What is conversion rate optimization?
Conversion rate optimization, usually shortened to CRO, is the work of getting a larger share of your existing visitors to take the step you want: send an enquiry, tap WhatsApp, call, or ask for a quote. It is not about more traffic.
If 1,000 people visit your clinic's site in a month and 10 book a consultation, moving that to 20 doubles your leads without spending a baht more on getting people there.
For most Bangkok service businesses the honest method is simple. Decide what counts as a conversion, measure it properly with consent, watch where people give up, fix the likely causes, and judge results over weeks rather than days. Split testing has its place, but small sites rarely have the numbers for it, and pretending otherwise leads to decisions made on noise. You will see "conversion-focused design" listed among current web design trends; underneath the label it is mostly this discipline applied steadily.
Decide what a conversion is
A shop has one obvious conversion: a paid order. A service business has several, and they are not equal. Write them down before you touch a tool.
Primary conversions (a real lead)
- Contact or enquiry form submitted
- Quote or booking request completed
- Tap on a WhatsApp or LINE chat button
- Tap on a phone number on mobile
- Click on the email address
Secondary actions (interest, not yet a lead)
- Viewing the pricing or packages page
- Downloading a brochure, menu or rate card
- Opening the map or directions
In Thailand, chat taps matter more than many foreign guides assume. Plenty of customers would rather send a LINE or WhatsApp message than fill in a form, so a site that only counts form submissions can look like it is failing when it is not.
Working out the rate
Conversion rate is conversions divided by sessions (or by users; pick one and stick to it), times 100. Calculate it per page type and per traffic source, not only for the whole site. Your blog readers and people who searched your company name behave very differently.
Be wary of benchmark figures from blogs. They mix industries, countries, definitions and site sizes, so "the average is 3%" tells you little about a dental clinic in Sukhumvit. The only comparison that means much is your own site against its own past.
Measure it properly
You cannot improve a number you are guessing. The free standard is Google Analytics 4 (our GA4 setup guide for small businesses walks through it), where the actions that matter to your business are marked as key events.
A practical set-up
- Form submissions: fire an event when the thank-you message or thank-you page appears, not when the Send button is clicked. Clicks on a button with a validation error are not leads.
- Chat and phone taps: track clicks on links starting with wa.me, line.me and tel:, each as its own event, so you can see which channel people prefer.
- Mark them as key events in GA4 so they appear in reports by page and by source.
- Test every event in GA4's DebugView before trusting the numbers. A broken tag reports zero, and zero looks like a design problem.
Consent comes first
Thailand's Personal Data Protection Act requires a lawful basis for collecting personal data, and analytics cookies are normally set only after the visitor agrees. Your cookie banner should load analytics tags after consent, with a real option to decline, which is part of building a PDPA-compliant website. Two practical consequences:
- Your analytics will under-count, because some visitors say no. That is fine; you are watching trends, not auditing.
- Count leads in a second place too, such as the enquiry emails or chat messages you actually receive, and compare each month.
Find where visitors give up
Once events are tracked, look for the places where interest drains away. Three reports do most of the work.
Landing pages
List the pages where sessions start, with their key event rate. A service page that attracts plenty of visits but almost no enquiries is your first suspect. Compare it with a page that converts: what does the good one show above the fold that the weak one does not?
Exit points
Pages where many visitors leave are not always a problem; people leave after finding your phone number. Worry about exits from pages that sit in the middle of a decision, such as pricing or a service detail page.
Form abandonment
Track when someone starts a form as well as when they finish. If lots start and few finish, the form is the leak, not the page.
Watching real behaviour
Numbers tell you where; recordings and heatmaps suggest why. Microsoft Clarity is a free tool that records anonymised sessions and shows click and scroll maps. Twenty recordings of people on your pricing page often reveal the obvious: a button hidden below a huge photo, a tap target that does not respond on small phones, a menu that covers the page.
These tools collect behavioural data, so under PDPA load them only after consent, mask form fields, and name them in your privacy notice. Pair them with a few conversations with real customers; our guide to UX research for small businesses shows how to run five short interviews without a research budget.
Fixes that usually move the number
No list replaces looking at your own data, but on service sites the same causes come up again and again. Work down this list roughly in order; the early items are cheap and tend to matter most.
