Sometimes, but clicks without leads do not prove the landing page is at fault. The page gets blamed first because it's the part everyone can see. Check the reporting, destination, traffic and offer before you test the page and the enquiry route. Fix every break the evidence shows, because more than one thing can fail at once.
Check whether leads or just reporting are missing
Start with what the business is actually claiming. No form submissions, no answered calls, no bookings, no qualified leads and no sales are five different problems. The first job is to find out which one you have.
Reconcile the records
Pick one date range and read it across every system that could hold an enquiry:
The Ads account.
The analytics property, if there is one.
The form tool.
The booking software.
The CRM or inbox.
A campaign can have leads that never reached Ads. A broken tag or a booking tool on another domain can cause that.
That gap isn't rare. Our audit sample is 180 Google Ads account audits from August 2021 to September 2026. We checked the 90 days before each audit. In about 1 in 4 accounts, no primary conversion action had recorded a genuine enquiry. But the business had received enquiries in that time.
Ads can also show conversions that are nothing more than button clicks. Neither picture on its own says anything about the page.
Review the conversion action
Open the conversion action the campaign is judged on. Check where the action comes from, how it counts and whether it's primary or secondary.
Primary actions steer bidding when their goal is used for bidding. Secondary actions are only reported, unless they sit in a custom goal. A custom goal bids on them regardless, as Google's page on primary and secondary conversion actions explains.
Do not mark every micro action primary so the column looks busy. One enquiry recorded by the site tag and again by a CRM import is still one lead.
Our guide to Google Ads conversion tracking covers the setup and the repairs. For this diagnosis, the only question is whether the number means what the owner thinks it means.
Test the success event
Run the real enquiry journey through Google Tag Assistant's conversion verification. A passed test can leave the action showing "No recent conversions" until genuine conversions arrive. After a test, that status does not by itself mean the tag failed.
Make sure the success event fires only after a real successful action, not on the button press. A debug visit verifies firing. But that visit isn't a conversion attributed to an ad click.
Nobody should click their own live ads or manufacture enquiries to see a number move. Test the consent states visitors actually see as well. Don't force a tag to fire by overriding anyone's choice.
Allow for lag
Ads files each conversion under the date of the search that led to it. That's not the day the enquiry came in, according to Google's page on conversion lag. So the latest days look worse than they'll end up. Cost per acquisition reads high and return reads low until the late conversions catch up.
Time zones, attribution settings and each system's own definitions widen the gap further. A week that looks empty on Monday can be respectable by the following Friday. Third party booking tools and cross domain measurement fail in ways the Ads interface cannot show. So they need whoever manages the tracking.
Confirm clicks reach the intended page
A click is a billing event. It is not proof that a usable page loaded on the device the person was holding.
Read the Landing pages report
Google's Landing pages report sits under Campaigns, Insights and reports, Landing pages. It lists the pages ads sent traffic to, including sitelink URLs for Search. The View expanded landing pages option shows the URLs people actually reached under each one.
That view is where a tracking template, a redirect or URL expansion reveals itself. If the campaign uses final URL expansion, read this report before assuming the configured URL is the destination.
One caution: Google's help page for the report still describes a Mobile friendly click rate column. It also points readers to the Mobile Friendly Test. Google retired the Mobile Friendly Test in December 2023. So don't build a step on that column.
Walk the route yourself
Open the destination directly, or through an appropriate preview, on a phone and on a desktop. Then trace the route:
Follow every redirect to the page it ends on.
Watch for a page that is unavailable or shows the wrong location.
Confirm the query parameters the measurement depends on survive the redirects.
Follow the handoff to any external booking or payment domain. Keep tracing it to the point where the measurement either survives or quietly stops.
Expect clicks and sessions to differ
Ads clicks and analytics sessions rarely match one for one. Repeat clicks, interactions that aren't page visits, blocked or unconsented measurement and different definitions all open a gap.
A gap is a prompt to investigate, not a universal percentage of wasted spend. Before anyone claims visitors never arrived, check the loading and event evidence that would show whether they did.
Check whether the traffic fits the offer
Say the leads really are missing and the page really does load. Then give the ad side the same scrutiny the page is about to get. The wrong traffic produces clicks without leads just as reliably as a broken form.
