Web DesignMarketing

Your Website Gets Traffic But No Enquiries: Finding the Leak

Kage Works9 September 20269 min read

Two thousand visitors a month and three enquiries feels broken. Whether it is broken depends on a number most owners never work out, and on four places an enquiry can vanish between someone landing on your site and your phone ringing. One of those places sits after the enquiry arrives, which is why a redesign sometimes changes nothing.

Work through them in order. The cheapest checks come first, and one of them often turns "no enquiries" into "enquiries nobody was counting".

Work out the number you should be seeing

Ruler Analytics publishes the most usable benchmark set for this, drawn from more than 110 million sessions and 5 million conversions across 13 industries, published on 26 May 2026. A conversion there means a qualified lead or a sale, counted across both web forms and tracked phone calls. The average across all industries came out at 5.13%, with wide spread underneath it: automotive and legal at 7.9%, software 7.6%, education and finance 6.3%, professional services 6.1%, real estate 2.8%, retail and ecommerce 2.4%, travel 1.9% (Ruler Analytics).

Read those as a ceiling rather than a target. Ruler's sample skews toward businesses running tracked paid campaigns with call tracking already installed, and the spend behind it is UK-weighted. A New Zealand site living on organic and Google Business Profile traffic will usually sit below its industry row. The useful part is the order of magnitude and the gap between industries, which tells you that comparing a trades site to an ecommerce store proves nothing.

Run the arithmetic once. An accountancy firm taking 900 sessions a month, held against the 6.1% professional services figure, would be looking at roughly 55 enquiries at the top end. Something in the twenties would still be a working site. Four is a leak worth chasing.

The same arithmetic settles the opposite case. A firm counting 25 enquiries off 900 sessions is converting fine and short of visitors, which makes a redesign the wrong purchase and traffic the right one. Our SEO guide for NZ small businesses covers that half.

About half your enquiries may never touch a form

The same Ruler data breaks conversions by channel, and the split changes what "no enquiries" means. Legal took 56.3% of its conversions by phone. Professional services took 52.6%. At the other end, marketing and advertising ran 95.1% through forms and beauty and cosmetics 92.2%.

A New Zealand plumber or accountant sits at the phone end of that range, and Google Analytics 4 cannot see a phone call by default. It records the tap on a tel: link if someone set that up, and it records nothing at all when a customer reads the number off your header and dials from a different phone. If half your enquiries arrive as calls and your dashboard only counts form submissions, your real conversion rate is roughly double what you are looking at, and you are about to redesign a page that works.

Two checks settle it inside a week. Ask whoever answers the phone to write down where each caller says they found you. Then confirm your tel: link taps are firing as a GA4 event, which takes a developer about twenty minutes. Our local SEO guide for trades goes further into call tracking, because the same blind spot ruins most attempts to measure whether a Google Business Profile is earning its keep.

Check whether the traffic is the buying kind

Traffic volume and traffic intent move independently, and a rising session count with flat enquiries usually means the mix changed rather than the page.

Split your top landing pages into two groups: pages someone reads while deciding to buy, and pages someone reads while looking something up. A page ranking for "how much does rewiring cost in Auckland" brings people who might hire you. A page ranking for "how does a home switchboard work" brings people who never will. Blog traffic flatters a report and converts near zero, which is fine as long as nobody is judging the site on the blended average.

Google's AI Overviews have made this worse and better at the same time. Pew Research tracked nearly 69,000 Google searches by 900 US adults in March 2025 and found people clicked a traditional result in 8% of searches carrying an AI summary, against 15% without one (Pew Research Center). The clicks disappearing first are the informational ones, since those are the questions a summary can finish. If your session count fell while your enquiry count held steady, the site is converting better than it was and losing traffic that was never going to buy. That is a different diagnosis and a different response. Our guide to getting cited in AI answers covers what to do about the visibility side.

