Conversion Audit: 15 Checks That Find Your Lost Enquiries

2. What a Conversion Audit Actually Is

Quick Answer: A conversion audit is a structured inspection of the enquiry path, run against a real device and a real submission. The output is a ranked list of things that are broken, ordered by how many people each one affects, rather than a redesign proposal or a traffic report.

The word “audit” is doing real work here. An audit has a fixed scope, a fixed checklist, and a result someone else can reproduce. That is the difference between an audit and a designer telling you the hero image feels tired.

What the audit produces, and what it does not:

• Produces a defect list, not a wish list. Every item is something a visitor actually hit: a form that failed, a button nobody could tap, a page that took nine seconds.
• Produces a ranking, not a to-do pile. Each defect carries how many visitors it affects and how long the fix takes.
• Does not decide the fix. Naming the leak and choosing the patch are separate jobs.
• Does not require a rebuild. On most sites the top three findings are configuration and copy, not architecture.

It sits one step before the optimisation loop. The audit tells you what is wrong now; the six-step CRO process is what you run afterwards, and conversion rate optimisation services are what that looks like month to month.

3. Checks 1–5: Can Visitors Even Reach the Enquiry?

Quick Answer: The first five checks ask whether a motivated visitor can physically get to an enquiry point. Run them on a phone, on mobile data, from a cold start. Most failures here are invisible on a designer’s desktop, because the desktop layout hides them.

Malaysia is a phone-first market, with internet penetration at 98.0% and 35.4 million users at the end of 2025, so the phone pass is the one that matters:

1. Is an enquiry point visible on a phone without scrolling? If the first screen shows only a logo, a menu and half a headline, the enquiry is already several seconds away.
2. Does the main content render in about 2.5 seconds on mobile data? That is Google’s threshold for a good Largest Contentful Paint, measured at the 75th percentile of loads. Past four seconds it counts as poor.
3. Is there one primary action, or three competing ones? A page offering a call, a form, a brochure and a newsletter offers no clear next step at all.
4. Does every one of your top five landing pages have its own enquiry point? Traffic rarely lands on the page you optimised.
5. Does the mobile menu open, and is Contact reachable in one tap? Broken hamburger menus survive for months because nobody on the team uses the menu.

Checks two and five hold the cheap wins. If the page is slow, start with what a slow website costs and how to fix it; if the phone experience is weak, why phone visitors never enquire covers the pattern. Twelve landing page fixes that lift leads deals with check four.

4. Checks 6–10: The Form and the Phone

Quick Answer: Checks six to ten test the enquiry mechanism itself: field count, whether the form submits on a phone, whether the notification email arrives, and whether the phone and WhatsApp links are really tappable. This is where the audit finds breakage rather than friction.

The second pass ends with you submitting a genuine enquiry and watching where it lands:

6. How many fields does the form ask for, and how many does sales actually use? Anything nobody reads is friction with no return.
7. Does the form submit on a phone, on mobile data, first attempt? Not on office wifi, and not on desktop. Submit it properly and time the response.
8. Does the notification email arrive in a monitored inbox? Check spam, and check the address still belongs to someone who works there.
9. Are the phone number and WhatsApp real links? A number typed as plain text is a number nobody dials.
10. Is anything gating the enquiry? A captcha, a required “how did you hear about us”, a login wall. Each gate costs enquiries, and most were added for internal convenience.

Check six is the highest-yield item on the list, so settle whether multi-step forms really get you more enquiries before redesigning anything, and let field-level drop-off data tell you which question loses people. Check eight fails more often than owners expect, because forms that silently stop sending emails can run for months. For check nine, weigh the sticky call button most SMEs skip against live chat versus a WhatsApp button.

5. Checks 11–15: Trust, Proof and the Aftermath

Quick Answer: The last five checks cover why a visitor who can enquire still decides not to, and what happens in the minutes after they do. An audit that stops at the submit button misses every enquiry lost to a slow reply.

11. Is there real proof within one screen of the enquiry point? A named client, a specific result, a licence number. Stock badges do not count.
12. Does the page say what happens next? Who replies, how fast, on which channel. Silence here is the quietest reason people close the tab.
13. Is there a dedicated thank-you page? A real URL, not an inline message that disappears, which also gives you a clean conversion event.
14. Is every enquiry action counted in one place? Form submits, tap-to-call and WhatsApp taps. If two of the three go uncounted, every number after this is guesswork.
15. How long does a real enquiry wait for a real reply? Measure it from the test enquiry you submitted in check seven.

Check thirteen is the cheapest of the fifteen and the most often skipped, as the free conversion win most sites miss explains. Check fourteen decides whether any of this is measurable, so pair it with a proper GA4 conversion tracking setup. Check fifteen sits outside the website: replying within five minutes wins the sale more often than any button colour, provided the enquiries are qualified in the first place. Only once those fourteen pass is it worth asking whether exit intent popups still work in Malaysia.

