Why is my website not converting? The diagnosis and the fixes

Traffic with no enquiries means a break between the visit and the reply, or conversions nobody counts. Before anyone touches the design, test whether a real enquiry reaches the business and your tracking records it. Then work through traffic fit, the offer, proof, the contact task, speed and follow up. Fix what the evidence shows.

Are enquiries missing or just missing from reporting?

Start with the business record, not the analytics report. Ask whoever answers the phone and the inbox what is actually arriving. Relevant enquiries that your reports don't show point to measurement. A pile of spam and duplicates points to the form and its filters. Genuine silence points to the site and its traffic.

Then define the action the page exists to produce. That could be an accepted enquiry, an answered call worth taking, a completed booking or a purchase.

Scrolls, button clicks, taps on a call link and submitted requests are earlier stages. They aren't interchangeable. A tap doesn't prove anyone answered. A booking request isn't a confirmed appointment.

Only the business can say what counts as suitable. That's a request for a service you sell, in an area you cover, from someone you could help. Analytics alone can't tell you whether a lead arrived, qualified or got a reply.

Run a labelled test enquiry

Send a test through a priority page, once on a phone and once on a desktop. Label it clearly so nobody quotes on it. Check that the form validates and shows a receipt. Then follow the request to the staff notification or CRM entry. See who it was assigned to and whether someone would have followed it up.

Delivery is where the cheapest failures hide. Across 100 website audits from May 2023 to September 2026, we sent labelled test enquiries. About 1 in 5 sites with an enquiry form failed the test. The request didn't reach the person who answers it within one business hour. The count covers a spam folder, a wrong notification address, a stopped integration and a queue nobody owned.

Then use Tag Assistant or debug mode to check that the intended analytics event appears once. Google's Monitor events in DebugView page shows how to watch events arrive in real time. Neither tool proves a person received the lead or acted on it. Test with the consent choices your visitors make. Keep your test activity out of any performance conclusions.

Expect this step to turn something up. Google's About key events page defines a key event. It's one that measures an action that's particularly important to the success of your business.

Here's what the same audits found across all sites. About 3 in 4 had no working GA4 key event for the main enquiry. That count covers sites where:

• GA4 wasn't collecting

• no key event existed for the enquiry

• the event fired on page load

• the event fired twice

• the event counted the wrong action

When the event is missing or doubled, look at the mechanics before you blame the page. None of these is automatically a fault. So compare what's built with the success condition you wrote down.

Our guide to GA4 for small business covers the setup itself. Respect consent choices and ad blockers rather than hunting for a workaround that collects data without permission. Our guide to Australian privacy law changes for websites covers what applies to your site.

Keep names, emails, phone numbers and free text messages out of URLs, page titles and event parameters too. Send Google nothing it could use or recognise as personally identifiable. That's what Google's Best practices to avoid sending Personally Identifiable Information (PII) page asks.

Calls need the same treatment, because a site counting only forms can look dead while the phone rings. In Elev8d's service accounts, calls are commonly 30% to 50% of leads, more for emergency trades. That's our observation, not a market statistic. It covers the 45 service accounts we managed from May 2024 to September 2026, all with call tracking. Our guide to call tracking for Google Ads explains how calls get counted.

• Suspect it when: staff mention enquiries the reports don't show. Or reported enquiries fell overnight while traffic held steady.

• Check: a labelled test on a phone and a desktop, followed to the person who answers it. Then look for the event in DebugView.

• Change: repair delivery first, then the event. That way each real enquiry is recorded once.

• Measure: a second test, then a month of reported enquiries compared line by line with what staff received.

Use a rate with a stated denominator

Once the outcome is defined and recorded, turn it into a rate you can repeat. For a session rate, divide the sessions that produced the defined action by the eligible measured sessions. Then multiply by 100. User rates, form completion rates and lead to sale rates answer different questions. So label each one and don't mix them.

Show the working with every rate:

• numerator

• denominator

• dates

• pages

• channel

• device

Compare like with like and allow for a long sales cycle. A rate can move while nothing real changes. Low counts, a season, a change in traffic mix or a tracking fix can all do it.

There's no universal good website conversion rate. We treat 5% to 6% of clicks becoming enquiries as standard. With enough conversion history and a landing page built for the campaign, 8% to 9% is realistic. Below 3% is worth investigating.