Check | Evidence to inspect | Action when the evidence supports it |
|---|---|---|
Search intent | Visible queries, their matched targeting and the qualified outcomes behind them | Exclude unsuitable demand or separate a valid offer, and keep research and price questions |
Match and negative controls | Whether expansion brought unsuitable queries or exclusions suppress useful ones | Adjust the relevant control and scope, without blaming broad match or prescribing exact only |
Geography | Location options, reported locations and the real service or job address | Correct the service footprint, since presence is not a location detector and city words are not geo settings |
Campaign and network mix | Search, partners and other inventory compared by meaningful outcomes | Investigate the segment causing the problem instead of judging all traffic by a blended rate |
Timing and capacity | Ad promise against operating hours, call answering, availability and fulfilment | Match the promise to what can be delivered, or offer an accurate after hours path |
Read the queries against outcomes
Google's search terms report shows the searches that triggered your ads. Each search needs enough activity to be shown. Low volume searches are left out for privacy. So treat the report as a sample of the evidence, not a list of everything you paid for. A high click through rate can still mean the wrong promise attracted the wrong person.
Read the queries beside the matched keyword and the qualified outcome. Do not exclude a valuable search after one click that did not convert. Our guides to Google Ads keyword match types and Google Ads negative keywords cover how to set them. Here the only question is whether the traffic is the fault.
Check the actual service area
Compare the campaign's location options with the report's locations and with where the jobs or customers actually are.
Presence targeting is not a perfect reading of where someone stands. A suburb in a keyword does not fence the campaign. A business whose ads run across a state it cannot serve will get clicks and no leads. The reasons have nothing to do with the page. Our guide to Google Ads location targeting covers the settings.
Split the traffic before judging it
One conversion rate can blend Search and search partners, phones and desktops, three services and two states. That blend hides the segment that's actually failing.
Segment by each of these before attributing a whole account change to one page:
service intent
device
browser
network
geography
actual destination
If the campaign isn't serving at all, the page isn't even in play. The same goes for a billing hold or a policy restriction. Our guide to why Google Ads is not working covers campaigns with no delivery and account wide faults.
Campaign service and available offer: [what is advertised and what can actually be delivered]
Search intent and visible query evidence: [queries, matched keywords and outcomes]
Match and negative controls: [match types, negatives and their scope]
Customer or job location evidence: [where enquiries and jobs actually come from]
Device, destination and network: [the segments compared and the destination each reached]
Qualified outcome and reporting delay: [the outcome used and the lag allowed]
Proposed change and reason: [one change and the evidence for it]
Match the ad promise to a credible offer
Here is an example case. A Melbourne electrician who only takes planned weekday bookings runs an ad that promises 24 hour emergency response. The clicks arrive at night. The phone rings out or the form asks for a preferred weekday. Nobody becomes a customer.
The page is not the problem. The promise is.
The ad promises 24 hour emergency service but the business only takes planned weekday bookings
The ad and targeting describe the planned service and accurate booking availability
Fix the promise before the page
That is an ad side diagnosis with an ad side fix. Rewriting the page to repeat an unavailable promise would make the mismatch worse. A dedicated emergency landing page would be an expensive way to disappoint people faster.
Contrast it with a different case. Relevant traffic for a service the business does provide lands on a page about a different service. That one is a destination problem. Our guide to a Google Ads landing page vs the homepage walks through the choice. The options are a homepage, a service page and a campaign page.
Check the page answers the five questions
Whichever page the traffic lands on, a visitor needs to understand:
The service.
Whether they are eligible or inside the area covered.
The likely cost basis.
The evidence that the business can do the work.
The next action.
Proof has to be genuine and relevant to the claim beside it. Keep the information and the navigation the decision needs, rather than stripping a page to a form.
Sometimes the diagnosis does point at the page. Then our guide to building a landing page that converts covers the build section by section. That build is not the default prescription.
Check the offer can be delivered
Pricing, service coverage, opening hours, appointment capacity and response speed can stop useful leads arriving. That can happen even when the page and the tracking are perfect. A quote that takes four days produces clicks without leads. So does a phone staffed until 3pm for an ad that runs until 9pm. A price the market will not pay does the same.
No amount of page work fixes them. Check the operations before the layout.
