Why a website does not generate enquiries: check the funnel

“The website is not working” can mean several things: nobody arrives, visitors misunderstand the offer, a form fails, enquiries disappear or the contacts do not fit the business. Find the specific failed transition before commissioning a redesign. A website can accept enquiries correctly while sales fail because of terms or follow-up. Each stage needs its own evidence, an accountable owner and a hypothesis that can actually be checked.

1. Verify that you measure a real enquiry

Follow the path on a phone: open the landing page, complete the form, send it and locate the record with sales. Match its time, source and identifier. A button click, a submission attempt and a saved enquiry are different events. Test repeated submission, a messenger transition and recovery after an error. If analytics celebrates every click while the server rejects the form, the report records conversions the business never received. Improving the site using that report starts from the wrong evidence.

2. Separate traffic volume from intent

Segment visits by source, entry page, device and language. Research queries, a search for a supplier and visits from existing customers expect different answers. Do not judge a campaign landing page against an average for the whole website. Check service and geography alignment: an enquiry about delivery in Samarkand should not lead to terms covering Tashkent only. When relevant visits are scarce, address demand distribution before changing the colour of a button.

3. Check whether the offer is understandable

On the first screen a visitor should identify the service, whether it fits and the next step. Complex products benefit from scope, exclusions, a pricing process and a concrete example of the deliverable. For an individual quote, explain which requirements are needed. Compare the advertisement with the page: an offer of a one-day review cannot become a generic company presentation. Ask people from the intended audience to describe the offer in their own words and compare it with the alternatives they genuinely consider.

4. Find technical and language friction

• Test the first screen, form and buttons on an ordinary phone with an unstable connection.
• Try a local telephone number with spaces, contact paste and clear field errors.
• Check Russian, Uzbek and English in the form, receipt and terms, as well as headings.
• Remove required fields that are unnecessary for the first conversation.
• Confirm that an error keeps the input and offers a working retry.

Speed affects usability, but a PageSpeed score alone does not explain missing sales. Reproduce the obstacle: unreadable content, a covered button, an image wait or a broken submission. The correction should improve the visitor journey, rather than merely change a testing-tool report. Keep the same device and task when verifying it.

5. Inspect delivery and initial follow-up

Illustrative funnel: 2,000 relevant visits, 100 form starts, 70 submissions, 60 saved contacts, 20 qualified enquiries and five customers. Losing ten records between submission and storage is a technical issue, not insufficient demand. Investigate logs and identifiers. Then check assignment, response time and reachability in the chosen channel. Agree the real handling schedule: a night-time enquiry is not being worked on merely because a notification appeared in a manager chat.

6. Assess qualification separately

Define a suitable enquiry before rating channels: a relevant need, service area, contactability and an agreed next step. Mark unreachable contacts, spam and rejection over timing as separate reasons. Otherwise sales might call every unclosed opportunity bad, while advertising calls every completed form good. Both reports conceal the actual loss. Review ambiguous examples together and use one definition across the team.

7. Change one testable cause

Record the hypothesis, metric, segment and observation period according to enquiry volume. For example, a shorter form should increase saved contacts without increasing spam. With a small sample, start with task testing and conversations rather than declaring a winner after two enquiries. Advertising scope: https://samsa.agency/en/services/paid-media. Technical speed: https://samsa.agency/en/services/performance. Published work: https://samsa.agency/en/cases.

Measurement sources

Metrica goals: https://yandex.com/support/metrica/en/general/goals. Qualified leads in Google Ads: https://support.google.com/google-ads/answer/11459091. A configured goal and a quality assessment answer different questions. The sequence above is an investigation plan, not an industry conversion benchmark.