6. Where the Lost Enquiries Actually Hide

Quick Answer: Across ZenWeb conversion audits, form design and mobile call access together account for roughly half of every recoverable enquiry. Trust and proof, the thing owners usually want to spend on first, accounts for about a seventh.

The order is stable across industries and rarely matches what owners expect. Almost nobody asks about field count; nearly everyone asks about the design. That gap is the reason to audit before signing a quote, and why benchmark conversion rates for Malaysian SMEs belong beside this table.

7. Why the Same Check Fails Differently on Mobile

Quick Answer: Per 1,000 visitors who reach the enquiry point, mobile loses roughly twice as many as desktop, but the gap is not spread evenly. Call access loses almost seven times more on mobile, while trust and delivery failures hurt both devices equally.

The bottom two rows are the useful ones. Trust and delivery failures cost the same on both devices, so they are content problems. Everything above them is layout, and layout is where heatmap analysis shows what visitors never tell you.

8. What Each Fix Costs and What It Returns

Quick Answer: On a services site with around 4,000 monthly visitors, the four cheapest fixes from the audit take about two days of build time and return more enquiries than a five-day page rebuild. Effort and return are close to inversely related.

The bottom row is the one worth arguing about. A five-day rebuild returns fewer enquiries than a half-day sticky bar, and eats the budget that would have funded the other five rows. A healthy enquiry rate with poor sales is a targeting problem instead, which is the comparison in what inbound and outbound leads cost to close.

9. How Long Before the Numbers Move

Quick Answer: Nothing moves in the first fortnight after the audit, because that fortnight is spent building. Half the eventual lift lands by week four and about nine-tenths by week eight. Judging the work before week four is judging an empty window.

The flat fortnight is what gets the work cancelled. It is not a failure signal; it is the build. Reading the curve honestly needs the counting fixed before week zero, which is check fourteen doing its job.

10. Running the 15 Checks Yourself in One Afternoon

Quick Answer: A first-pass conversion audit takes one person about three hours: pull the baseline, run the phone pass, run the desktop pass, submit a live test enquiry, then score and rank. No tool licence is needed for the first three findings.

Do it in this order, and write each result down as pass or fail before arguing about fixes:

1. Pull four weeks of baseline numbers. Visitors, enquiries and the split by device. If the split is unavailable, that is finding number one.
2. Run the phone pass. A real handset, mobile data, cold cache. Work checks one to five, noting where you had to think.
3. Run the desktop pass. Same checks, compared against the phone. Differences are findings in themselves.
4. Submit a live test enquiry through every channel. Form, tap-to-call and WhatsApp. Follow each to a human and time the reply.
5. Score the fifteen checks and rank the failures. Order by visitors affected per build day, using the device split from step one.
6. Ship the cheapest three fixes first, one at a time. That keeps the result readable when you review at week eight.

Free tooling gets you a long way, and a comparison of the heatmap tools worth using shows which free tier is enough. If too few people reach the enquiry point at all, the answer is upstream: lead magnets a Malaysian SME can build in a week give visitors a reason to identify themselves earlier.

11. Conclusion

Quick Answer: Run the fifteen checks on a phone, submit a real enquiry, rank the failures by visitors affected per build day, ship the cheapest three first, then review at week eight. That sequence finds most lost enquiries without a rebuild.

The exercise is deliberately unglamorous, and most of the list it produces is small. But a half-day fix that returns six enquiries a month beats almost anything a redesign brief promises, and you know within eight weeks whether it worked.

If you would rather not run the checks yourself, ZenWeb handles the audit, the build and the measurement inside our web design services, so no finding waits on a second supplier.

12. Frequently Asked Questions

1. What is a conversion audit?

A conversion audit is a structured inspection of the path a visitor takes to enquire, run against a real device and a real submitted enquiry. It produces a ranked list of defects, not a redesign proposal. The fifteen checks cover reach, the form and phone, and what happens after submit.

2. How long does a conversion audit take?

A first pass takes one person about three hours, on top of four weeks of baseline data. A full audit with recordings and field-level drop-off runs about a week. Fixes take one to three weeks to build, and the result reads by week eight.

3. How is a conversion audit different from an SEO audit?

An SEO audit asks why people are not arriving. This one asks why the people who did arrive are not enquiring. They share a few technical checks, page speed in particular, but the outputs are different lists and the fixes sit with different people.

4. Do I need a lot of traffic for a conversion audit?

No. The fifteen checks work at any traffic level because most findings are defects you can see, not differences you must test for. Traffic only decides how you read the result: below roughly 150 enquiries a month, use before-and-after windows instead of split tests.

5. How often should I run a conversion audit?

Once every six to twelve months for a stable site, and always after a redesign, a platform migration or a change of form plugin. Those events break tracking and delivery more often than anything else, and the breakage is silent.