If a website is not generating leads, the cause sits on one of four funnel levels: no traffic reaches the site, the wrong traffic arrives, the page fails to convince the visitor, or the form, speed and lack of trust block the final step. You can check all four levels in about 30 minutes using GA4 and Google Search Console reports — no vendors and no "40-page comprehensive audits". Below is the step-by-step method: which report to open at each step, which numbers count as normal, and what to fix when they are not.
Why "12 reasons" won't help: you need a method, not a list
Search for "why my website doesn't sell" and you get dozens of listicles: outdated design, no USP, slow loading, few reviews, clunky forms… All true, and all useless. A list never answers the only question that matters: which of the reasons is yours. Rebuilding the site against all twelve points at once means spending a redesign budget when the actual problem might be that only 40 people a month visit the site.
A website produces a lead only when four links fire in sequence: the visitor arrived → arrived for what you sell → the page convinced them → the form did not stop them from leaving a contact. That is a funnel, and it always tears in one specific place — the weak link you have to find. Fixing a link that is not torn is pointless: form conversion will not save a site with no traffic, and more traffic will not save a page people abandon after ten seconds.
Preparation: what the diagnostic cannot work without
You will need access to your site's Google Analytics 4 property. Before starting, verify three things:
- The tag is installed and collecting data. Open the reports overview: if the visits graph is empty or there is no tag — the diagnostic is postponed two weeks while statistics accumulate. Installing the tag is 15 minutes of work.
- Conversion events are configured. The minimum: a key event for form submission and one for clicks on the phone number or messenger. Without them GA4 sees visits but not leads — there is nothing to compute conversion from. If there are no key events, that is the audit's first finding: you have been steering the site blind.
- Session recordings are enabled. Install Microsoft Clarity (free) or a similar tool: recordings of real sessions show how a person scrolls, where they stall, what they type into the form and where they give up. Steps 3 and 4 are impossible without it. Data accumulates from the moment you switch it on.
Analyze the last 90 days: shorter windows are noisy. And pin down the base math we will keep checking against:
Step 1. Is there any traffic at all
The step's question: do enough people reach the site for leads to be statistically possible?
Where to look: GA4 → Reports → Acquisition → Traffic acquisition, for 90 days. Check total monthly sessions and the channel split: organic search, paid, referral, direct, social.
Checkpoints:
- Fewer than ~300 visits a month — diagnosis made, no need to go further: the site has an acquisition problem, not a conversion problem. Even a perfect page converting at 3% yields 5–9 leads from such traffic, and a typical 1% yields three. No redesign will change that.
- Almost all traffic is direct and "unassigned" — real visitors are even fewer than they look: direct traffic hides your own employees and yourself.
- Traffic exists but one channel delivers 90%+ — you can proceed with the funnel check, but the business is sitting on a single needle; log it as a risk in the plan.
What to fix if there is no traffic. That is separate work with two tools of different speeds: PPC advertising delivers targeted visits from week one, but every visit is paid; SEO takes 3–4 months to ramp, but afterwards organic traffic stops charging you per click. How the budget for the second path works and when it pays off, we covered in the article on the cost of SEO; for the paid path, the same math is in our breakdown of the Google Ads budget. Important: before pouring paid traffic in, still run steps 3–4 of this diagnostic — advertising into a page that does not convert is buying visits into a leaky bucket.
Step 2. Is it the right traffic
The step's question: are the people arriving actually looking for what you sell — or just "people"?
Where to look: three reports.
- GA4 → Traffic acquisition → engagement rate and bounce rate per channel. A bounce in GA4 is a session with no engagement: the visitor glanced and closed. For a commercial site a healthy per-channel bounce rate is roughly under 40%. Above ~60% on a channel — that channel is bringing the wrong audience. The classic: paid traffic bouncing at 70% while organic bounces at 25% — the problem is not the site, it is the campaign targeting.
- Google Search Console → Performance → Queries. Read them with your own eyes: what do people actually come from search for? If you sell custom web development but your traffic collects on "how to build a website for free", those visitors never intended to order anything — no conversion rate will produce leads from them.
- GA4 → User attributes → Demographic details → geography. If you serve one metro area and half the visits come from other regions and countries, your real audience is half the reported one — and conversion must be computed against it.
What to fix if the traffic is wrong. Not the page — the aim: in paid campaigns, keyword cleanup, negative keywords, geo and audience corrections; in SEO, rebuilding pages around commercial queries instead of informational ones. And check query-to-page match: someone who clicked an ad for "kitchen remodeling" must land on the kitchen remodeling page, not on a generic homepage.
Step 3. Does the page convince
The step's question: the right visitor arrived — why do they leave without becoming a lead?
Where to look: GA4 → Engagement → Conversions (the rate for your form-submit and phone-click key events), then Clarity's scroll maps and click maps, then session recordings.
Checkpoints:
- Engagement is fine, conversion below 0.5% — people read but are not convinced. That is the "page doesn't sell" zone.
