Your website gets traffic but no enquiries. Here's what to check first

Traffic is up. The phone is quiet. The contact form has produced three spam messages and nothing you can quote.

Before changing the homepage, separate two questions:

Are the right people arriving?
Can they take the next step?

Those are different problems. A site can attract thousands of people who were never likely to buy, or lose ten ideal prospects to a broken form. Both look like “website traffic but no leads” in a monthly report. Follow the path from measurement to enquiry instead of guessing.

1. Make sure the problem is real

Start with measurement because a missing enquiry and an unrecorded enquiry look identical in analytics.

Write down what counts as a lead for your business: a completed form, a phone call, a booking, a quote request or perhaps a qualified email. Then compare three places for the same period:

your website analytics;
the actual form inbox, booking system and phone log;
your CRM or whatever list you use to follow up.

They may disagree. A form can reach the inbox without firing an analytics event, or a “conversion” can fire even though the form fails. Messages may also be sitting in spam.

Google Analytics recommends generate_lead when somebody submits a form or requests information. It does not send that event automatically; it needs to be set up, and Google recommends checking events in Realtime or DebugView after setup (Google Analytics). Treat that as a measurement convention, not proof that the lead arrived. The confirmation page, inbox and CRM still need to agree.

If nothing is measured yet, start a dated spreadsheet with source, landing page, action and whether the enquiry was qualified. A small reliable record beats a polished report built on the wrong event.

2. Check the intent behind the traffic

“Traffic” describes people, bots and intentions that have almost nothing in common.

A visitor searching “roof repair cost in Orlando” may be comparing contractors. Somebody reading “how roof trusses work” may be a student or a homeowner doing it themselves. Both count as organic sessions; only one looks close to an enquiry.

In Google Search Console, open the Performance report and inspect:

the queries producing clicks;
the pages receiving those clicks;
the countries and devices involved;
changes over a useful period, not one unusual day.

Search Console is free and shows how the site appears in Google, including performance, indexing and page-experience information (Google Search Console). If you have not set it up, start there. Our separate guide explains what to check when your website is not showing up on Google.

For paid traffic, compare the ad promise, search terms, location targeting and landing page. For social, ask whether the post attracted buyers or broad curiosity. An irrelevant referral is still irrelevant traffic.

High-traffic pages that answer early research questions are not necessarily failing. Give readers a route to the relevant service. The SEO work is to attract demand your business can serve, not merely enlarge the session graph.

3. Compare the promise with the page

Now open the most common landing page as if you had never heard of the business. Does its first screen answer four things?

What do you do?
Who is it for?
Where do you do it?
What should the visitor do next?

The page should continue the conversation that brought the person there. An ad for emergency air-conditioning repair should not land on a general homepage that gives equal space to unrelated services.

This is page-message match: partly copy, partly structure and only sometimes visual design. People scan online rather than reading every word; Nielsen Norman Group’s eyetracking research across more than 500 participants found that this behaviour persisted through changing designs (NN/g). Put the answer in descriptive headings and plain text, not a slogan the visitor must decode.

Good web design gives each page a job. The examples in our work are deliberately different: a contractor needs a scoped bid request, a nail studio needs a short booking path, and a lawn service wants a phone call. The job decides the page.

4. Put proof where doubt appears

Visitors need evidence for the risk they are considering.

For a local service business, that might be:

a real address, service area and working phone number;
current reviews on an independent platform;
photographs of completed work, the premises or the people who will arrive;
licences, insurance or professional memberships where relevant;
a clear process, so the first call does not feel like a trap;
case examples that say what was done without inventing an outcome.

Place proof beside the claim it supports: “fully insured” near the service decision, a relevant review near hesitation, and a portfolio before the quote request.

Do not replace contact details with a form. In NN/g’s study of 20 business professionals using 40 corporate websites, people expected phone numbers and email addresses on contact pages and looked for contact access in the navigation or footer (NN/g contact-page research).

5. Make the offer concrete

Sometimes the site is clear and credible, but the offer remains hard to evaluate.

“Tailored solutions for all your needs” says nothing. A useful service page explains what the customer gets, who it suits, what happens first, what affects price and what commitment is required. If you cannot publish a fixed price, give a starting point, common scope or the factors that move it. If you offer a free consultation, say what happens and what the visitor leaves with.

6. Give each visitor a relevant next step

A call to action is not a brighter button. It is a clear answer to “what happens if I click this?”

Match it to the page and the visitor’s readiness:

Request a roof inspection is clearer than Get started.
Check appointment availability is clearer than Contact us.
Send the drawings for a bid is clearer than Learn more.

Use one primary action per page, with a phone or email alternative where useful. Repeat it after the information needed to decide. Five equal buttons ask the customer to design their own journey.

7. Test the form, call and booking flow yourself

Do this before debating button colour.

Open a private window on your phone, turn off Wi-Fi and arrive through the same page a customer would. Then:

Tap the phone number. Check that it opens the dialler with the correct number.
Submit the form using a real external email address and realistic answers. Try one validation error as well.
Confirm that the button visibly responds and cannot create accidental duplicate submissions.
Check that errors explain what to fix without deleting completed fields.
Check the success state. It should confirm receipt and say when and how the business will reply.
Verify that the message reaches the correct inbox, includes the submitted details and does not land in spam.
Complete the booking flow through the final confirmation. Check time zone, availability, confirmation email and calendar entry.
Confirm the lead appears in analytics and the CRM once—not zero times and not twice.

