Web Design Case Study: How to Write One, With Example

A web design case study is a short account of one client project. It names the client, shows what was wrong with the old site, says what you changed and reports the result in numbers the client has approved. This guide gives the six parts, the numbers to record before you start building and one full example. The example is made up.

Most pages that rank for this search are finished case studies or galleries of them. They show what one looks like. This guide is for the agency that has to write its own.

What is a web design case study?

A web design case study is the story of one project, told as problem, work and result. A portfolio piece shows what a site looks like. A case study shows why the site was rebuilt and what happened next.

The reader is a business owner deciding whether to trust you. That owner has one question. Did this work for a business like mine?

So the best web design case studies are specific. They cover one client, one trade, one town, one problem and one measured change.

What goes into a web design case study?

Six parts go into it, in the order below. Together they fit on one page.

Do not skip part 6. A page that admits a limit is easier to believe than a page with none.

Which numbers should you record before the build starts?

Record the baseline before you change anything. You cannot measure the old site once it is gone, so a case study starts on the first day of the project.

Ask for three full months of each business number. One month can be a fluke, and a roofer's March looks nothing like a roofer's July.

Save every figure with its date and its source in one file. You will need them when the client approves the page. Our website audit checklist lists what else to look at on the old site.

How do you report the result honestly?

Report raw numbers for equal periods, name the source and say what else changed. An honest result is smaller than a boastful one, and far easier to believe.

1. Give the raw numbers. "From 9 enquiries a month to 16" tells the reader more than "78% more enquiries".
2. Compare equal periods. For a seasonal trade, set three months against the same three months a year earlier.
3. Name the source. Say whether a number comes from the client's booking system, from Search Console or from your own test.
4. Say what else changed. New ads, a second location and a price change all move the numbers.
5. Claim only what the site can explain. Speed and mobile fixes are yours. A rise in revenue has many causes.

US guidance on advertising points the same way. The FTC says an endorsement that claims specific results will usually be read as what others can expect, and that a line such as "Results not typical" does not change that reading (FTC). It also says an endorsement must reflect the endorser's honest opinion. The FTC describes an endorsement as an advertising message that people are likely to believe reflects the opinion of someone other than the advertiser. In our reading, a client quote in a case study fits that description, so write it with that guidance in mind. This is not legal advice.

Our own research is reported this way. Our October 2026 study found that companies with the weakest websites replied positively to 1.60% of generic emails and to 2.35% of emails that named the problem. It then states that the gap was not statistically significant, with p = 0.238. The test covered Dutch companies and email only, and we say so each time we quote it. Limits like these take two sentences, and they are why the other numbers can be trusted.

What does a finished website case study look like?

A finished website case study holds the six parts on one page. The one below is made up. The clinic, the people and every number are invented to show the format, and none of it is a real result.

Example, made up. Larkspur Veterinary Clinic, Boise, Idaho

The client. Larkspur Veterinary Clinic is a two-vet practice in Boise that treats dogs and cats. Most new clients find it by searching on a phone.

The problem. We reviewed larkspurvet.example on March 3, 2026 and found four things.

1. The main content took 6.8 seconds to appear on a phone. Google's mark for a good page is 2.5 seconds.
2. The "Book a visit" button sat below three screens of text.
3. The phone number, 208-555-0142, was an image and could not be tapped.
4. All services shared one page, so the clinic did not show up for "dog dental cleaning Boise".

The work. We rebuilt the site on a lighter theme and compressed every image. We put a booking button and a tap-to-call number at the top of each page. We wrote one page for each of the clinic's 7 services. The new site went live on March 30, 2026.

The result. Business numbers are monthly averages for April to June 2025 and April to June 2026, from the clinic's booking system and Google Search Console. Speed was tested with PageSpeed Insights.

• Largest Contentful Paint on a phone went from 6.8 seconds to 2.1 seconds.
• Online booking requests went from 11 a month to 19 a month.
• Calls from the website went from 34 a month to 41 a month.
• Search clicks for service terms went from 120 a month to 210 a month.

The client's words. "People used to phone to ask if we were open. Now they book." Dr. Erin Castillo, owner

The limits. The clinic added Saturday hours in May 2026, which will have helped bookings. We measured three months, and a full year would say more.

Figures confirmed by the clinic on July 8, 2026.

Look at what the example leaves out. It makes no claim about revenue. It gives no percentage without the two numbers behind it. It names the Saturday hours, even though they weaken the story.

How do you get the client to agree?

Ask at the start of the project, and ask again when you have numbers. A client who agreed in the contract already expects the request.

1. Put one line in the contract. It says you may write up the project and that nothing is published without the client's approval. Our web design contract template has a portfolio clause to start from.
2. Record the baseline in week one. Tell the client why you are asking for the numbers.
3. Draft the page yourself, about 90 days after launch. Take the quote from what the client said on a call or in an email, and let them edit it.
4. Send the whole page for approval. Ask for a written yes to the figures and the quote, and keep the email.
5. Offer a version without the name. "A two-vet clinic in Boise" still works if the owner wants privacy.

None of the five pages we read for this post says who approved its numbers. Yours should, and one dated line at the foot of the page is enough.

Where do you use a case study once it is written?

Use it wherever a stranger is deciding whether to trust you. That means your site, your proposals and your outreach.

On your site. Give each case study its own page, linked from the service page for that trade.

In proposals. Put the closest match above the price in your web design proposal.

In outreach. Send it as a second or third message to a business in the same trade with the same problem. Our cold email follow-up templates include one built around a similar business's result.

Next to a price. A cold prospect has two questions, which are whether you can do the work and what it will cost. The case study answers the first. Our guide on how much to charge for a website helps with the second.

Match the case study to the prospect by the finding. If it is about a slow site with a buried booking button, send it to businesses whose sites have that problem.

Leadodo covers the first half of that match and leaves the proof to you. Twip, the AI inside Leadodo, scores every website in a search out of 100 across 10 domains. Its findings become the opening line of each letter and email in your Outbound Machine. Leadodo does not write case studies.

Questions

How long should a web design case study be?

As a guide, 300 to 500 words on one page. An owner will read a short page about a business like theirs. Put the result in the first two lines and the detail under it.

What if I have no results to show yet?

Start with what you can measure yourself. Load time, mobile layout and the number of taps to reach the booking button need no client data. Record them before and after, and add business numbers once the client has three months of them.

Can I write a case study without naming the client?

Yes, with the client's permission. Describe the trade, the town and the size of the business. A named client with a quote is more convincing, so ask for that first.

What is the difference between a case study and a testimonial?

A testimonial is the client's opinion in a sentence or two. A case study is the whole project, with the problem, the work and a measured result. A good case study holds a testimonial inside it.

---

Next step. A case study is most useful when it reaches an owner with the same problem. Leadodo finds businesses, scores their websites and runs outreach to the ones with the weakest sites. See the features.