Find friction in the enquiry journey
When the traffic fits and the promise is honest, the page finally earns a proper look. Test it the way a visitor meets it, not the way the designer sees it.
Bottom line Sometimes. Rule out the certain faults first, then test one page change
Run the form on the devices that matter
Use the account's own device and browser split to decide what to test. Don't rely on an assumption about who searches on a phone.
On each device and browser that carries real traffic:
Fill the form in by touch, then again using only a keyboard.
Read every label.
Enter something wrong, and see whether the validation says what is wrong and how to fix it.
Submit and confirm what the visitor sees.
Try a double tap on the button.
Watch for a CAPTCHA or chat widget that fails. Check that no banner or sticky bar sits over the fields.
The W3C's form validation guidance is the standard here:
Mark required fields in the label as well as the code.
Accept the different ways people type the same data, such as a phone number with or without spaces.
Validate on the server as well, because client side checks can be bypassed.
Shorten or clarify fields when they obstruct the task. But keep the ones that qualify the enquiry. More raw submissions of poorer quality is a worse result wearing a better conversion rate. Our guide to enquiry form design best practices has the detail.
Record every fault in the same format so anyone can reproduce and retest it after the fix:
URL: [the exact destination tested, after redirects]
Device: [make, model and screen size]
Browser: [browser and version]
Consent state: [accepted, declined or not shown]
Steps: [what was done, in order]
Expected result: [what should have happened]
Actual result: [what happened, with the exact message or screen]
Retest: [date, who retested and the outcome]
Reproduce loading problems with field data first
Treat speed as a measured usability question. Use the field data for the page where it exists. Field data describes real visitors on real connections.
Use lab tools to reproduce a fault once the field data points at one. Google's Core Web Vitals guidance sets the thresholds. It also says lab measurement is not a substitute for the field.
A good Lighthouse score does not prove the page converts. A poor one does not prove it cannot. Measure the page you have rather than chasing a fixed image size or load time. Our guides to why a website is slow and Core Web Vitals explained cover the repair work.
Check the phone and booking routes
The form is one of three routes. Call the number from the page during the hours the ads run. Check the right line picks up. Ring again after hours and hear what a caller hears.
Then open the booking tool from the page on a phone. Take a booking through to the confirmation screen.
Each route has its own way of failing. A page can pass the form test and still lose every caller.
Trace enquiries through to the person answering
A success screen is the halfway point. The rest of the journey is invisible to the visitor and often to the business.
Complete an authorised, clearly labelled test enquiry and follow it:
Is it stored in the form tool?
Does it arrive at the inbox, the booking tool or the CRM?
Is it assigned to a named person?
Does that person respond?
Inspect CRM filters, spam queues, failed integrations, booking confirmations that never send and leads sitting unassigned. Calls need the same trace, because a missed call nobody returns is a lost lead too. Our guide to Google Ads call tracking covers the technical side of calls.
The distinction that matters is between a lost enquiry and a received one that nobody answered. A lost enquiry never reached anyone. The first is a technical fix. The second is a management one.
A labelled test enquiry can show its success message yet never reach anyone who would respond. That happened in about 1 in 5 accounts we have audited. The failure points are dull ones:
a spam filter
an integration that has stopped
a wrong notification address
a queue nobody owns
We'll reconcile your tracking with your inbox. We'll follow a test enquiry to the person who answers it. Then we'll tell you which fault we'd fix first.
Choose a fix and measure what happens
Priority follows the kind of fault. A reproducible failed submission or a broken destination is repaired today, before any cosmetic test. It is certain and it is losing every lead.
A misleading offer or plainly unsuitable traffic is corrected next, because it is also certain. Only once the route and the measurement work do page hypotheses get ranked. Each is ranked by commercial impact, the evidence behind it and the effort to test it.
A ticked checklist is not a diagnosis. Nor is a page fixed because an arbitrary number of boxes are ticked.
We work in this order because of what our own clicks but no leads diagnoses have found. In about 1 in 2 cases, the first confirmed fault was measurement, destination or delivery. It wasn't the page itself.
For an uncertain page improvement, write the plan in the fix record at the end of this section. Do it before you touch anything.
Google's experiments overview covers custom experiments for Search and Display. It says they're typically used to test landing pages, among other things. That makes them the tool when the volume can carry a comparison. Google's own guidance for an undecided result is to let it run at least four to six weeks. So don't call it early.