The form is rarely the villain people name

"Cut your form to three fields" gets repeated as settled fact. HubSpot's own analysis of more than 40,000 customer landing pages, updated in June 2025, is more measured than the advice built on top of it: conversion rates fall as fields increase, but the decline is slight, and the field type matters more than the count. Extra single-line text inputs barely register. Multi-line text areas and dropdown select boxes are what depress conversion (HubSpot).

Unbounce's analysis of its own customer data points the other way past a certain length, with conversion climbing again beyond the four-to-seven field dip, which fits the idea that people will answer more questions when the answers visibly shape what they get back (reviewed by Cobloom). Nobody has published a New Zealand dataset on this at all.

The honest reading: drop the free-text box asking people to describe their project and the dropdown asking them to categorise themselves, then keep every field that lets you quote without sending a second email. Field count is not the lever. A phone number visible without scrolling, and a form that behaves on a phone screen, will each outperform any amount of trimming.

The same caution applies to adding a chatbot as a fix. Humii's Online CX Index, which analyses real online shopping journeys across Australia and New Zealand, added chatbot obstruction as an abandonment category at the start of 2026 at 3.4% of pre-checkout abandonment reasons, and had it near 7% by mid-year, reaching 35% at some retailers (Inside Retail New Zealand). That is retail data rather than lead generation, so read it as a warning about burying the human contact route rather than a verdict on chatbots. Our piece on what chatbots cost and when they pay goes into where they earn their place.

The leak that happens after the form

This is the one that no redesign touches, and the evidence behind it is old but unusually clean.

Oldroyd, McElheran and Elkington audited 2,241 US companies by submitting a web-generated test lead to each and timing the reply. 37% responded within an hour. 16% took between one and 24 hours. 24% took longer than a day, and 23% never replied at all. Among the firms that did respond within 30 days, the average response time was 42 hours (Harvard Business Review, March 2011).

The same authors ran a second study across 1.25 million sales leads received by 29 B2C and 13 B2B US companies. Firms that tried to make contact within an hour of a query arriving were nearly seven times as likely to qualify the lead, which they defined as having a meaningful conversation with a decision maker, as firms that tried an hour later. Against companies that waited 24 hours or more, the multiple was over 60.

The related lead response management study, using three years of data from six companies covering more than 15,000 leads and over 100,000 call attempts, found the odds of qualifying a prospect dropped roughly 21-fold between a five-minute and a thirty-minute response (Lead Response Management).

Both are US datasets, and the more recent one was published in 2011. Buyer behaviour has not slowed down since. Map it onto a New Zealand trades business: an enquiry lands at 9:40am while the owner is under a house, gets read at 6pm, and gets answered the next morning. The website did its job. The enquiry was already cold, and the customer had emailed two competitors at the same time.

Fixing this is cheap compared with a rebuild. An automatic acknowledgement inside a minute that names a real person and a real callback window buys you most of the first hour. Routing enquiries to a phone rather than an inbox nobody opens on site buys the rest. At higher volumes, sorting enquiries by urgency before a human reads them is one of the more defensible uses of workflow automation for a small business.

Run the checks in this order

Count your real enquiries first, calls included, for two weeks. Divide by sessions and put the result next to your industry row above. If you land within range, the site is fine and the budget belongs in traffic.

If the gap is real, look at your top five landing pages and separate buying intent from curiosity before you change anything. Then time your own reply speed by submitting a test enquiry through your own form on a Tuesday morning and seeing how long it takes to hear back. Most owners are surprised, and that number is the cheapest thing on this list to fix.

Page load time belongs in the same pass, since Google, Deloitte and 55 measured 30 million user sessions across 37 brand sites and found a 0.1 second improvement in mobile speed lifted retail conversions 8.4% and travel conversions 10.1% (web.dev case study). We cover the diagnosis in why your website is slow.

Redesign last. A rebuild is the right answer when the structure itself blocks people from buying, and our comparison of redesign against rebuild covers when that point arrives. Reaching for it before you have counted your phone calls and timed your replies is how businesses spend five figures to fix a problem that lived in an inbox.