROAS" might be real, might come from a tiny starting point, or might measure something that didn't help the business at all. The way to judge one is to ask what was measured, against what, over how long — and what the agency actually did.
What a credible case study shows
- The starting point — the baseline numbers before the agency started
- The time frame — dates and duration, not just "within months"
- The metric — revenue, qualified leads, new-customer CAC or margin, not only clicks or platform ROAS
- Absolute numbers or ranges as well as percentages
- What the agency changed — tracking, structure, creative, pages, offers
- What else changed — price, product, seasonality, other channels, the sales team
- How it was measured — the data source and attribution method
Common ways results get flattered
- Percentages on tiny bases — doubling from five leads to ten is "+100%"
- Platform ROAS as proof — it counts sales other channels or brand search drove; see why your ROAS lies
- Cherry-picked periods — comparing festive season with a slow month
- Ignoring spend — more revenue because the budget tripled
- Cost per lead without quality — cheaper leads nobody bought from
- Brand search included — customers who were already searching for the company
- One-off wins — a single viral ad presented as a repeatable method
Questions to ask
- "What was the baseline, and over what period?"
- "Which metric did the client judge you on, and from which system?"
- "How much did spend change over the same period?"
- "What specifically did you change, in what order?"
- "What else changed in the business at the same time?"
- "Did you test incrementality, or is this attributed?"
- "Can I speak to someone on the team who did the work?"
Reading percentages correctly
An illustrative comparison:
| Case A | Case B | |
|---|---|---|
| Headline | "ROAS up 250%" | "New-customer CAC down 28%" |
| What changed | ROAS from 1.0x to 3.5x on a retargeting campaign | CAC from ₹1,450 to ₹1,045 across all paid media |
| Spend | ₹40,000 a month | ₹12 lakh a month |
| Business impact | Unclear — retargeting may have claimed existing sales | Likely meaningful, measured on blended data |
The smaller-sounding headline can be the stronger result.
Anonymised case studies are normal
Many clients don't want competitors to see their numbers or know who their agency is. Anonymised examples — "a Mumbai D2C skincare brand" — are common and legitimate, as long as the method and measurement are clear. PMG doesn't publish client names on this website; on a call we can walk through relevant work within the limits of our confidentiality agreements.
What matters more than case studies
- The diagnosis — what the agency found in a written audit of your account
- The plan — what they'd do first and why
- The people — who would run your account, and what they've done
- The measurement — whether they'll report blended results such as MER from your own data
Our marketing audit is a practical way to see how an agency thinks before you commit.
Turning a case study into a test of fit
Instead of asking whether a case study is impressive, ask whether it's relevant: similar business model, similar spend, similar channels, similar problem. A brilliant result for a ₹5-crore-a-month app brand tells you little about how an agency will handle a ₹4-lakh-a-month clinic. The closer the match, the more the example tells you.
What to ask for instead of logos
- A walkthrough of how they diagnosed a similar problem
- The test log or a sample report from a comparable engagement, anonymised
- A sample audit of your own account
- The people who would work on your account, and what they've run before
- How they measure results, and from which systems
Watch for survivorship bias
Agencies publish their best results, not their average ones. That's normal, but it means case studies show what's possible, not what's typical. Ask what usually happens in the first 90 days, what goes wrong, and how often engagements end early. Honest answers to those questions are more useful than any headline number.
Frequently asked questions
Are agency case studies reliable?
Some are, many are selective. Judge them on baseline, time frame, metric, spend and method rather than the headline percentage.
Should we ask for client references?
You can, though many clients prefer not to be contacted. A conversation with the team who'd do the work is often more useful.
Why do agencies anonymise case studies?
Clients usually ask them to, to keep results and supplier relationships confidential. Anonymised examples are fine if the method and numbers are clear.
What's the most important number in a case study?
Usually the business outcome — new-customer CAC, cost per qualified lead or revenue against total spend — with the baseline and spend alongside it.
What if an agency has no case studies?
Look at their diagnosis of your account, the clarity of their plan and the experience of the people involved.
Should we trust video testimonials more than written case studies?
They show a real client was willing to speak, which helps, but they rarely include numbers. Combine them with clear method and measurement.
Can an agency show results from our own account?
If you've worked with them before, yes. Otherwise a short paid or free audit of your account is the closest equivalent.
Keep reading
Services behind this
Terms used on this page
Talk to a strategist
Thirty minutes. Your numbers. A straight answer.
Book a strategy call with the people who'd actually run your account. We'll tell you what we'd do — or that you don't need us yet.