Based in Arlington, TX. Serving any business, anywhere. 
(901) 490-4750 
contact@uicompass.com
The salon and the roofer
A hair salon puts online booking on its homepage. Bookings go up, no-shows go down slightly because of automated reminders, and the front desk stops answering the phone during appointments. It is an unambiguous win.
A roofing company copies the idea. Six weeks later they have a calendar full of "Roof inspection — 30 min" slots, half of which are for jobs that need two hours and a ladder, a quarter of which are at addresses two counties away, and a handful of which are people who booked an inspection when what they wanted was a price for something they saw on a neighbour's house.
Same feature. Opposite outcome. The difference is not the software.
The question that decides it
Can the customer, without your help, correctly choose what they need and how long it takes?
If yes, a calendar is the right tool. Fixed duration, fixed scope, fixed price. A haircut is 45 minutes for everyone. A dental cleaning is a known slot. An oil change, a massage, a piano lesson, a consultation with a defined length — all of these fit a grid.
If no, the calendar is a liability, because a booking that is wrong costs you more than a booking you never got. You have blocked a slot, driven somewhere, and started the relationship by telling the customer that what they booked is not what they need.
Most trades, most professional services, and anything involving a site visit or a quote fall on the "no" side. That does not mean they should never use scheduling — it means scheduling belongs after the first conversation, when you know what the job is. The right sequence is enquiry, quick call, then a booking link for the slot you both agreed on.
Four fields
For everyone on the "no" side of that question, the front door is a form. And the form is almost always too long.
The purpose of a first-contact form is not to collect a complete job file. It is to start a conversation with someone who is currently comparing you against two other companies. Everything you can ask in your reply, ask in your reply.
Our default is four fields:
- Name. One field, not two. Nobody has ever been lost by a first-name-only submission.
- One contact method. Phone or email — let them pick which. Requiring both is a common and costly habit; some people will not give a phone number to a stranger, and some will not give an email.
- What do you need? A single open text box. Not a dropdown of your service categories, which forces the customer to translate their problem into your vocabulary before they are allowed to talk to you.
- Where is the job? A postcode, suburb, or address, depending on trade. This one earns its place because it lets you triage instantly.
That is enough to call someone back. Budget, timeline, property type, how they heard about you — all of that is conversation, not gatekeeping.
Every field you add is another place to stall. That is not a claim about a specific percentage; it is a claim about direction, and the direction is consistent.
The details that cost you submissions silently
A form can have the right four fields and still leak, usually on a phone.
Wrong keyboards. A phone-number field without type="tel" or inputmode="tel" gives the visitor a full QWERTY keyboard to type ten digits on. An email field without type="email" gives them a keyboard with no @. Both are one attribute.
No autocomplete. The autocomplete attribute lets a browser fill name, email, phone, and address from what the visitor has already saved. A form without it makes a returning customer type everything by hand. This is documented in the HTML standard and supported everywhere; it is free.
Targets too small. Fields and buttons that are comfortable with a mouse can be genuinely hard with a thumb. WCAG 2.2 sets a minimum target size of 24 by 24 CSS pixels with specific exceptions, and platform guidance from both Apple and Google recommends considerably more than that for primary actions. Treat the WCAG figure as the floor, not the goal.
Validation that fights the user. Rejecting a phone number because it contains spaces, or an address because it has a hyphen, is a self-inflicted wound. Strip formatting on your side.
No confirmation. A form that clears itself and shows nothing leaves the visitor genuinely unsure whether it worked, so they either submit again or call. Show a real confirmation, and send an email that says what happens next and when.
A confirmation email that never arrives. This is the one nobody tests. Auto-replies sent on your behalf by a form provider need SPF and DKIM records under your own domain to be trusted, which is one of several reasons business email belongs on your domain. Send yourself a test from an address on a different provider and check the spam folder, not just the inbox.
The rest of the failure modes — silent delivery failures, forms posting to an address nobody reads any more — are worth working through properly in a contact form audit.