Those rates are for service lead generation from high intent paid traffic. Our guide to what a good conversion rate is explains them. They come from our account experience, not a published benchmark.

No percentage alone makes a page high converting. A page converts well when suitable people complete the action it exists for. The cost also has to be one the business can carry.

Here is one example cohort, followed from visit to booking.

Of 1,000 observed people, 40 enquire, 20 qualify and 10 book within the same follow up window. Each stage is a subset of the one before. Half the enquirers qualified (20 of 40), but only 2% of the visitors did (20 of 1,000).

Both rates are true but answer different questions. The 4% is arithmetic, not a target. And never stitch unrelated analytics sessions, CRM leads and bookings together as if they were one matched group.

For paid traffic, check which conversion actions the ad platform optimises toward. Say the same enquiry is imported from GA4 and measured by a native Google Ads tag. That's one business outcome, not two. Our guide to Google Ads conversion tracking covers that setup.

How does a website work as a sales system?

Treat the website as one sales system, not a brochure. A suitable person arrives and finds a page that answers their buying questions. They see evidence they can believe. They make a request without a fight, then get a useful reply from a real person. When one link fails, the whole system underperforms, however good the rest looks.

A small business marketing funnel is simply a way of organising that progress. It runs from awareness and evaluation to an enquiry and a commercial outcome.

The flow is a model, not a measurement. Real journeys loop. People leave and come back, or compare you with two competitors. They switch from phone to laptop or ring the number on your van. The model still shows you where to look when one stage goes quiet.

Give each page its job

Different pages do different jobs. Judging them all on one enquiry rate hides what each one is for:

• An article answers its question and offers a relevant route when the reader needs a service.

• A homepage helps people understand the business and choose a path.

• A service page answers buying questions.

• A campaign landing page continues one specific promise from an ad or email. Our guide to building a landing page that converts covers it properly.

• An online store has its own friction, from product detail to checkout.

Each service also needs a clear path. A visitor shouldn't have to infer that a broad services page covers their exact need. Picture a business that hires out equipment and also does landscaping. Someone hiring a mini excavator for the weekend and someone planning a new garden ask completely different questions. So the business needs separate, clearly labelled routes.

A bookkeeper who looks after sole traders and growing companies has the same problem in a quieter form. Use meaningful labels and genuinely distinct content. Don't manufacture duplicate location or service pages just to create more routes.

Finding where suitable visitors drop out of this system, then fixing it, is what conversion rate optimisation involves. The table maps the usual symptoms to the first thing to check.

Are the right people reaching the right pages?

Compare each page's purpose with where its visitors come from:

• the source

• the campaign promise

• the searches that bring them

• their location

• whether you serve them

Then sort the traffic. Some visits are a poor fit and never could have converted. Some are suitable people doing early research. Some are ready to enquire now.

Because each group needs a different remedy, diagnose before you fix. Informational traffic isn't worthless, and some of it converts later. An urgent repair search and a DIY guide need different next steps, not the same hard sell.

An example case: a Melbourne appliance repairer gets steady traffic from Geelong and Ballarat, outside its service area. That calls for clearer targeting and eligibility, such as naming the suburbs it covers. Meanwhile a customer inside the service area lands on a generic appliance services page and leaves. This one calls for a more relevant destination. City, search and device alone can't tell you which problem you have.

For campaign visits, check the targeting and whether the landing page continues the ad's promise. Our guide to Google Ads clicks but no leads works through that diagnosis. For visits from organic content, keep the helpful answer and add a relevant service route. Don't turn the article into a sales pitch. Our guide to high intent vs low intent keywords explains how to read what a search wants.

Don't read too much into engagement either. Traffic volume, a ranking or a strong engagement rate doesn't prove anyone meant to buy. Nor does a high bounce rate prove they rejected you. A visitor who found your number on a laptop and rang it can still count as a bounce. Our guide to what bounce rate means in GA4 explains how the figure is calculated.

• Suspect it when: visits grow while enquiries don't. Or the enquiries you get are out of area or for services you don't offer.

• Check: search terms, campaign targeting, landing pages by source and where your enquiries come from.

• Change: tighten targeting and send each audience to the page that matches its promise. Give research content a relevant next step.

