Website Conversion Audit: 7 Questions to Find Your Leaks

A website conversion audit finds the places where visitors who could become customers leave without calling, filling in a form or buying. I run it as seven questions, one for each layer of the C.O.N.V.E.R.T. Method from my book Decoding The Click. Each question has a test you can run yourself, a number to watch and a free tool to watch it with.

On any new site, I start the same way. I act like a visitor. What is this site trying to tell me or sell me? What is its main goal, and does the page achieve it? The seven questions turn that first look into checks you can measure.

This is the half of website work that SEO alone does not cover. One piece of common advice I disagree with is “optimize your website for Google first, and the user experience will follow.” It runs the wrong way. Rankings bring visitors to a page. The page decides whether they become customers.

The website conversion audit checklist

Here are the seven questions, in the order of the C.O.N.V.E.R.T. Method. Answer each one for every key page, on a phone first, and note the evidence behind your answer.

One check sits under all seven: speed on a phone. In my book I put it this way: “A website that is clear but slow is a promise unfulfilled.” Google’s Core Web Vitals guidance sets three targets. Aim for Largest Contentful Paint within 2.5 seconds and Interaction to Next Paint under 200 milliseconds. Keep Cumulative Layout Shift under 0.1. I check these in PageSpeed Insights, one of the free Google tools I use most.

Make sure GA4 counts conversions before you judge a page

An audit without numbers is an opinion. Before you run the seven questions, decide what counts as a conversion on your site. Then check that Google Analytics 4 records it.

In GA4, a key event is an event that measures an action particularly important to your business. Any event you collect can be marked as one. For most service businesses, that means:

• Form submissions. Google’s recommended event for this is generate_lead, for when someone submits a form or a request for information.
• Clicks on phone, WhatsApp and email buttons. A visitor who calls you has converted, but form tracking will not count the call.
• Bookings or purchases, if the site takes them.

GA4’s enhanced measurement can record form starts and submissions automatically. Send a real test inquiry and confirm it shows up before you trust the numbers. Keep names, phone numbers and emails out of Analytics, because Google requires that no personally identifiable information is collected.

For a large Gurugram real estate developer, my team at Adtric set up GA4 events for inquiry forms, contact clicks and project filters. They went live in June 2025, so the monthly report showed leads, not just visits. From June to September 2025, it recorded 2,414 inquiry form completions from 1,782 people who came from Google (case study).

Once key events work, compare each channel’s share of visits with its share of conversions. The two can be far apart. Here is how it looked on three sites my team at Adtric has worked on.

• An Indian maker of transmission towers: organic search brought 30% of sessions but 75% of key events, March to August 2026 (case study).
• A 3PL warehousing company: organic search brought 106 of 170 contact form submissions, or 62%, July 2025 to September 2026 (case study).
• A small finance bank: organic search brought 5,194 of 5,623 Apply Now forms, or 92%, January to June 2026 (case study).

Judged by sessions alone, the manufacturer’s organic search would have looked like under a third of the result. If your traffic grows but inquiries do not, also read why website traffic isn’t converting.

Then choose the pages to audit. GA4’s Landing page report (Reports, then Engagement, then Landing page) shows sessions and key events for the first page of each visit. Audit the pages where most visits start, plus services, pricing, contact and the form itself. When I have access, Search Console and GA4 are my most useful audit tools.

Questions 1 and 2: can visitors understand you and see what to do next?

Clarity: the five-second test

Clarity comes first because every other layer depends on it. A visitor who cannot tell what you do never reads your offer or your proof.

To test it, show the page for five seconds to someone who has never seen your site. Hide it, then ask three questions. What does this company offer? What do you remember? Who is it for? Write the answers down word for word. With three to five people, a pattern shows quickly. I explain the full test in the 5-second rule.

When the test fails, start with the headline. My book’s formula is outcome plus qualifier: the result the visitor wants, plus a detail that makes it specific. In the book, I use the example of two landscapers. One says “Quality Landscaping Services”. The other says “Get a Barbecue-Ready Lawn in Under 14 Days”. Only the second tells a hurried homeowner what they get and by when.

For pages that get search traffic, add one more check. Search Console’s Performance report shows which queries bring visitors to your site. Read each landing page’s headline against those queries. If they do not match, the visitor has to work out whether they are in the right place.

Obvious action: one clear next step

Once visitors understand you, the next step should be the easiest thing on the page to find. Check these on a phone:

• One primary button per page. If Call, WhatsApp, Download and Subscribe all look equally important, none of them is.
• A label that says what happens. It should complete the sentence “I want to…”. “Get my quote” beats “Submit”.
• A short form. Ask only what you need to reply. Every extra field is one more reason to stop.
• A familiar layout. Logo top left and linked home, menu where people expect it. In my book I call a clever, unfamiliar layout a backward bicycle.

For a quick visual check, step back from the screen and squint. The main button should still stand out. Then compare form starts with form submissions in GA4. Many starts and few submissions point to the form itself. Microsoft Clarity, a free tool, has click maps that show clicks on things that are not links. Those clicks mean visitors expected something to happen.

