Why your website isn't converting: a diagnostic

"Converting" just means a visitor does what you built the site for. They buy, book, call or fill in your form.

The order matters because it saves you money. A new headline can't help if your contact form is broken. A redesign can't help if only 40 people saw the old site.

Start with what you're seeing

Don't know your numbers? Start at Check 1 anyway. All you need for it is a phone.

Check 1: Is something broken?

This takes five minutes, and it's the most painful thing to find out late. Pick up your phone, turn off wifi and use your site like a stranger:

• Fill in your own form. Does a thank-you message appear? Does the message reach you? Look in your spam folder too.
• Tap your phone number. Does it dial? What does a caller hear after 5pm?
• Open your booking page. Are there any open times?
• Buy the cheapest thing you sell, if you sell online. Use a real card, then refund yourself.

Next, compare your reports with real life. Count last month's actual calls, emails, bookings and orders. Then see what your analytics tool says. That's the program that counts your visits, usually Google Analytics.

"Sometimes the website is converting and analytics is wrong." Neil Webley, founder of FreeCROTool

Here's one way that happens. Someone reads your site, then picks up the phone, and no report counts that as a win. So for the next two weeks, ask every new customer, "How did you find us?" Write down the answers.

Last, check how fast you reply. Someone who fills in your form may have filled in two others that same evening. If you answer two days later, the site did its job and the sale was lost after that. Try to reply the same day.

If you find something broken, fix it today. You don't need to test a broken thing.

Check 2: Have enough people seen it?

Look at the page that sells, not your whole site. In Google Analytics, go to Reports, then Engagement, then Landing page. It lists the pages people arrive on and how many came to each.

Now ask whether that's enough visits to judge. Here's an example. Say your page would normally get 2 out of every 100 visitors to act. Plain bad luck could still leave you with nothing:

So under about 150 visits, you don't know anything yet. Changing the page now is guessing. After 300 visits with nothing, something really is wrong, so move on to Check 3. If you sell something expensive, fewer people will act, so give it longer.

If you're short on visitors, leave the site alone and go get some:

• Message 20 people who could really use what you sell. Send them the page and ask what would stop them from buying.
• Ask your last ten happy customers who else they know. Then make it easy for them to pass your link along.
• Answer questions where your customers already hang out. Local Facebook groups and trade forums count. Read the rules first.
• Run one small ad to one page. Spend only what you can afford to lose. You're buying information.

Check 3: Are they the right people?

Ten visitors who might buy are worth more than a thousand who never will. Look in three places:

• Where they came from. In Google Analytics, go to Reports, then Acquisition, then Traffic acquisition. Which of those places ever brings a sale or a call?
• What they typed. Search Console is a free Google tool. Its Performance report lists the searches that found you. "How to fix a dripping tap" and "emergency plumber near me" both bring visits. Only one of them brings a call.
• Where they live. A dentist in Denver gains nothing from 400 visits from Germany.

Watch for one trap. Your busiest page may be an old blog post. Those readers wanted an answer, they got it, and they left happy. Don't count them with the people who reached a page that sells.

If you run ads, read your ad and then open the page it leads to. If the ad says "Free quote in 24 hours" and the page says "Welcome to our family business," people will think they clicked the wrong thing. Put the ad's promise in the page's headline.

Check 4: Do they get it in five seconds?

Show the top of your page, on a phone, to someone who doesn't know your business. Give them five seconds and take the phone back. Then ask three questions: What do we sell? Who is it for? What would you do next?

People often leave a web page within 10 to 20 seconds, says website researcher Jakob Nielsen. You don't get long.

A common problem is a headline that sounds nice and says nothing. Take "Tailored solutions for your unique needs." A plumber, a bank and a wedding photographer could all put that on their homepage, and somewhere all three already have.

Say what you do, who it's for and one thing they care about:

Only promise what's true, of course. Then give the page one main button, and make its words say what happens. "Get my quote" beats "Submit."

Check all this on a phone, not your big office screen. If far fewer people act on phones than on computers, look at your mobile conversion rate next. That's how many phone visitors end up buying or booking.