When the volume can't carry a comparison, fix the evident usability problems. Also record the baseline and report the remaining uncertainty. Do not manufacture a conclusive split test out of 40 clicks.
Do not impose a universal 100 click, 14 day or four week rule. The right decision point depends on the cost of a click and the lag. It also depends on what the business can afford to learn.
The numbers below are a sensitivity example, not an achieved improvement or a forecast. An assumed $3,000 of media at an assumed $5 a click buys 600 clicks. Hold that fixed and let only the enquiry rate move.
Now cost each row, then judge it on the leads worth having:
The row with the cheaper enquiries lost. That counterexample governs every page test. A page change does not promise a move between those rows. These are enquiry costs, not customer costs or return. With no enquiries at all, the cost per lead is undefined, not zero.
Here's a rule of thumb from experience, not a published benchmark. 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.
Quality Score belongs here only as a footnote. Google's page on Quality Score describes it as a diagnostic, not an input in the ad auction. Its landing page experience rating is a clue about relevance, not proof of conversion performance. Lifting the displayed score does not mechanically cut the cost per click. Our guide to Google Ads Quality Score explains why.
Observed symptom and supporting evidence: [what was seen and the records behind it]
Cause confirmed or hypothesis to test: [deterministic fault or the hypothesis]
Change and responsible owner: [the one change and who makes it]
Valuable outcome and qualification rule: [the outcome measured and what counts as qualified]
Traffic and measurement kept comparable: [what was held constant]
Budget ceiling, review date and lag: [allowance, date and days allowed]
Result, uncertainty and next action: [counts, denominators and the decision]
What we recommend at Elev8d
Don't touch the page first. When clicks stop turning into enquiries, we start with a labelled test enquiry and the business's own records. That's because a broken route loses every lead. And a dead tag hides the ones you do get. Both are quick to confirm.
Then we read the traffic. We compare the search terms, locations and networks with what the business can actually deliver. We fix the promise in the ad before anyone rewrites a headline. Only when the route works, the reporting is honest and the traffic fits do we test the page. We make one change at a time, judged on qualified enquiries after the lag.
We'd change that order in two cases. First, nobody set the account up to record real enquiries, or several basic settings are wrong at once. No patching: we rebuild the foundations, following our guide to setting up Google Ads for a small business. And if the first test on a phone shows an obvious fault, we fix it that day. Forms that error and numbers that ring out can't wait for the full diagnosis.
What we won't do is sell a new landing page to fix a promise problem. A rebuild comes after the diagnosis, never instead of it.
FAQ: remaining troubleshooting questions
Is the destination page the fault?
Only when relevant searches land on a page that does not answer them. A brand search can land on the homepage. But a search for one service belongs on that service's own page. A specific offer can justify a campaign page. The Landing pages report shows where each click went.
Can conversion rate optimisation help alongside Google Ads?
Yes, when the evidence points at the page journey. CRO investigates and improves how visitors move from arrival to enquiry. It cannot fix a wrong promise, a broken tag or a lead that lands in a spam folder. Those need their own fixes first.
That diagnosis is the first step in our CRO service.
What should I check first if enquiries stopped suddenly?
Whatever changed just before they stopped. Line that day up against website and plugin updates, plus any new form or booking tool. Do the same with any email or domain change and the account's change history.
Then send one labelled test enquiry from a phone through the ad's actual destination. Follow it to a person. A sudden stop points to a break, not to a page that slowly stopped persuading people.
Should the page be rebuilt from scratch?
Only when the diagnosis says the page is the fault and the fault is structural. A mismatched message, a broken field or a failing widget is a repair, not a rebuild. The campaign side may need more investigation than a business can do itself. We run that investigation for clients, as our Google Ads agency in Melbourne page explains.
Next steps
Reconcile the records for one date range
Ads, analytics, the form tool, the booking software and the inbox. Decide whether the leads are missing or unrecorded before anyone starts editing the page.
Walk the click and the enquiry yourself
From the ad to the destination on a phone. Through every redirect, into the form, the phone line and the booking tool. And out the other side to a person.
Fix the certain faults, then test one hypothesis
Broken route and wrong promise first. Then one page change, written down, measured on qualified enquiries after the lag, with the uncertainty reported.
