On this page
The job the page is doing
Someone has typed a problem into a phone, tapped a result, and arrived. They have not agreed to read anything. They are deciding whether this firm is plausibly the answer, and they will make that decision on the first screen and roughly one scroll, then either act or go back and tap the next result. That window is the page's entire opportunity.
Which makes the first mistake obvious once you look for it: sending paid traffic to a homepage. A homepage has to serve eleven audiences at once, so it introduces the company, lists everything, and answers nobody's specific question. The visitor who searched for gutter replacement in Redditch has to work out where that lives. Some do. Most tap back, and you paid for the click either way.
The second mistake is message mismatch. If the ad said “emergency drain unblocking, Coventry, 60 minute response” and the page headline says “Welcome to our family-run business”, the visitor experiences a small jolt of doubt about whether they have arrived in the right place. Say the same thing the ad said, in nearly the same words, at the top. It looks crude on a screen and it works.
The seven questions
In roughly the order they occur to someone with a problem:
| The question | Where the answer belongs | What a weak answer looks like |
|---|---|---|
| Do you do this exact job? | The headline | A company name, or a list of eleven services |
| Do you come to where I am? | The subheading, in named towns | “Serving the Midlands and surrounding areas” |
| Can you come when I need you? | First screen, next to the number | Nothing at all, which reads as no |
| Roughly what will this cost? | Within the first scroll | “Competitive prices” and a quote form |
| Are you a real, safe firm? | Immediately after the price | An unnamed accreditation badge with no number |
| What happens if it goes wrong? | Alongside the proof | “Fully guaranteed” with no terms |
| What do I do now? | Everywhere, and it is one action | Three competing buttons of equal weight |
Question four is the one firms fight hardest and lose most from. The fear is that publishing a range invites price shopping. What actually happens is that the people who leave over the number were never going to buy at your price and would have taken forty minutes of your time to reach the same conclusion, and the people who stay have pre-qualified themselves. An honest range with the variables named — access, scale, materials, whether scaffolding is needed — is more persuasive than silence, because silence reads as expensive.
A structure that answers them in order
- Headline: service plus place. Plain, matching the search. “Gutter replacement in Redditch and north Worcestershire.”
- Subheading: the qualifier and the timing. Who you work for, how quickly you get out, and any minimum job size. Filtering here saves you calls you did not want.
- The action, immediately. A tap-to-call number at a size you cannot miss, with the hours beside it. On desktop, the number plus a short form.
- Three specifics, not adjectives. Years trading, the registration number of your accreditation, the length of the guarantee.
- Price guidance. A range, and the three things that move it.
- Two or three recent jobs with photographs, the town, what was involved and roughly how long it took.
- Reviews with a first name and a place. Four is plenty. Pick ones that describe the work rather than praise the attitude.
- The four objections you hear on the phone, answered plainly. You already know what they are.
- The action again, and a footer that repeats the number, the areas and the accreditation.
Removing the site navigation from a paid landing page is standard advice and it does help slightly, mostly by removing exits. It is a refinement, not a strategy, and it will not rescue a page that has failed to answer question one. Get the order right first.
Proof that survives scepticism
Assume the visitor has been let down by a trade firm before, because most have. Everything on the page is read through that. Specificity is the only thing that gets past it.
- Registration numbers, not badges. A Gas Safe or NICEIC logo is a picture. The number beside it is checkable, and checkable is the point.
- Your own photographs. Taken on site, on a phone, in ordinary light. Stock imagery of a smiling man in a hard hat is recognised instantly and quietly discounts everything around it.
- Named places. “Flat roof replacement, Alcester, March” is evidence. “Recent project” is not.
- Guarantee terms in numbers. What is covered, for how many years, and what voids it.
- Insurance stated as a figure, because “fully insured” is what everybody writes.
- A face and a name. One photograph of the person who will turn up does more for a domestic enquiry than a page of copy.
The awkward part is that this is production work rather than writing. Most trade pages underperform because nobody has taken forty decent photographs, not because the copy is bad. That is worth knowing before you commission a new site and expect the design to carry it.
The form, and why shorter is not always better
The received wisdom is that fewer fields mean more submissions, and that is true. It is also incomplete, because it optimises for the wrong number. Fewer fields raise the count of enquiries and lower the average quality, and a firm that already cannot get round its quotes does not want more enquiries, it wants better ones.
