A testimonial tells a seller you were pleasant to work with. A case study tells them what happened when the first offer collapsed three days before closing. Only one of those answers the question a nervous seller is actually asking.
What a case study does that a review cannot
Our post on testimonials that convince covers how to collect and present client praise, which is worth doing. This is a different asset with a different job.
A review is an opinion about you. A case study is a record of a transaction. The seller reading your site at 10pm has already assumed you are competent and friendly, because every agent's site says so. What they have not assumed is that you have dealt with a house like theirs, in the condition theirs is in, with the complication theirs has.
That is why specificity does the work here. "Sarah made the process stress-free" could be about any agent in your market. "The house had sat for 94 days with another brokerage, the photos were shot in November on an overcast afternoon, and we relisted in March after a reshoot and a price correction based on three comparable sales the first listing had ignored" could only be about one deal. The second one is checkable, and the reader can feel that.
There is a timing reason this matters more than agents expect. NAR's 2025 Profile of Home Buyers and Sellers found that 80 percent of sellers contacted only one agent before hiring. The long listing presentation where you finally get to tell your best story often never happens. The page tells it instead, or nobody hears it.
The four-part structure
Every case study worth reading has the same four parts, in the same order. Skipping one is what makes most attempts read like a press release.
The situation. One paragraph on where the client was when they reached you. Property type, rough location, condition, and the circumstance driving the sale. Specific enough that someone in the same spot recognises themselves in it.
The actual obstacle. This is the part most agents soften into nothing. "The market was challenging" is not an obstacle. "The seller needed to close before their new mortgage commitment expired in 60 days, and the house had a finished basement without permits" is an obstacle. Write the real one, including the awkward ones: an overpriced expectation, a tenant refusing showings, a failed inspection, a co-owner who kept changing their mind.
What you did. The decisions, in order, with reasoning. Not activities. "Marketed the property aggressively" is an activity and means nothing. "Pulled the listing for 11 days, had the basement drywall opened to document the framing for the inspection file, and relisted 4 percent lower with a clean disclosure package" is a decision anyone can evaluate. Include the decision the client pushed back on, and say how that conversation went.
The outcome. Numbers, with dates. Days on market, final price against list price, how many offers, what the previous attempt had produced. If it did not end perfectly, say so and say why. A case study where everything went right reads like marketing. One that names a tradeoff reads like a record.
Pick the hard deal
The instinct is to write about the best result. The house that got six offers in a weekend and closed 9 percent over asking feels like the obvious choice.
It is usually the weakest one you could pick. A skeptical seller reads that story and concludes the market was hot, which it was, and that the house would have sold without you, which it probably would have. Nothing in it demonstrates a decision that changed the outcome.
The deal that proves something is the one that nearly did not happen. The listing that had already expired with another brokerage. The property with a legal complication. The seller who insisted on a price you disagreed with, and what you did after the first six weeks confirmed the problem. Our post on expired listings covers why those sellers are the most receptive audience for this kind of proof, because they have already lived the failure your case study describes.
Difficulty is what makes your judgement visible. Write about the deals where your judgement was the variable.
Permission, privacy, and how to anonymise
Never publish a client's story without asking, and never treat a verbal yes as the record.
Send a short email that lists exactly what will appear: the figures, whether they will be named, whether the street will be mentioned, and a link to a draft if you have one. Ask them to reply with a yes. That email protects you, and it surfaces the detail they are uncomfortable with before it goes live rather than after. People agree warmly at closing and feel differently when they see their divorce on a website.
The sale price is usually public record in North American markets, so republishing it is rarely the exposed part. The risk is the combination. Price plus street plus month plus a sentence about why they were selling can identify one household on one block, even with no name attached.
Anonymising is straightforward once you know what carries the value. Strip the name, the exact address, the precise date, and any detail a neighbour would recognise. Keep the property type, the rough area, the numbers, the obstacle, and the sequence of decisions. "A 1970s split-level in a north-side neighbourhood, listed the previous spring with no offers" tells the reader everything useful and identifies nobody. The decisions are the content. The address never was.
Privacy rules for client information vary by province, state, and real estate board, and some brokerages have their own policy on publishing transaction detail. Check yours, and take proper advice rather than ours.
Three beat thirty
Agents who accept the case study argument often try to produce one for every deal they have closed. That is the wrong shape.
Three to five thorough case studies outperform thirty thin ones, for two reasons. The first is attention: nobody reads thirty, and a long index of similar stories signals volume rather than judgement. The second is that thin ones are worse than none. A four-sentence summary that says a house sold quickly for a happy client is a testimonial wearing a case study's clothes, and it dilutes the pages that actually do the work.
Choose by coverage instead of by count. Pick the three or four client situations you are asked about most often, and write one case study for each. A relisted expired property. A downsizing seller managing a move. A first purchase with a tight budget. An investment property. Four stories that each match a different visitor's situation will outwork twenty stories that all match the same one.
Where case studies belong in the site
Give every case study its own page with its own web address, and link them all from a single index page.
Putting five stories on one page splits that page's relevance five ways, so it matches nothing in particular when someone searches. Separate pages let each story compete for a different search: a neighbourhood name, a property type, or a phrase like a house that did not sell the first time.
Then link them where the relevant visitor already is. A case study about a condo sale belongs on your condo content and on the neighbourhood page for that building's area. A case study about a relisted property belongs on your page for sellers whose listing expired. Your sold listings page should link to the case studies that go deeper on individual sales, because a grid of sold photos is a claim and the case study is the evidence behind it.
They also belong in the listing presentation. The same page you send as a link before the meeting is the page you open during it.
Why AI assistants can quote a case study
When someone asks an AI assistant which agent handles difficult listings in a particular area, the assistant has to produce sentences it can stand behind. It reads pages, finds statements with a subject and a verifiable outcome, and repeats those.
A star rating gives it nothing. There is no sentence an assistant can construct from four-point-nine stars that is useful to the person asking, and assistants avoid repeating unattributed praise about a named person.
A case study gives it plenty. "Relisted after 94 days with no offers, repriced against three comparable sales, under contract in 19 days" is a complete factual statement about a specific outcome. That is the shape of thing an assistant can quote in an answer, with a link back to you.
Two habits make this more likely. Write in complete sentences rather than bullet fragments, because fragments lose their subject when lifted out of context. And put the outcome in plain words near the top, so the summary is available without the assistant having to infer it. Our post on content that earns AI citations covers the wider pattern.
What to do this week
Open your transaction history and find the deal you still tell people about at the office. That is almost always the right first case study, because the reason you retell it is that something in it was genuinely hard.
Write the four parts before you write anything else: situation, obstacle, decisions, outcome. Keep it to one page. Then email the client the draft and ask for a written yes, or anonymise it and ask anyway.
Do that three times over a month and you have the only proof asset on your site that a competitor cannot copy, because they did not do the deal.
The takeaway
A case study earns trust by being checkable. Pick the transaction that nearly fell apart, write the obstacle honestly, show the decisions in order, and publish the real numbers with the client's written permission. Three of those on their own pages will do more for a hesitant seller than another twenty reviews, and they give AI assistants something they can actually repeat about you.
