Short answer: When a website gets traffic but no leads, the cause is almost always one of five things: leads are happening but not recorded, the traffic has the wrong intent or location, the page does not make a clear offer, the page is slow or awkward on mobile, or enquiries are not followed up. Test them in that order using the 21 checks below, starting with tracking, because a broken form or event makes every other number meaningless. As a yardstick, Unbounce's 2024 analysis of 41,000+ landing pages found a median conversion rate of 6.6%; a blog-heavy site will sit far below that and still be healthy.
Why does my website get traffic but no leads, and what is normal?
First separate "no leads" from "fewer leads than I expected". Two benchmarks help:
- Landing pages built to convert: Unbounce's 2024 Conversion Benchmark Report analysed 57 million conversions across 41,000+ landing pages and found a 6.6% median across industries. Its industry medians range from 3.8% (SaaS) to 12.3% (entertainment), with education at 8.4%, financial services 8.3%, legal 6.3% and professional services 6.1%.
- Whole websites with a blog: much lower, because most blog readers are researching, not buying. A site where 70% of sessions land on articles can have a site-wide rate under 1% while its service pages convert well.
So never judge the site-wide number. Judge conversion rate per landing page type. The same report found about 83% of visits were on mobile while desktop converted 8% better, which is why so many checks below are mobile tests.
Each check has a test you can run yourself in Google Analytics 4, Search Console, PageSpeed Insights or a phone, and the red flag that tells you it is the problem.
Checks 1–3: Are leads happening but not being recorded?
| # | Check | How to test | Red flag and fix |
|---|---|---|---|
| 1 | Lead events actually fire | Submit your own form, tap call and WhatsApp on a phone, with GTM Preview and GA4 DebugView open | No event, or an event on button click rather than success. Fix tracking before anything else. |
| 2 | Leads arrive somewhere | Count last month's enquiries in the inbox, CRM, WhatsApp and call log; compare with GA4 key events | Real enquiries exist but are not recorded, or forms silently fail (spam filter, broken SMTP). |
| 3 | Thank-you step works on mobile | Submit from Android Chrome and iPhone Safari on mobile data | Error page, endless spinner, or a success message nobody sees below the fold. |
Start here because it is the fastest to test and the most embarrassing to miss. A common pattern: the form "works" for the visitor, but notification emails go to an unmonitored inbox or to spam. Our guide to form tracking in GA4 with Tag Manager covers reliable success events.
Checks 4–8: Is the traffic the right traffic?
| # | Check | How to test | Red flag and fix |
|---|---|---|---|
| 4 | Which pages get the traffic | GA4 > Reports > Engagement > Landing page, sorted by sessions; tag each page as blog, service, location or home | Most sessions land on blog posts or the home page, not on pages that sell anything. |
| 5 | What people searched | Search Console > Performance > Queries, filtered to your top 10 landing pages | Queries are definitions, "free", "jobs", "salary", "course" or another city. That is traffic, not demand. |
| 6 | Where the visitors are | GA4 Landing page report with Country and City added; Search Console Countries | Large share from outside your service area (common for blogs ranking in the US or Pakistan). |
| 7 | Paid traffic quality | Google Ads > Search terms; Meta placement breakdown | Broad match matching irrelevant searches; Audience Network or in-app placements with high clicks and no enquiries. Add negatives, exclude placements. |
| 8 | Engagement on money pages | GA4 engagement rate for service and landing pages | Engagement rate far below site average on a page that should convert: wrong traffic or a page that fails in the first screen. |
GA4 counts a session as engaged when it lasts longer than 10 seconds, has a key event, or has two or more page views; engagement rate is the share of sessions that qualify. In Search Console, the Performance report shows clicks, impressions, CTR and average position by query, page, country and device, which is enough to answer checks 5 and 6.
A traffic rise from AI-era informational content deserves extra scepticism. Pew Research found that Google users clicked a traditional result in 8% of visits when an AI summary appeared, against 15% without one, so the clicks that do arrive on "what is" queries are often from students and researchers. If checks 4–6 fail, the fix is not more traffic but different pages: service + city pages, pricing pages and comparison pages that match buying queries. Our SEO service builds exactly those.
Checks 9–14: Does the page make an offer worth responding to?
| # | Check | How to test | Red flag and fix |
|---|---|---|---|
| 9 | The first screen says who, what, where | Show the page on a phone to someone outside the company for 5 seconds; ask what you sell and to whom | They cannot say. Replace slogans with "Physiotherapy at home in Noida and Greater Noida, same-day visits". |
| 10 | There is one clear next step | Count calls to action above the fold | Three or more competing buttons, or only "Submit". Use one action that names the outcome: "Get the price list", "Book a site visit". |
| 11 | The offer is worth giving a number for | Ask: would I fill this if I were comparing three providers? | Generic "Contact us". Offer something useful: price range, availability check, sample report, callback within 10 minutes. |
| 12 | Price expectations are addressed | Search the page for ₹ or "starting from" | No price signal at all. Visitors leave to find one. A range or "from ₹" filters out poor-fit leads too. |
| 13 | Proof is specific | Look for named clients, results with numbers, reviews, address, registration numbers | Stock photos and "100% satisfaction". Add verifiable evidence; in regulated sectors show RERA or council registration. |
| 14 | Copy is easy to read | Paste the main copy into a readability checker | Long paragraphs and jargon. Unbounce found pages at a 5th–7th grade reading level converted at 11.1% vs 5.3% for professional-level copy. |
Check 12 is the one Indian service sites most often skip. Buyers expect to negotiate, but they still want to know if you are in the ₹5,000 or the ₹50,000 range before calling. Real estate sites have a specific version of these problems (missing price sheet, no RERA number, a brochure gate) covered in real estate landing page optimisation.