• Measure: suitable enquiries by source and landing page, not total visits.

Is the offer clear and worth acting on?

Clarity and appeal are different problems. Clarity means a visitor can tell what the service is, who it's for and where you work. They can also tell what's included, what makes you different and what to do next.

Appeal is whether that service solves their problem on terms they'll accept. Strong copy can fix a clarity problem. It can't create capacity, suitability, a competitive offer or evidence that doesn't exist.

Clarity comes from practical buying answers:

• scope

• exclusions

• pricing guidance or how a quote is worked out

• realistic timing

• availability

• eligibility

• the process

• what happens after someone gets in touch

Your inbox is the best source. The questions people ask in their enquiries are the questions the page should have answered. Don't invent a discount, a guarantee, urgency or a free offer to make the page feel stronger.

Try a stranger test. Show a priority page for about 10 seconds to someone who doesn't know the business. Then ask what you offer and how they'd take the next step.

It's a quick comprehension check, not a proven attention span. And it's no substitute for analytics and real user testing. Our small business website guide covers the planning that makes those answers obvious.

On a homepage the job is routing. Our guide to writing a homepage that converts covers that structure. On a service page the job is the decision. Here is an example anatomy of one with a clear enquiry path.

It's a structure, not a ready made offer. So fill each part with real facts about your business. Our guide to service page structure and copy takes each part further.

One clear priority doesn't mean stripping out useful navigation. Nor does it mean forcing everyone to call or putting the same button after every paragraph. A long page can repeat a relevant action where people decide. And a homepage can route to several services.

If the words are the problem, swap vague claims for specifics. Our guide to writing website copy that doesn't sound like a robot shows how.

Sometimes the page is perfectly clear and people still don't bite. That's an offer problem and a business decision. It could be the price, the minimum job, the service area, the lead time or the terms. Be honest about which problem you have before you spend money on the page.

• Suspect it when: relevant visitors arrive and leave. Or enquirers keep asking what the page should already have told them.

• Check: the stranger test and the questions in your last 20 enquiries. Then look at the alternatives a customer would weigh you against.

• Change: answer the buying questions on the page. If the terms are the problem, change the offer itself.

• Measure: suitable enquiries from that page, and fewer basic questions in the ones that arrive.

Does the page give people enough reason to trust you?

Proof works when it answers the visitor's real doubt. Can you do this kind of job? Will you turn up? Is the price fair for what's included?

Answer those questions with project details, approved reviews, credentials, real people, service boundaries and a clear process. Put each one beside the question it answers. A stock photo of a smiling tradie isn't evidence of completed work. It isn't automatically a usability problem either. It just proves nothing.

The same audits found about 3 in 5 sites with no proof near the main call to action. On those sites, none of these sat in the first two mobile screens or beside the main button:

• a review

• a testimonial

• a rating with its source and count

• a licence

• an accreditation

• a named customer example

The count includes about 1 in 4 sites with no reviews or testimonials anywhere on the site.

The sites with reviews kept elsewhere have the easiest fix. The reviews already exist. They just sit nowhere near the decision. Vague claims are the other gap to close. The cure is a plan for evidence, not a better adjective.

Real case studies need dates, scope, a baseline, the result, the limitations and the customer's permission. Our guide to E-E-A-T for small business covers recording that evidence and showing trust on the page.

Display reviews accurately. The ACCC's Online reviews for product and services guidance makes two points. It's against the law for a business to create fake or misleading reviews. A business may also mislead consumers if it suppresses or edits negative reviews, or removes genuine ones.

Our guide to how to get more Google reviews legally covers the rest of the rules. It also shows how to collect reviews the right way.

Skip the universal claims about buyer psychology while you're at it. People weigh price, speed, quality and risk differently. So explain your real prices and service conditions and let them choose.

• Suspect it when: enquirers ask whether you've done this kind of job before. Or people only get in touch after a referral.

• Check: what sits in the first two mobile screens and beside the main call to action. Does it answer the visitor's real doubt?

• Change: place relevant, approved evidence beside the decision it supports.

• Measure: suitable enquiries from the page, and whether enquirers mention the evidence when they get in touch.

Can visitors complete the next step?

Test the real task from start to finish, the way a customer would. For a service business, that's usually an enquiry. The steps are finding the right service, deciding, making the request and knowing it went through. Do it on a phone and on a desktop.