Questions 3 and 4: is there an honest reason to act now, and is the offer worth it?

Now: real urgency only

Interested visitors default to “I’ll think about it.” Give them a true reason not to wait. Good reasons are real: an admission deadline, limited slots each month, a price change on a stated date, or stock that is actually low.

Fake urgency costs more than it earns. In my book, I use the example of an online coffee seller whose 15-minute countdown restarts every time a buyer returns. The buyer who notices never trusts that brand again.

In India it is also a legal risk. On November 30, 2023, the Central Consumer Protection Authority issued guidelines listing 13 specified dark patterns. The government’s answer in the Lok Sabha names false urgency as one example. So check that every deadline, counter and “only a few left” message on the site is true today.

Value: sell the result, not the service list

Visitors do not buy a list of features. They buy the change the service makes for them. Check whether each service page describes the visitor’s situation after the work, before it lists what you do.

Then check what lowers the risk of saying yes: a guarantee you can honor, a free first call, clear refund or cancellation terms. If you cannot publish a price, say what it depends on. In my book, I call the next step a value stack. List every part of the offer, so the visitor compares the price with everything it includes, not with nothing.

To measure it, use GA4’s scroll event, which records when someone reaches 90% of a page’s depth. Few scrolls on a long service page, or recordings of visitors who reach the price and leave, point to a value problem.

Question 5: can visitors check your claims where they hesitate?

Visitors arrive skeptical. Before they act, they look for proof that you are real and that your claims hold up. Check three things:

• Specific testimonials. A name, a role or company, and a result. My book’s test is before, after and bridge: the problem, the outcome and what made the difference. “Great service” from “John D.” proves nothing.
• Checkable facts. A full address, a phone number someone answers, the named people behind the business, and case studies with dates.
• Placement. Proof next to the form, the price and the checkout, where hesitation peaks. A testimonials page only reaches visitors patient enough to look for it.

Dates are what make proof checkable. Every case study I link on this page gives a time window with its numbers. A reader can see what was measured and when.

To measure it, use a Clarity scroll map to see whether visitors scroll far enough to reach your proof. If you move a strong testimonial next to the form, compare form completions over equal periods before and after.

Questions 6 and 7: where do visitors leave, and did your fix work?

Refinement: find the leak before you fix it

Numbers tell you where visitors leave. Recordings tell you why. You need both.

In GA4, a funnel exploration shows the steps users take to complete a task. Build one from landing page to service page to contact page to form submission. Use an open funnel if visitors can enter at any step, for example straight onto the contact page from search. Break it down by device category, and start with the step that loses the most people.

Then find out why with Clarity’s session recordings. Watch recordings of visitors who reached that step and left. Look for hesitation, scrolling back and forth, and fields people give up on.

Turn what you see into a hypothesis before you change anything. My book gives a template for it.

“Based on [your qualitative insight], we believe that making [a specific change] will result in [a measurable improvement], because [the reason you believe it will work].”

Testing: let visitor behavior decide

A fix you have not measured is still a guess. In my book, I call the usual alternative the HiPPO: the highest-paid person’s opinion. It sounds confident, and it is often wrong.

An A/B test splits visitors between the current page and one changed version over the same period. Google Optimize is no longer an option: Google says it has not been available since September 30, 2023.

Test without hurting your search traffic. Google’s A/B testing guidance says:

• Show Googlebot the same versions people see. Showing it something different is cloaking.
• Point variant URLs to the original with rel="canonical", rather than using noindex.
• If the test redirects visitors, use a 302 temporary redirect, not a 301.
• Remove every element of the test as soon as it ends.

Many small business sites do not get enough conversions for a fair split test. Then make one change at a time and write down the date. Compare key events over equal periods before and after. It is weaker evidence than a split test, but far better than opinion.

Fix the leaks in this order

Do not fix everything at once, because each change muddies the measurement of the others. Work in this order:

1. Tracking. Until key events work, you cannot tell whether any fix helped.
2. Clarity. In my book I call it the gateway to every other layer.
3. Obvious action. A hidden or competing next step loses people at the last moment.
4. Evidence. People who understand you still need a reason to believe you.
5. Value and Now. Sharpen the offer, then add an honest reason to act today.
6. Refinement and testing. These never finish. Each test points to the next leak.

If you only have time for three fixes, choose these. Make the headline pass the five-second test, give every key page one obvious next step, and put proof beside every form. These three come first because each one can stop a visitor who already wants what you sell.

Run the audit again after a redesign, a change of offer, or a shift in where your traffic comes from.

Run the seven questions on your own site

Start this week. Confirm that GA4 counts your inquiries and calls. Pick the pages where most visits start. Run the five-second test on each, then work through the table above.

The questions rest on four instincts every visitor brings: they are lazy, impatient, judgmental and skeptical. Chapter 1 of my book explains them, and you can read it free.

If you would rather have an outside view, try my website conversion review. It runs these seven questions on your key pages, phone first, and backs its findings with your GA4 data.