Checks 15–18: Is the page fast and easy on a phone?
| # | Check | How to test | Red flag and fix |
|---|---|---|---|
| 15 | Core Web Vitals pass on mobile | PageSpeed Insights for your top 5 landing pages; read the field data | LCP over 2.5 s, INP over 200 ms or CLS over 0.1 at the 75th percentile. |
| 16 | Mobile layout works | Use the page on a mid-range Android phone: tap targets, sticky bars, pop-ups | Chat widget covering the form, pop-up on entry, phone number not tappable. |
| 17 | Form length fits the offer | Count fields; check GA4 for form_start vs generate_lead if tracked | More than 4–5 fields for a first enquiry. Ask name, phone, one qualifying question; collect the rest on the call. |
| 18 | Form errors are friendly | Enter a 10-digit number with +91, with spaces, with a leading 0 | Valid Indian formats rejected, or errors that wipe the form. |
Google's Core Web Vitals thresholds are LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less, measured at the 75th percentile of page loads. PageSpeed Insights shows field data from real Chrome users over the previous 28 days, plus a Lighthouse lab test; trust the field data where it exists. The usual Indian culprits are uncompressed hero images, several chat and analytics widgets loading at once, and heavy page builders on shared hosting. Converting images with our WebP converter is a quick first win.
Checks 19–21: Are enquiries followed up?
| # | Check | How to test | Red flag and fix |
|---|---|---|---|
| 19 | Response time | Submit a test enquiry at 11 a.m. and at 7 p.m. on a weekday; time the first reply | No reply within an hour. Many "no leads" complaints are really "no follow-up". |
| 20 | Contact channels match the audience | Check for click-to-call, WhatsApp and form on every service page | Only a form for an audience that prefers to call or chat, or only WhatsApp for B2B buyers who need a quote by email. |
| 21 | Source reaches the CRM | Look at the last 20 CRM leads | Source blank or "website" for everything, so you cannot tell which pages produce leads. |
If check 19 fails, no website change will help. An enquiry that waits until tomorrow has usually spoken to two competitors. Automated WhatsApp acknowledgement plus a named owner fixes most of it; see WhatsApp lead management automation.
Which checks should you fix first?
Run the checks in the order below. It is sorted by how cheaply each group can be tested and how badly a failure distorts the other checks, not by page position.
| Priority | Checks | Time to test | Typical fix effort | Why this order |
|---|---|---|---|---|
| 1 | 1–3 Tracking | 30 minutes | Hours | A recording failure makes every other diagnosis wrong |
| 2 | 19–21 Follow-up | 1 day (test enquiries) | Days | Leads may exist but die unanswered |
| 3 | 4–8 Traffic intent | 1–2 hours | Weeks (new pages, negatives) | Tells you whether the page or the audience is the problem |
| 4 | 9–14 Offer and page | 2 hours | Days | Highest conversion impact once traffic is right |
| 5 | 15–18 Speed and forms | 1 hour | Days to weeks | Matters most when most traffic is mobile |
Do not redesign the whole site first. A redesign changes 15 variables at once, so you never learn which one mattered, and it often breaks tracking (check 1) in the process.
Worked example: how the checks change the diagnosis
Illustrative numbers, not client data. A Noida interior-design firm has 10,000 sessions a month and 12 enquiries, a 0.12% rate.
- Check 4: 7,000 sessions land on blog posts about "modular kitchen design ideas"; 2,000 on the home page; 1,000 on service pages.
- Checks 5–6: 40% of blog sessions come from outside India. Blog posts produced 1 enquiry.
- Check 2: the CRM shows 12 enquiries, but the call log shows 19 more calls from the website number that nobody logged.
- Service pages: 1,000 sessions produced 11 form leads plus about 15 of the calls, so 26 enquiries, a 2.6% rate.
The problem is no longer "the website does not work". It is "service pages get 1,000 sessions and convert at 2.6%". Two actions follow: build city and service pages to grow that 1,000 (checks 5–6), and lift the 2.6% towards the 6.1% professional-services median with a price range and a clear offer (checks 11–12). At 1,000 sessions, moving from 2.6% to 5% adds 24 enquiries a month. Plug your own numbers into our lead funnel calculator, and if you want an outside team to run the checks, our conversion-focused website team does this audit before any redesign.
Sources
- Unbounce: 2024 Conversion Benchmark Report: 6.6% median, sample size, mobile share, reading-level finding.
- Unbounce: What's a good conversion rate?: industry medians.
- web.dev: Web Vitals: LCP, INP, CLS thresholds and 75th percentile.
- PageSpeed Insights: About: field data from CrUX over 28 days; lab data from Lighthouse.
- GA4 Help: Engagement rate: engaged session definition.
- Search Console Help: Performance report: metrics and dimensions.
- Pew Research Center (July 2025): click rates with and without AI summaries.