- Scroll map: fewer than half reach the middle of the page — the first screen gives no reason to read on. Look at it honestly: is it clear within 5 seconds what you offer, for whom, and how you differ? "Innovative solutions for your business" is not an offer, it is a screensaver.
- Click map: visitors click on non-clickable things (images, plan names) — the page is not answering a question people actually have. Most often it is pricing: if there are no prices at all, part of the audience leaves to find them at a competitor's.
- Session recordings: watch 10–15 sessions longer than a minute with no conversion. The most honest report of all: you see where a person freezes, what they re-read, where they backtrack and on what they close the tab. The exit point usually repeats from session to session — that is your broken block.
What to fix. In descending order of impact: the first-screen offer (what, for whom, the benefit, the next step), prices or at least honest "from" ranges, social proof next to the price (case studies with numbers, reviews, portfolio), a clear CTA on every screen. This is work on meaning, not on "beauty" — sometimes a targeted repackaging of the page solves it, sometimes it runs into the site being simply obsolete and easier to rebuild: that is a web development conversation, and the boundary is simple — when the improvement quote crosses half the cost of a new site, improving is a bad deal.
Step 4. Form, speed, trust
The step's question: the visitor decided to leave a request — what stopped them from finishing?
This is the most painful level of loss: the person had already decided, the money was on the table. Check three things.
The form
In your session recordings, filter for sessions with focus on a form field but no submission. The typical lead killers: more than 3–4 required fields (every extra field cuts submissions), a captcha, hostile validation ("enter a valid phone number" — valid how?), and a form with no success message — the person cannot tell whether the request went through and does not write again. Separately, test it by hand from your phone: submit a request yourself. Surprisingly often the form turns out to be simply broken after some update — and "the site stopped generating leads" has exactly that cause.
Speed and mobile
In GA4, check the share of mobile traffic (Tech → Overview) — for most niches it is 60–80% of visits. Then run the key pages through PageSpeed Insights: the target is LCP (main content load) under 2.5 seconds on mobile. A slow site loses visitors before the first screen, and those people barely register in analytics — they left before the tag loaded. If the mobile version is awkward (tiny text, overlapping elements, a form you cannot fill with a thumb), you are losing the majority of your audience silently.
Trust
Before leaving a phone number, a person looks for confirmation that you exist and will not let them down: real company details and an address, live photos instead of stock, verifiable case studies and reviews, working social links and a maintained Google Business Profile. Check the outer perimeter too: search "your company name + reviews" — if the first thing a customer sees is two-year-old negativity with no reply, leads are lost before they ever reach the site.
Summary table: symptom → cause → how to check → what to fix
The whole diagnostic on one screen. Find the row with your symptom — and you have the step to start from.
| Symptom | Likely cause | How to check | What to fix |
|---|---|---|---|
| Fewer than ~300 visits a month | No traffic: invisible in search, no ads running | GA4 → Traffic acquisition; Search Console → indexing and queries | Acquisition channels: SEO for the long game, PPC for leads now |
| Traffic exists, per-channel bounce rate above ~60% | Wrong traffic: untargeted queries, ads missing the audience | GA4 bounce/engagement by source + Search Console queries, geography | Keywords, negatives, geo; landing pages matched to query intent |
| Engagement fine, conversion below 0.5% | The page does not convince: weak offer, no prices or trust | Scroll map, click map, 10–15 session recordings | First screen, offer, prices, case studies and reviews, CTA |
| They reach the form but do not submit | The form: extra fields, captcha, validation errors, breakage | Session recordings with form focus; submit a request yourself from a phone | 3–4 fields max, fix validation, add messengers as an alternative |
| Heavy mobile traffic, conversions only on desktop | Speed or the mobile experience | PageSpeed Insights: LCP under 2.5 s; walk the site by hand on a phone | Speed optimization; if the platform is old — a new site |
| Analytics shows leads, but "there are no customers" | Handling: leads get lost after submission | Reconcile GA4 key events with CRM/email; send a test lead and time the response | Notifications, a CRM, a 15–30 minute response policy |
If you checked everything and there are still no leads
Walked all four steps, the numbers are normal, and customers still are not coming? Two zones remain that lie outside the website.
- Leads arrive but get lost. Compare the number of form-submit key events in GA4 with the number of leads your sales team actually saw. Any gap is emails in spam, a broken integration, or a rep who calls back the next day when the customer has already bought elsewhere. Send a test request and time the response: anything over 30 minutes during business hours is lost money.
- The site is fine — the offer's economics are not. If your offer is objectively weaker than competitors' (more expensive with no explainable reason, slower, no guarantees), the site merely reports that honestly. Then the work is not markup but meaning: positioning, packaging, differentiation — a separate discipline we cover in our brand positioning service.
And the main thing: this diagnostic is not a one-off event. Set a monthly reminder to open the four reports from this article — 15 minutes of routine control catches problems (a broken form, a sagging channel, rising bounces) weeks before they turn into "we somehow have no customers".