Repeat important flows on desktop and a second phone if possible. If the form uses an integration, retest after either system changes. Reliable forms, booking and CRM connections are development work, not something a finished-looking page proves.

8. Use the site like a customer on a phone

A page can be technically responsive and still be unpleasant to use.

Check whether the main message is visible without dismissing a stack of banners. Make sure buttons are easy to tap, labels remain visible while typing, the right keyboard appears for phone and email fields, and the form is not asking for information the first conversation could collect later. NN/g’s mobile-input checklist begins with the useful question: is each field absolutely necessary? (NN/g).

Use a real device, not only a desktop browser narrowed to phone width.

9. Check speed in context

A slow page can interrupt every step above, but a fast irrelevant page still will not generate leads.

Run the main landing pages through PageSpeed Insights. Google’s current Core Web Vitals cover loading, interactivity and visual stability. Its “good” thresholds are Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint at 200 milliseconds or less, and Cumulative Layout Shift at 0.1 or less, assessed at the 75th percentile of page loads (web.dev).

Use those numbers to find problems. A late booking widget, oversized hero image or moving button deserves attention. A score of 96 instead of 99 probably does not explain an empty inbox.

10. Inspect what happens after the enquiry

The website’s job does not end at Submit.

Check who receives the notification, owns the reply and covers after hours. Send an immediate confirmation with a realistic response window. If enquiries sit unassigned in a shared inbox, the conversion problem is operational.

Review quality too. Ten requests outside your service area are not better than three jobs you can quote. Feed those patterns back into targeting and copy.

A symptom-to-test table

Symptom	Likely cause	Evidence to inspect	First test
Traffic rises, enquiries do not	New traffic has weak buying intent	Queries, sources, locations and landing pages	Isolate one high-intent source and compare its enquiry rate
One campaign gets clicks but no leads	Ad and landing page do not match	Ad copy, search terms and first screen	Send the campaign to a page dedicated to that promise
People visit service pages but do not act	Message, offer or proof is incomplete	Page headings, pricing context, reviews and recordings if available	Rewrite the first screen and add one relevant proof point
Form starts but few completions	Too many fields, poor mobile input or errors	Field-level behaviour and manual tests	Remove every field not needed for the first reply
Form reports success but inbox is empty	Delivery or integration failure	Form logs, spam, routing and CRM	Submit from an external address and trace it end to end
Mobile traffic performs worse	Tap, typing, layout or speed friction	Real-device flow and device-segmented data	Fix the first blocking task on the most-used landing page
Leads arrive but sales say “website leads are poor”	Broad targeting or weak qualification	Search terms, enquiry content and service-area fit	Tighten the page promise and add one qualifying question
Analytics reports no leads but staff receive them	Measurement failure	Realtime events, confirmation state and CRM	Trigger one test lead and reconcile all three records

Traffic problems that look like conversion problems

You may not have enough qualified visits to expect a steady stream of enquiries. A blog post can inflate sessions while commercial pages remain almost unseen. An ad platform can report clicks from the wrong geography. Search visibility can disappear while direct and referral visits make the overall chart look stable.

If the right service pages receive little relevant traffic, work on discovery and targeting before rebuilding them. See our Google visibility checklist.

Conversion problems that look like traffic problems

The reverse happens too. A business sees few enquiries and buys more traffic, but the contact link is buried, the offer is vague or the form fails on mobile. More visitors then produce a larger version of the same disappointment.

Test the full path before increasing spend. If qualified people arrive and cannot understand, trust or contact the business, acquisition is not the first repair.

A one-afternoon audit with no paid tools

Set aside three hours and use what you already have.

First hour: prove the numbers. Compare 30–90 days in analytics, Search Console, inboxes, bookings, call logs and your CRM. List the top five landing pages and sources, then mark which could produce a customer.

Second hour: follow the journey. On a real phone, open the top commercial pages, write down each promise, then test calls, forms and bookings through confirmation, delivery and analytics.

Third hour: make a short repair list. Score each issue by impact and confidence. Fix broken delivery and misleading measurement first, then high-intent mismatch, missing information and mobile friction. Leave decorative preferences at the bottom unless they obstruct the task.

You should finish with one diagnosed constraint and one test, not a 46-item redesign wishlist.

Run experiments your traffic can support

High-traffic sites can split visitors between two versions and wait for enough outcomes. Low-volume service sites often cannot.

If a page gets 200 qualified visits and four enquiries, one extra enquiry moves the apparent rate from 2% to 2.5%: a 25% lift caused by one person. It may be the change, seasonality, a referral or chance. Tiny percentage movements are weak evidence when the count is tiny.

Optimizely’s low-traffic guidance makes the same point: small differences need more visitors and conversions to separate effect from noise (Optimizely).

So match the method to the volume:

Very low volume: fix failures, speak to recent customers, watch people attempt the task and make one substantial change.
Moderate volume: change one meaningful part, record the date and compare qualified enquiries across ordinary weekly variation.
High volume: calculate the sample needed before an A/B test and choose a change large enough to matter commercially.

Do not cycle button colours every Friday. Test a stronger hypothesis: a campaign-specific page, a specific service and area, clearer pricing context or a shorter form.

The diagnosis should tell you whether the next job is targeting, copy, trust, design, development or follow-up. Sometimes the honest answer is that the website is fine.

If you would like a plain review, send us the site and tell us where the traffic comes from. We will trace the route to an enquiry and tell you what appears to be failing first, including when the answer is not a redesign.