Offer two routes, always
Some people will never fill in a form. Some people will never make a phone call. These are not the same people, and both are buying.
So put both on the screen, and make it obvious which is which:
- A phone number that is a real
tel:link, in the header and near every call to action. - A form that is short, on the contact page and, ideally, embedded at the bottom of each service page so nobody has to navigate to use it.
A third route is worth adding if — and only if — you will actually watch it. Text messaging is genuinely popular for trades and often gets answered faster than either of the other two. A chat widget is fine when a person is behind it and a liability when it is an unattended bot adding weight to every page for no conversion benefit.
What does not help is a wall of options. Phone, form, chat, WhatsApp, Messenger, a booking calendar, and an email address all on one screen is not generosity, it is a decision you have handed to the customer.
Speed beats design
Here is the uncomfortable part.
You can spend a week on form design and lose to a competitor with an uglier form who answers in fifteen minutes. Somebody with an active problem — water on the floor, no heat, a broken storefront — is contacting several businesses in one sitting. The first useful reply usually sets the frame for the whole job.
This is an operations question wearing a web design costume, and it is worth being honest that the website can only do part of it. What the site can do:
- Route the enquiry to somewhere a human actually looks — a monitored inbox on your own domain, not a form-provider dashboard nobody logs into.
- Send an instant auto-reply that sets an expectation: "Got it. We reply to every enquiry the same business day. If it is urgent, call 214-555-0142."
- Include the job location in the notification so whoever is holding the phone can triage without opening anything.
Those three take an afternoon and they change the response time more than any layout change will.
A short decision guide
- Fixed-duration, fixed-price, customer knows what they want → online booking on the site, with reminders.
- Needs a quote, a site visit, or scoping → short form plus a prominent phone number. Booking link after the first call.
- Emergency or urgent work → phone first and loudest, form as backup. Put your hours and your after-hours arrangement in plain text next to the number.
- High-value, long-cycle work → form with a slightly longer set of fields is acceptable here, because the enquirer is already invested. Still fewer than you think.
None of this requires a platform decision. It requires deciding what happens in the first ten minutes after someone decides they want to talk to you, and then building the shortest possible path to it.
Frequently asked questions
Should I put online booking on my site or not?
Ask one question: can a customer, unaided, correctly choose what they need and how long it takes? A haircut, a dental cleaning, a 60-minute massage, an oil change — yes. A roof repair, a custom kitchen, a legal consultation, a commercial network install — no, because the customer does not know the answer and a wrong booking costs you a wasted trip. For those jobs the calendar belongs after the first conversation, not instead of it.
How many fields should a contact form have?
Four is our working default: name, one contact method, what they need, and where the job is. Every additional field is another chance for someone to close the tab, and most of what gets added — company name, budget range, how did you hear about us, preferred contact time — is information you could ask for in your reply, once they are already talking to you.
Is a phone number still worth showing prominently?
Yes, and for a large share of local service businesses it is still the primary route. The mistake is showing it as plain text. On a phone, a number that is not a tel: link means the visitor has to memorise or copy it, and some fraction of them simply will not — see tap-to-call for the fix, which takes about a minute.
Do multi-step forms convert better than single-page ones?
Sometimes, and it depends on why the form is long. Splitting a genuinely long form into steps with a visible progress indicator can help, because the first screen looks manageable. Splitting a four-field form into three steps is theatre and adds taps. If a multi-step form is your solution to a form that has fourteen fields, the real fix is fewer fields.
What about a chat widget?
It works when a real person answers within a minute or two during the hours you claim to be available. It works against you when it is a bot that collects the same information a form would have, adds a script to every page, and delays an answer that a form would have delivered identically. If nobody is watching it, take it off.
How fast do I actually need to reply?
Faster than the competitor the customer contacted at the same time. Someone with a leak, a broken AC, or a wedding in six weeks is not sending one enquiry — they are sending three. Whoever answers first frames the job. We tell clients to treat same-business-day as the floor and under an hour as the target during working hours; those are operating targets we set, not an industry benchmark.
