Website not generating leads
Traffic but no enquiries: find where the website loses people
Four things can be true when a site has visitors and no enquiries: nobody useful arrives, the wrong people arrive, people arrive and leave, or the form has been quietly failing. Check the last one first. It takes twenty minutes and it is more common than anyone expects.
Do this before anything else
The twenty-minute check that finds most of these
A surprising proportion of dead enquiry pipelines are a delivery failure rather than a marketing failure, and the site looks perfectly healthy the entire time.
Work through this list in order, with your own phone, off the office network. If the pipeline resumes at step four you have saved yourself a redesign.
The reason this matters is that a broken form is invisible from the inside. The visitor sees a confirmation message, the analytics records a conversion, and nothing arrives anywhere.
- Submit a genuine enquiry from your phone on mobile data, using an email address nobody at the company recognises
- Include a distinctive phrase in the message so you can search every system for it afterwards
- Check the destination inbox, then its spam and promotions folders, then any shared or delegated mailbox it forwards to
- Check whether the form plugin or platform stored the submission even though no email arrived
- Confirm the notification address is not a former employee, a disabled alias or a mailbox nobody opens
- Check whether the email authentication records for your domain were changed the last time anyone touched DNS
- Repeat the test on a different browser and on an ad blocker, since some form scripts fail silently under blocking
- Time how long the page takes to become usable on a mid-range phone, not on your laptop
- Count the fields on the form and remove every one that will not change whether you call the person back
- Reconcile last month's analytics conversion count against the number of enquiries a human actually saw
Differential diagnosis
Four failures that look identical from the dashboard
- Nobody useful arrives.
- The visit count looks respectable and almost none of it is commercial. Blog readers, students, competitors, other countries, referral spam. The tell is a healthy total with almost no sessions on the pages that sell anything. This is not a website problem and no amount of conversion work will touch it — the traffic itself has to change.
- The wrong people arrive.
- Commercial traffic exists and it does not match your business: out of area, out of budget, or looking for a service you do not sell. The tell is a low conversion rate on service pages plus a sales team that dismisses the few enquiries that do come. Usually the messaging is too broad, or campaigns are optimising toward a conversion event that does not represent a sale.
- The right people arrive and leave.
- Sessions on the service pages are decent, time on page is short, and nothing happens. The page has failed to answer a question that mattered — usually price, usually proof, sometimes whether you cover their area. On a phone it may be that the page took so long to settle that the person left before it finished loading.
- People try to convert and cannot.
- The most damaging and the least visible. The form throws a validation error nobody sees, the submit button fails under an ad blocker, the confirmation shows but no email is sent, or notifications route somewhere nobody reads. Form field analytics or a session recording exposes it in an afternoon, and it is often weeks or months old by the time anyone notices.
Isolating it
Four questions that tell the four causes apart
The four failures look the same on a dashboard. They are separated by asking about a different point in the visit each time, and each answer sends you somewhere else.
Do the pages that sell anything get real sessions?
Yes: Commercial traffic exists, so the loss happens somewhere below this point.
No: Nobody useful arrives. That is a traffic problem, and conversion work cannot touch it.
Do the few enquiries you get match what you sell?
Yes: Targeting is sound. Keep going down the list.
No: The wrong people arrive. Look at the messaging, and at what campaigns optimise toward.
Do people reach the form and then abandon it?
Yes: A question went unanswered — usually price, sometimes proof, sometimes coverage.
No: They leave before the form. The page has not said what it is quickly enough.
Does a test submission reach a human inbox?
Yes: The delivery path works, so one of the three findings above is your cause.
No: The form is failing silently. Fix this before you trust any other number on the site.
Sequence
Fix them in cost order, cheapest first
The order below is deliberately not the order agencies propose, because the cheapest fixes are also the least profitable to sell.
Confirm enquiries reach a human
The twenty-minute check above, done properly and repeated for every route in — form, phone, chat, email link. Until this passes, every other measurement on the site is unreliable and every conclusion drawn from it is unsafe.
You get: A verified path from submission to a person
Segment the conversion rate
Separate commercial pages from everything else and calculate the rate for each group. A site average that includes the blog is not a number anyone should act on, and this step frequently reveals that the pages you were about to rebuild are fine.