Check your own GA4 device report to see how your visitors split between phones and desktops. Start the test on the device most of them use.

Here's an example task review: the friction a tester might hit at each stage and its fix. Not every page has all four problems.

Then repeat the task the hard way. Use the keyboard only, leave the cookie banner open and make a deliberate mistake in the form. Look for visible labels, readable text and a visible focus outline. Check that error messages explain the problem and a confirmation says what happens next.

A chat widget or sticky bar that covers the button is a broken task, not a design detail. Our guide to website accessibility for Australian businesses covers the standards behind these checks.

The first screen matters as well. Take the first mobile screen of the homepage or main service page. In the same audits, about 2 in 3 sites showed no clear call to action there. That means no button, tap to call link or form start.

Forms, calls, bookings and checkouts

For forms, ask for what you need to respond to or route the enquiry. Explain any required question and leave the rest for later. There's no universal field count.

The same audits looked at forms too. About 1 in 2 sites with an enquiry form required at least one unused field. Unused here means the business didn't use it for its first response.

Give useful error messages and keep answers when something fails. Allow uploads where a photo helps and offer another way to get in touch. Our guide to website enquiry form design covers the field decisions, spam and privacy.

For phone led services, check that:

• the number is visible

• the tap to call link works

• the hours are clear

• any after hours option is one the business actually supports

Count answered, suitable calls separately from taps on the link. Our guide to getting more phone calls from your website covers placement and measurement.

For bookings, test the handoff to any external scheduler. Count a request separately from a confirmed appointment. For an online store, check account requirements, delivery costs that appear late and payment failures. And check how people recover from a checkout that fails. Ecommerce conversion rate optimisation has its own checks beyond these.

A new design can help when the platform or layout throws up the same barrier again and again. It isn't the default answer to low enquiries. When the problem is local and you can verify it, repair it. And ignore anyone who says a particular button word, colour or position always lifts conversions. Nothing always does.

• Suspect it when: people start the form or tap the number and no enquiry arrives. Or customers say they tried and gave up.

• Check: the full task on a phone and a desktop. Then repeat it with a keyboard, with the banner open and with a deliberate error.

• Change: fix the step where the task broke, then run the whole task again.

• Measure: completed requests and answered calls, counted where the task happens.

Is speed making that task harder?

Speed matters when it makes the task harder. So measure it on the page and device that carry suitable visitors.

Open PageSpeed Insights for a priority page. When available, the top section shows field data from real Chrome users over the previous 28 days. The Lighthouse lab run below it helps you find causes. Google's About PageSpeed Insights explains the difference.

Check whether the field data covers the page or the whole origin before you draw conclusions. Missing field data doesn't mean the page passes or fails.

Field results are judged on three Core Web Vitals. LCP is how long the main content takes to appear. INP is how quickly the page responds to taps, clicks and key presses. CLS is how much the layout jumps around. Core Web Vitals explained covers each metric, its thresholds and how to read field and lab data.

Google's Lighthouse bands put 0 to 49 as poor and 50 to 89 as needs improvement. Scores of 90 to 100 count as good. The score is lab data.

The field Core Web Vitals are what Google's ranking systems use. In our view, a mobile score in the 70s is usually fine for a small business. That holds when the field Core Web Vitals pass.

Interaction to Next Paint became a Core Web Vital in March 2024. About 1 in 2 sites we've audited since then failed at least one Core Web Vital on mobile. Across all the audits, about 2 in 5 sites had a mobile LCP slower than 2.5 seconds.

Investigate what the diagnostics show:

• heavy images

• content that arrives late

• scripts

• a slow server response

• a layout that moves as it loads

Don't assume cheap hosting is the culprit. Our guide to why your website is slow works through the causes and fixes.

Keep speed in proportion, too. Google's Understanding page experience in Google Search results says good Core Web Vitals results don't guarantee top rankings. Nor does it tell you how many enquiries your business loses for each second of delay.

Elev8d ties the final payment on some builds to achieving green Core Web Vitals. Speed still sits beside the other checks in this guide, not above them.

• Suspect it when: field data fails on a page that carries suitable visitors. Or customers mention waiting for pages to load.

• Check: PageSpeed Insights for that URL on mobile and desktop, field data first and lab diagnostics second.