- Say what you do and for whom in the first screenA visitor should know within seconds whether they are in the right place. Replace slogans with a plain statement of service, area and audience.
- Show a price signalA starting price or a range filters out poor fits and reassures good ones. Some people never enquire because they fear an unknown cost.
- Put the main action where thumbs areOn mobile, a sticky WhatsApp or call button at the bottom of the screen is often easier to reach than a button in the header.
- Offer the channel people preferForm, WhatsApp, LINE and phone. Let the customer choose instead of forcing one route.
- Add real trust signalsReal photos of your premises and people, clear address and hours, qualifications you can prove. Never invented reviews.
- Cut form fieldsAsk only for what you need to reply. Name, contact and one message box is often enough for a first enquiry.
- Make it fastSlow pages on mobile data lose people before anything loads. Compress images and drop scripts you do not use.
- Answer objections on the pageShort FAQs about price, timing and process close doubts that would otherwise end the visit.
Several fixes touch other disciplines. Clear contrast and labelled buttons overlap with making your website accessible, because a button some visitors cannot read is a button they cannot press. Consistent calls to action across dozens of pages are far easier with design systems for websites, and a defined accent colour reserved for actions belongs in your brand guidelines for the website. Trust also rests on images: stock photos of smiling strangers undo a lot of good copy, so see our business website photography tips before the next shoot.
How do you test changes when traffic is low?
A/B testing shows version A to half your visitors and version B to the rest, then compares results. It is the gold standard when you have the numbers. Most small service sites do not.
Why small samples mislead
Take a hypothetical page with 600 visits a month and 12 enquiries. Split that in two and each version gets around six enquiries. A difference of two or three leads between them is well within ordinary chance; you could run the same page against itself and see gaps that size. To detect a modest real improvement with confidence you typically need hundreds of conversions per version, which a site like this would take years to collect. Any online sample-size calculator will show you the gap for your own figures.
What to do instead
- Make bigger, clearer changes. A rewritten first screen with a price signal is easier to judge than a new button colour.
- Compare longer periods. Four to eight weeks before and after, with the same season. Songkran, year-end and school holidays shift behaviour for many Bangkok businesses, so compare like with like.
- Change one area at a time, so you know what caused the shift.
- Keep a change log with dates, so a jump in the chart has an explanation.
- Use qualitative evidence. If five recordings show people missing the button, move it. You do not need statistics to fix something plainly broken.
No dark patterns
Fake countdown timers, invented "only 2 slots left" messages, pre-ticked consent boxes and hidden decline buttons may lift a number briefly. They damage trust, and the consent tricks can put you on the wrong side of PDPA. Urgency is fine when it is true: a real promotion end date or a genuine booking limit.
Prioritising and next steps
You will end up with more ideas than time. A simple scoring method keeps the list honest. Rate each idea from 1 to 5 on three questions, then add the scores:
- Impact: how many visitors does it touch, and how close to the enquiry does it sit? A fix on your most visited service page scores higher than one on an old blog post.
- Confidence: how strong is the evidence? Recordings, customer comments and a clear drop in the data score high; a hunch scores low.
- Ease: how quickly and cheaply can it be done? Rewording a heading is a 5; rebuilding the booking flow is a 1.
Work from the top, one or two changes at a time, and note the date of each.
When the problem is the site itself
Sometimes the scoring points at structural issues: a theme that is slow on mobile, a page builder that makes every change fiddly, no room for a second language, or no clean way to track chat taps. Patching those endlessly costs more than fixing them once. That is where proper web design for service businesses earns its keep: clear first screens, visible actions, trust content and tracking-ready buttons planned from the start.
Our Starter package, from THB 10,000, covers up to five pages with WhatsApp, LINE and contact buttons plus PDPA cookie consent. The Business package, from THB 30,000, adds up to 15 pages in English and Thai with forms and booking or enquiry flows. Both are indicative starting prices, and our pricing page sets out what each includes before you ask for a written quote.
Who reviewed this guide
Who builds your site
Litesh Dhir
Builds and ranks the websites: the systems behind the 17 client sites whose Search Console numbers we publish.
More about usMore on web design and UX
Sources
- Google Analytics Help: Mark events as key events
- Microsoft Clarity
- Ministry of Digital Economy and Society: Personal Data Protection Act B.E. 2562 (2019), English translation