You get: Conversion rates by page group
Watch people actually use it
Session recordings and form field analytics on the pages that matter. Where do they stop scrolling, which field do they abandon, what do they click that is not a link. Twenty recordings usually produces more actionable findings than a full audit document.
You get: A short list of observed friction points
Remove friction before adding persuasion
Cut fields that do not change whether you call someone back, fix the mobile experience, put the phone number where a thumb can reach it, and answer the price question in whatever form you honestly can. These need no test to justify.
You get: Shipped changes with a dated before-and-after
Then work on the offer
What the visitor is being asked to do, what they get for doing it, and what proof stands behind it. A page that asks a stranger to book a sixty-minute call may convert far worse than one offering a specific, small, useful first step.
You get: A revised primary and secondary call to action
Only now consider structural change
If the pages needed do not exist, the site cannot be edited safely, or the mobile experience cannot be repaired within the current build, a rebuild is justified. Reaching this point with evidence behind it is very different from starting here.
You get: A rebuild case supported by the previous five steps
The standard
What a page that converts actually does
None of this is about design taste. These are the four jobs a commercial page has to do before anything else about it matters.
Says what it is, immediately
A visitor who lands from search decides in seconds whether this page is about their problem. If the first screen is a slogan and a stock photograph, they are working it out from the navigation instead, and many of them will not bother.
Answers the question that stops people
Usually price, sometimes coverage, sometimes whether you handle a case like theirs. You do not have to publish a price list, but a page that dodges the question entirely sends people back to compare you against someone who did not dodge it.
Proves it with something checkable
Real work, real names where permitted, real detail, real credentials. Vague superlatives read as filler to a buyer who has already visited four competitors saying the same words in the same order.
Makes the next step small and reachable
One obvious action, phrased as what the person gets rather than what you want. A tappable phone number, a short form, and a secondary option for people who are not ready to speak to anyone yet.
Where to go next
What each finding turns into
If the site is the problem
If the traffic is the problem
If the numbers cannot be trusted
Questions
What people ask about a site that will not convert
How do I know if my contact form is actually working?
Test it as a stranger would. Use your own phone on mobile data rather than office wifi, submit from an email address nobody at the company recognises, and write something distinctive in the message so you can search for it later.
Then check every destination: the inbox it is meant to reach, that inbox's spam and promotions folders, the CRM, and the form plugin's own stored submissions if it keeps any. If the entry exists in the plugin but never reached an inbox, you have a delivery problem rather than a form problem, and the two are fixed differently.
What is a reasonable conversion rate for a website?
There is no single honest number, and anyone quoting one across all industries is quoting an average that includes businesses nothing like yours. Rates vary enormously by category, price point, urgency and traffic source.
The comparison worth making is against yourself. Segment your commercial pages from your blog, establish your own baseline, and measure change from there. That is more useful than a benchmark you cannot verify.
Do we need a full redesign?
Usually not, and it is the most expensive thing to be wrong about. A redesign is justified when the structure genuinely prevents the fix — the site cannot support the pages you need, is unusable on a phone, or is built on something nobody can safely change.
When the problem is an unclear offer, a long form, a missing proof element or a broken notification, a redesign solves it accidentally at ten times the cost, and often loses the pages that were working.
We do not have much traffic. Can we still improve conversion?
Yes, though not with statistical testing. Below a few hundred conversions a month, split tests will not reach a conclusion before the market changes underneath them, and running them anyway produces confident decisions based on noise.
At low volumes the work is qualitative: session recordings, form abandonment data, asking recent customers what nearly stopped them, and fixing the obvious problems that do not need a test to justify. Save the testing for when the volume supports it.
Our analytics shows conversions but sales sees nothing. How?
Because the analytics is almost certainly counting an arrival at a thank-you page, not an email reaching a person. Those come apart the moment a notification address breaks, and the tracking carries on reporting success.
It also happens when the same visitor is counted repeatedly, when a test submission fires the event, or when a bot completes the form. Reconcile the conversion count against what the sales team can actually see in their inbox or CRM, for the same dates, before trusting either number.
Find the failure point before you pay for a rebuild
Send us the site and read access to your analytics. We will run the checks on this page, tell you which of the four causes is live, and say plainly if the answer is a broken notification rather than a project.
If we don't deliver the work we agreed to deliver for reasons within our control, you don't pay for the undelivered work. Read our guarantee
Last updated · Reviewed by Zubair Afzal