• Change: the specific causes the diagnostics point to, page by page.

• Measure: field data again after the next 28 day collection window, alongside enquiries from that page.

What happens after someone enquires?

Winning an enquiry and turning it into a customer are two different jobs. If relevant leads arrive and sit unanswered, the failure is in routing, staffing, availability or follow up. No page fix will touch it. Track those stages separately before anyone blames the website.

Decide who owns each new lead and who covers when they're away. Set a response target the business can meet. Know the difference between an automatic acknowledgement and a useful human reply.

Publish clear business hours. Route missed calls and after hours requests somewhere they'll be seen. Whoever replies should know the service the page promised, the qualifying questions and the next step.

It's a conceptual workflow. Add a response target to it only when the business has set one it can keep. Reply promptly, but don't chase a magic number of minutes.

Record every enquiry in one place, with the same fields each time.

Keep the reasons separate. Each of these needs a different fix:

• unanswered calls

• spam

• duplicates

• out of area requests

• price mismatch

• no capacity

Collect only what you need to respond, and keep the records secure.

Follow up the way the person asked: by phone if they rang, by email if they wrote. Do it at the time they said suits them. A receipt email isn't a sales response.

Check that quotes explain the work clearly and that the team can deliver what the page offers. More leads can expose a capacity problem. Better qualification can cut volume while lifting profitable work. Both count as progress if you measure the right thing.

• Suspect it when: enquiries arrive but few become quotes, or customers say they never heard back.

• Check: the lead log, for time to first useful reply, who replied and why each lead was lost.

• Change: set ownership, cover and a response target the business can meet.

• Measure: time to first useful reply, then quotes and bookings per suitable enquiry.

What should you fix first and how will you judge it?

Fix what's broken before you polish what works. Repair a verified broken enquiry path, missing delivery or misleading offer before any cosmetic change or experiment. Then restore the measurement you need to judge outcomes. Then tackle the high impact problems your task tests and records support. Test the uncertain improvements last.

Investigate in the same order across the whole site. Start with measurement, then traffic, then whether people can understand and complete the task. Finish with lead quality and handling.

Our audits can't tell you which cause will turn out to be yours. Nor can a small set of observations set a universal fix order. So choose your order by weighing:

• the likely impact on suitable customers

• your confidence in the evidence

• the effort

• the risk

• what depends on what

Skip numeric priority scores. They make a judgement call look like maths.

Start with a 15 minute scan

If you want a starting point today, this scan gets you going in about 15 minutes. It produces observations to test, not a diagnosis or a score.

Write down what you saw, not a verdict. For the full version, our DIY website audit checklist scores 15 checks in about 30 minutes.

Three examples, not client results, show how the evidence points to the change and the measure.

Before you make any change, write it down.

Include the effect on lead handling and business outcomes, not only the click rate on a button.

Judge changes on suitable outcomes

An A/B test can help when you have enough traffic and the discipline to run it cleanly. Without the volume, use task observation, customer interviews and carefully qualified before and after comparisons. Neither method justifies calling a winner early. Inconclusive is a legitimate result.

When should you get help with conversion optimisation?

Get help when the problem spans the whole journey. The same goes when the fixes need skills or access you don't have. It also makes sense when nobody has time to run the tests and keep the records. Get it sooner if your tracking can't be trusted, because every other decision depends on it.

For businesses that want to try the service, Elev8d can build a demo homepage before they commit. We base it on their brand and services. A demo shows one possible approach. It isn't evidence that the finished website will lift enquiries.

A scoped engagement should give you:

• a reliable baseline

• a review of the priority journeys and how leads are really handled

• hypotheses backed by evidence

• an ordered change plan

• clear responsibility for implementation

• reporting on outcomes

Ask in writing whether development, tracking repair and testing are included. Be wary of anyone who promises a lift before they've seen your data. Our conversion optimisation work starts with that measurement, not a redesign.

What we recommend at Elev8d

FAQ

Next steps

Written by Ajay K.Co-founder, Elev8d

Ajay is the co-founder of Elev8d and writes most of what you will read here. Psychology grad turned marketer, which is a fancy way of saying he cares about why people click, not just where things rank. Every guide is written in plain English for Australian businesses, tested on real campaigns before it earns a spot on this site.