The useful approach is to keep the form short and add one deliberate qualifying field. Postcode is the best single choice for most trades: it filters people outside your area at no cost to anyone inside it, and it tells you where to send the van. Job type or rough timescale works where scope varies more than geography.
Some practical rules that hold up. Fire the conversion when the submission is confirmed, not when the button is clicked, or you will count every fumbled tap. Give people a route that is not a form, because a large share of trade customers would rather send a WhatsApp message with a photograph of the problem than type into boxes. And on genuinely urgent services, consider having no form at all — somebody standing in three inches of water is not completing fields. Whatever arrives has to reach someone quickly, which is the argument for automating the acknowledgement rather than hoping the inbox is being watched.
What mobile actually changes
Most trade enquiries begin on a phone, often outdoors, often in a hurry, sometimes on a poor connection. Design for that case and the desktop version looks after itself.
- A sticky call bar along the bottom of the screen, present from the first pixel of scroll. This is the single highest-impact change on most trade pages.
- Tap targets you can hit with a thumb, not links sized for a mouse pointer.
- The number as a real link, so tapping dials. A phone number rendered inside an image is a surprisingly common own goal.
- Photograph weight under control. Site photographs come off a phone at several megabytes each. Resize and compress them or the page will not load before the visitor gives up.
- Numeric keyboards on numeric fields, so a postcode or phone entry does not open a full keyboard.
- No carousels. People do not swipe them, and the content after the first slide is effectively invisible.
Test the finished page on a real phone on mobile data, away from the office, ideally in the sort of place your customers actually are. It is astonishing how many pages that pass a desktop review fail this.
Why most trade A/B tests tell you nothing
This is where a lot of marketing advice goes wrong for businesses of this size, so it is worth doing the arithmetic honestly.
To detect a genuine improvement from, say, a five per cent conversion rate to six and a half per cent, at conventional confidence and power, you need something in the order of three thousand visitors per variant. A trade landing page receiving four hundred visits a month would take over a year to finish that test, by which time the season has turned twice and you have rebuilt the site. Run it for a fortnight instead and you will get a result, but it will be noise, and acting on noise is worse than not testing at all because it feels like evidence.
So do the following instead.
- Make changes that are obviously correct without testing them. Adding a sticky call bar, answering the price question, putting the right headline on the page. Nobody needs a test to know these help.
- Compare over quarters, not weeks, and compare against the same quarter last year where you can, so seasonality is not doing the talking. The seasonality guide explains why that comparison matters more than it sounds.
- Use qualitative evidence, which is abundant at low volumes even though statistical evidence is not. Listen to ten call recordings. Watch ten session recordings and find where people stall. Ask the next five customers what nearly stopped them ringing.
- Test one big difference, not two small ones. If you must run a test, test a genuinely different page rather than a button colour. Large effects need far smaller samples to detect.
The pages that work in this sector are not clever. They say what the job is, where you go, when you can come, roughly what it costs, why you are safe to let in, and what to press. Everything we have measured across our scaffolding, roofing, glazing and surfacing work points the same way, and the terms used above are defined in the glossary if any of them are unfamiliar. If the page is carrying paid traffic, it is worth building it properly alongside the campaign rather than after it.
Rather have someone else do this?We run exactly this for trade and construction firms across the UK. Call and we will tell you what is realistic in your area, including when the answer is that it is not worth it.
Call 07443 392243WhatsAppCommon questions
Should paid traffic go to the homepage or a dedicated page?
A dedicated page, every time, for any service you are actively advertising. A homepage has to serve too many audiences to answer one searcher's specific question quickly, and the cost of that hesitation is paid for by the click you already bought.
How long should the page be?
As long as it takes to answer the seven questions and no longer. In practice that is usually one screen of decision-making followed by two or three screens of proof for the people who want it. Length is not the variable that matters; order is.
Do I really have to publish prices?
You have to answer the cost question in some form. A range with the variables named is usually enough, and it filters people you were never going to win before they take up your afternoon. Refusing to engage with it at all reads to most visitors as expensive.
Is a video worth adding?
A short clip of you explaining what happens on a typical job can work well, especially for domestic work where the concern is who is coming into the house. Keep it under ninety seconds, never autoplay with sound, and make sure it is not delaying the page load for the people who will not watch it.
What conversion rate should I be aiming for?
It varies too much by service and traffic source for a single target to be meaningful, and anyone quoting one figure across all trades is guessing. The useful comparison is your own page against itself over time, at the same time of year, with one thing changed.