Short answer: most sites that generate enquiries and do not book them are losing people to slowness, ambiguity and unnecessary form fields — not to weak copy. Speed up the reply, state plainly what happens next, and remove the steps nobody needs. Those three changes outperform a copy rewrite on almost every service site, and they are also what make the site easier for an assistant to summarise accurately.
Start with the four things that decide it
Before writing a word, it helps to have the list in the right order. A visitor comparing you against two competitors is running four checks, and only one of them is about your words:
- Do I trust them? Usually decided before they read — by how the site looks, whether reviews look real, whether the business is identifiable.
- Do I understand what happens if I get in touch? This is the one most sites fail, because “Contact us” and a form describe nothing.
- Are they available when I need them? If you only answer enquiries at 9am and they wrote at 6pm on a Friday, the rest is irrelevant — which for a local business is usually about the hours you are findable rather than about your copy. Being findable in the right place at the right time is a large part of local search, and it is covered in local search versus AI answers. [4]
- Does this page give me a reason to act now rather than later? Genuinely one of the harder ones.
Copy is a component of the first and fourth. It is not the lever people assume it is. A rewrite that does not address the other three produces a page that reads better and converts identically.
Key takeaway: if you fix one thing this quarter, fix the response time. It affects every enquiry and it is the one nobody optimises.
Strategy 1: the reply, not the pitch
The most valuable change on most service sites is not on the site at all. It is what happens in the ninety seconds after someone gets in touch.
An automatic acknowledgement that confirms receipt, states your response window and tells them what you need from them to quote — a photo, an address, a rough date — removes two anxieties at once: “did this send?” and “what happens now?” It also gets you the information required to quote without a second exchange. This is the same argument as the one in the integration edge, which covers which parts of a workflow are worth automating.
You do not need an assistant to reply immediately. You need the acknowledgement to be immediate and the answer to be honest about timing.
Strategy 2: make the next step a sentence, not a link
The most consequential line on most service pages is a navigation label. “Contact us”, “Get in touch”, “Learn more” — these are the default, and they describe nothing about what the visitor gets.
What works better is naming the thing: state the length, the cost if there is one, and what you need from them. A visitor deciding between three businesses is not avoiding your CTA; they are trying to find out whether it leads somewhere they want to go. Nielsen Norman Group’s heuristics put the same point differently — systems should speak the user’s language and make the consequence of an action clear before it is taken. [1]
Google’s helpful-content guidance frames the same thing as usefulness to someone with a real question. [2] This also has a second audience. When an assistant summarises your site for someone asking how to proceed with your service, it will carry across a sentence that says something concrete, and drop a label that says only “Contact”. Headlines and the line above the fold covers the top of the page for the same reason.
Strategy 3: cut friction you did not know you had
Friction is mostly administrative, and the useful examples are boring:
- Form fields that are not needed to do the job. Every one of them costs completions, and the cost is higher on a phone.
- Phone numbers buried in a header image, so they cannot be tapped or read by an assistant.
- A form that resets on error without preserving what was typed.
- A confirmation that promises a callback nobody has scheduled.
- A booking flow that sends you to a third-party platform which then asks for your email again.
- A PDF, or a page that explains you should fill in another form.
None of this needs a designer, and readers scanning rather than reading in order means the worst offender is often the one they pass. [3] Walk the path on a phone, on mobile data, and note every point where you hesitated — hesitation is the signal, and analytics will not show it to you because it only records what completed.
Once you have the list, the order of fixes is obvious: things that block completion before things that make it nicer.
Strategy 4: proof placed where the decision happens
Reviews do not need a dedicated page to work. They work when one of them describes the same problem the visitor arrived with, quoted directly under the service that solved it.
A review saying “they came out same day when the boiler died” is doing work for a reader currently looking at a broken boiler. A review saying “great service, very professional” is not — it could describe any company in any trade, and readers know it.
You almost certainly already have reviews. [5] Getting them onto your profiles covers the request mechanics; Google’s guidance on how to solicit them is worth reading first, because how you ask shapes what you get back.
Strategy 5: one clear action per page
Two competing calls to action on one page reliably produce less action than one, because the visitor is now deciding rather than doing.
Pick the single step you want from a given page and make that the page’s primary action. Everything else — the phone number, the directions, the hours — stays available but subordinate. If your service pages have six buttons, the page is asking the visitor to prioritise on your behalf.
The wording argument is in direct directives, and the structural argument — where actions sit in a site — is in schema and structured data.
Key takeaway: friction you can see costs you conversions; friction you have forgotten you added is costing you the same and nobody is measuring it.
The part most sites never look at
Booking is not the end of the transaction, and it is rarely the end of the funnel either. On most local service sites, a large share of enquiries do not become jobs on the first contact — they become jobs three weeks later, or they become nothing, and the site has no idea which.
Two numbers close that gap: how many enquiries you receive, and how many become booked work. A big gap between them is the most useful diagnostic on the site, and it is more informative than any conversion rate computed against total sessions. It also points straight at the cause — reply time, quote turnaround, unclear pricing, or capacity.
What to do first
- Time your own reply to an enquiry, on a realistic day. If it is over two hours, fix that before anything else.
- Read your contact page as if you were comparing three businesses. Does it say what happens next and when you will respond?
- Walk your booking path on a phone and remove every step that is not load-bearing.
- Pick one page. Pick one action. Delete the other calls to action.
- Start recording enquiries against booked work, even in a spreadsheet.
None of this is complicated, which is exactly why it is skipped. The version of this work that matters is not the copy — it is the ninety seconds after someone decides to trust you with their problem.
How this connects
- Reducing friction goes deeper on the specific steps that cost conversions.
- A justification web covers getting evidence to sit next to each claim.
- Local search versus AI answers covers being found for local intent specifically.
- Work smarter covers which parts of this are worth automating.
