When a founder is comparing agencies, the case studies do a lot of the persuading. They are the second-heaviest thing most buyers lean on to rank one agency over another, right after price. And they are the easiest thing on the whole website to dress up.

I have written our own case studies, and I have read hundreds of other agencies' case studies while helping founders make sense of a shortlist. The pattern is almost always the same. A big number in bold, a happy logo, and none of the context that would tell you whether the result is real or whether it would ever repeat for you.

This guide teaches you to read a case study the way a sharp buyer does. You separate the numbers that look like results from the numbers that are results, and you ask the few questions that make a weak case study fall apart. If you are still building a shortlist, pair this with how to choose a B2B Meta ads agency in India.

Case studies carry weight for a reason. In B2B, buyers trust the experience of another customer far more than anything the agency says about itself. Sopro's buyer research found that nearly 7 in 10 B2B marketers consider case studies their most effective form of content, and over three-quarters of buyers read user reviews before purchasing. So a case study is not just decoration. It is often the thing that tips a decision.

That is exactly why they get polished past the point of honesty. A case study shows you the win the agency wants you to see. It rarely shows the starting point, the budget, the sales motion behind the leads, or how the result was measured. Those missing pieces are where the truth lives.

None of this means case studies are worthless. It means you have to read them actively instead of being impressed passively. The rest of this piece is the checklist I would run in your seat.

The first thing to check is which kind of number the case study is built on. A vanity metric is one that looks impressive but does not inform any decision. Improvado defines it as a data point that is easily manipulated and appears impressive on a superficial level, but fails to provide meaningful insight into business performance or inform future strategy. Impressions, reach, clicks, likes, and raw lead counts almost always sit in that bucket for a founder trying to grow revenue.

Outcome metrics are the opposite. They map to money and they are harder to fake: qualified-lead percentage, cost per qualified lead, pipeline created, revenue, and return on spend that is actually tracked. A case study that leads with these is telling you the agency follows the lead all the way to a sale, not just to a form fill.

There is a simple test you can apply to any headline number. If that number doubled, would it change a decision you make? Impressions doubling changes nothing. Qualified pipeline doubling changes everything. Improvado frames the same test as whether you could make a specific decision based on the metric, and whether it directly influences a business objective like revenue or retention without multiple logical leaps. I go deeper on this split in qualified leads versus cost per lead.

Take the two most common case study headlines. 3x ROAS sounds like a clear win until you ask about margin. On a thin-margin offer, 3x can still lose money once you count the cost of delivery. And ROAS is only as trustworthy as the tracking behind it, which a case study almost never explains.

500 leads is worse. A lead count with no qualification behind it tells you the agency was good at collecting form fills, nothing more. If none of those 500 turned into sales conversations, the number is a cost dressed up as a result. In a long B2B cycle, that gap can hide for months before anyone notices the pipeline is empty.

The fix is to mentally attach four missing pieces to every headline: the baseline it grew from, the timeframe it happened over, the spend behind it, and whether the leads were qualified. A number without those four is not a result you can judge. It is a claim you are being asked to take on faith.

You do not need to be a marketer to interrogate a case study. You need five questions, and you need to watch how comfortably they get answered. The first is the baseline. What were the numbers before the agency started? A jump from 10 to 30 leads is a different story from a jump from 300 to 320, and a good case study tells you which one it is.

The second is timeframe. Over what period did this happen? A result across a clean twelve months is real. A result cherry-picked from the one strong month is a highlight reel. The third is spend. What was the monthly budget behind the number, so you can judge whether it maps to your own budget at all.

The fourth is quality. Were these qualified leads that the client's sales team accepted, or just raw volume the platform reported? The fifth is the one that settles everything: can I speak to that client? A short reference call cuts through every polished slide. An agency confident in its work will make the introduction. If that door quietly closes, you have your answer.

A case study can be completely honest and still be useless to you. The question is whether the win transfers to your business, and that comes down to three checks. The first is industry and buyer. Was the client selling to a buyer like yours, or is this a low-consideration consumer product with nothing in common with your sales motion?

The second is deal size. A stunning result on a five thousand rupee product tells you very little about whether the agency can generate qualified pipeline for a fifteen lakh contract. The buying behaviour, the objections, and the sales support required are different worlds. The third is sales cycle. A same-day purchase and a six-month B2B evaluation need different strategies, and an agency that only has short-cycle wins may never have run an account like yours.

This is why industry-adjacent proof matters more than a bigger headline from an unrelated space. I wrote about how much this actually weighs in whether your B2B ads agency needs industry experience. Judge a case study on how close it sits to your reality, not on how big the number is.

A few patterns reliably mean the proof is thinner than it looks. No numbers at all, just adjectives like massive growth and incredible results. Or numbers that stop at clicks and cost per lead, never reaching pipeline or revenue. Both are ways of looking busy without showing outcomes.

Next: no stated baseline, so you cannot tell what actually changed. A time window that is suspiciously short or oddly specific, which usually means the rest of the timeline was less flattering. And no client name plus no reference you can call, which removes the one thing that would let you verify any of it.

The deepest red flag is results reported purely in platform metrics, with no link back to the client's CRM or sales numbers. That tells you the agency optimised toward what the ad platform counts, not toward what the business earns. These overlap with the warning signs in Meta ads agency red flags. One flag is a caution. Several together mean the case study is decoration.

The reason surface metrics mislead is structural. Gartner finds that B2B buyers spend only about 17 percent of their entire buying journey meeting with potential suppliers, and when comparing several suppliers, just 5 to 6 percent of that time with any one sales rep. Most of the buying happens out of sight, so a case study built on top-of-funnel activity tells you almost nothing about whether real buyers moved toward a sale.

The metrics that survive scrutiny are the ones tied to money. Trade Press Services argues that vanity numbers like impressions, likes, and lead volume rarely demonstrate how marketing impacts pipeline, revenue, or customer acquisition, and should be replaced by conversion rate, cost per acquisition, return on ad spend, and lead-to-opportunity conversion. Those are exactly the numbers a strong case study will show, and the ones a weak one quietly avoids. These are US and global figures, but the logic holds in India too.

“A case study is only proof if you can trace the number back to a baseline, a budget, and a client who will pick up the phone.”

You can see how we present our own numbers, with context rather than bare headlines, on our results page.

Before your next agency call, run each case study through six questions. Does it show a baseline you can see the change from? Does it state a real timeframe and the spend behind the result? Does it report qualified leads or pipeline, not just clicks and cost per lead?

Then three more. Is the client similar to you in industry, deal size, and sales cycle? Is there a name and a reference you could actually call? And does at least one number tie back to revenue rather than a platform dashboard? If a case study fails three or more of these, it is telling you a smaller story than the headline claims. For the wider view of judging an agency once it starts work, read how to measure marketing agency performance.

If you would rather have a second pair of eyes, that is what our free audit is for. Send us the case studies you are weighing and we will tell you which numbers are real and which are decoration.