Check 5: Do they believe you?

Someone who has never heard of you has four quiet questions. Your page should answer each one.

Are you real? Show your face or your team, your name, the area you serve and a phone number that works. Skip the stock photo of people laughing at a laptop.

Have you done this for someone like me? One review with facts in it beats ten with none. Compare "Great service!" with "Furnace died on Sunday. They came by 3pm and the bill matched the quote." Use real reviews, with permission, and put the best one near your button.

What will it cost? If you can't give a fixed price, give a range and say what changes it. If your page says nothing about price, people have to guess. Some will guess high and leave without asking.

What happens after I click? People hold back when they fear a pushy sales call. Add one line by the button that tells them what comes next, such as "We'll call you within one working day. The quote is free."

Check 6: Is the last step easy?

Your form. Go through it box by box. Ask, "What do I do with this answer before our first call?" If the answer is nothing, delete the box.

Your checkout. If you sell online, show the full cost early. In Baymard's survey of US shoppers, 4 in 10 people who weren't just browsing said they left a cart over extra costs like shipping and tax. No other reason came close. Let people check out without making an account too.

Your phone. Try the whole thing with one thumb. Does a cookie banner or chat bubble sit on top of your button? Does the keyboard hide the next step? Does the page load before you lose patience?

Real people. Microsoft Clarity is free and lets you watch recordings of real visits. Look for the spot where people keep stopping. Better still, ask five people who could be customers to try it on their phones while you watch and say nothing. It's uncomfortable. Nielsen has long said that small tests with about five people teach you the most for the least effort.

Check 7: Do they want the offer?

Sometimes the right people understand you, trust you and can act easily, and they still don't. Then the page may be fine and the offer may be the problem.

Look up the three businesses your customer would compare you with. Are you pricier, slower or riskier? If so, either say why you're worth it or change something.

Try asking for less. A busy person may not book a "consultation," but they might take a free 15-minute call. They may skip a full job and try a small starter one. A shop can offer a sample or a smaller size.

Then ask the last three people who said no, "What made you go with someone else?" Ask kindly and you'll often get an honest answer. That answer is worth more than a month of reports.

Still stuck? Ask the people leaving

Everything above is you working out what strangers are thinking. The people about to leave already know. So put one question on your main page, and show it after someone has been there a while:

What's stopping you from booking today?

Change "booking" to whatever your page asks for. Offer answers people can tap, because tapping is easier than typing on a phone. Then use what you hear:

When they need an answer right there

A question tells you what's stopping someone, but they still leave. If people keep saying "I'm not sure this is right for me," they needed help while they were on the page. That's what we built Possible to do, so keep in mind that this is our own product.

Possible is an AI built on real sales psychology. It asks each visitor the right question at the right moment, remembers the answers and guides them to a sale, sign-up or booking. It's free to start. Here's our AI sales assistant for clinics and local services. It can't fix a broken form or bring you visitors, so do Checks 1 to 3 first.

What to leave alone for now

• A redesign. It costs the most and teaches you the least, because you change everything at once.
• More ads. If the page isn't working, you just pay to get the same result faster.
• Ten changes in one weekend. If things improve, you won't know why. Make one change, write down the date and your numbers, and look again in a few weeks.
• Fancy testing. Good ab testing tools compare two versions of a page, and they need plenty of visitors to tell you anything. With a few hundred visits a month, just fix what's clearly wrong.
• A giant to-do list. A conversion rate optimization checklist is handy after this diagnosis. Before it, you'd be fixing things at random.

Once the big problem is gone, you can start on website optimization. That just means making small improvements, one at a time.

If you only have an hour today

1. 10 minutes: Do Check 1 on your phone, and fix anything broken.
2. 10 minutes: Find the visits to the page that sells, and compare them with the luck table.
3. 10 minutes: Run the five-second test on one person.
4. 10 minutes: Delete every form box you don't need.
5. 20 minutes: Rewrite your headline so a stranger could say what you do.

Questions about a website that isn't converting