The short version
A quote form is not one decision, it is a sequence of them, and people quit at specific fields rather than quitting on the form as a whole. The fields that do the most damage are the ones that feel like an interrogation rather than a request: full address before anyone has said what the job costs, budget dropdowns, how did you hear about us, and anything marked required that the customer does not think is any of your business yet. The fix is not to cut the form down to a name and a number, because then you cannot quote anything. It is to ask only what you genuinely need to give a price, put the invasive fields after the submit, and measure which field people stop at so you are cutting the one that is actually costing you rather than the one you like least.
Forms do not fail all at once
When a form underperforms, the usual conclusion is that the form is too long, and the usual response is to delete fields until it feels short. That is guessing, and it often deletes the one field that made quoting possible while keeping three that nobody wanted to answer.
Forms fail field by field. Somebody starts typing, gets to a specific box, and stops. If you record which box that is, the whole argument is settled in a week of traffic, and you will usually be surprised, because the field people quit at is rarely the one the team assumed.
How to see it on your own site
You need two events: one when somebody focuses the first field, and one when each subsequent field gets a valid value. Any analytics tool can hold that. The output you want is a simple list, in form order, showing what share of starters reached each field. Read it top to bottom and look for the step where the number falls off a cliff instead of stepping down gently.
Do it for mobile separately. Mobile and desktop behave so differently on forms that averaging them tells you almost nothing. A date picker that is fine with a mouse can be genuinely unusable with a thumb, and if three quarters of your traffic is on a phone, the phone number is the only one worth arguing about.
The fields that cost the most
Four fields show up as cliffs over and over.
Full street address, asked before any price has been mentioned. From the contractor side this is obviously necessary, because you cannot schedule a visit to a postal code. From the customer side, they have known about your company for ninety seconds and you are asking where they live. Postal code alone is enough to confirm you cover the area, and you can get the street after they have agreed to a visit.
Budget, as a dropdown of ranges. This one is worse than it looks. The customer has no idea what the job costs, which is why they are on your site, so every option looks like a trap. Pick the low band and you look cheap, pick the high band and you are worried you just talked yourself into a bigger bill. A lot of people resolve that by closing the tab.
How did you hear about us. It helps you and does nothing for them, and it is sitting between them and a price. Ask it on the phone.
Anything asking them to describe the problem in a text box with no prompt. Staring at an empty box is work. People who were going to tap five buttons will not write you a paragraph, and the ones who do tend to write something you cannot quote from anyway. Give them four choices and an optional box.
Short forms are not automatically better
The overcorrection is a form with a name and a phone number, and it produces more submissions and worse days. You cannot price anything, so every lead becomes a discovery call, and half of them are for work you do not do in areas you do not cover. Your form is now a machine for generating phone tag.
The right question is not how few fields you can have. It is what is the smallest set of answers that lets you either give a price range or book a visit with confidence. For most trades that is four to six things, and they are specific to the trade. A furnace job needs the age of the unit and whether it is heating or cooling. A roof needs storeys, rough age and whether there is active leaking. A panel upgrade needs the current service size and the house era.
Notice what those have in common. They are all about the job, not about the person. Job questions feel like progress toward an answer. Personal questions feel like being added to a list.
Order changes the outcome more than length
The same eight fields in a different order produce different results, because early fields set the frame. Open with the job and you are having a conversation about their problem. Open with name, email and phone and you are collecting a lead, and everybody can tell the difference.
So put the job questions first, contact details last, and anything you want for internal reasons after the submit button on the thank you page, where it costs you nothing. People who have already submitted will happily answer how they found you, because the transaction is done and they are being helpful rather than being screened.
The details that quietly break things on phones
A few small things are worth more than any amount of copywriting.
Set the right keyboard for each field. A phone field that opens the alphabet keyboard adds two taps and a small moment of annoyance to every single submission. Email fields the same.
Turn on autofill properly, with the standard autocomplete attributes. Done right, a returning visitor fills their contact details with one tap. Done wrong, the browser offers nothing and they type an email address on a screen the size of a hand.
Never clear the form on a validation error. It still happens, and it is the single most reliable way to lose somebody who was actively trying to hire you.
Validate on blur, not on every keystroke. Telling someone their email is invalid while they are still typing the letter after the at sign is a small hostile act that makes people give up.
Show errors next to the field, not in a red bar at the top. On a phone, a message at the top of a long form is off screen, so the page appears to do nothing when they tap submit. They tap it twice more and then they leave.
What to do with the extra information you lose
Cutting fields does mean you know less at the moment the lead arrives, and that is a real trade, not a free win. The way to cover it is to move the questions into the callback rather than deleting them from the process. The person answering the phone can ask about access, parking, pets and the ten other things you actually need, and they can ask conversationally rather than as boxes on a screen.
That only works if somebody calls back quickly, which is why form length and phone coverage are the same project. A short form with a slow callback is a lot of leads you never reach. A long form with a fast callback is a small number of very well documented leads. The combination you want is a short form and a fast callback, and that is a staffing decision as much as a design one.
One thing worth testing that nobody tests
Try putting the phone number directly above the form on mobile, at the same size as the submit button. A lot of people who abandon a form were never form people in the first place. They were phone people who could not find the number, so they started typing because that was the only thing on the screen that looked like a way to reach you.
When we have done that on contractor sites, form submissions usually dip slightly and total contacts go up, because the calls more than cover the difference. That is a better outcome for the business and a worse looking number in any report that only counts form fills, which is exactly why it is worth checking who is reading your report and what they are being judged on, and whether the form is being optimised for the customer or for the chart.
Check it on your own site
Instrument every field on the form
Fire an event on first focus and one per field completed. Ten minutes of setup in any analytics tool. Without this you are rearranging a form based on opinions, and everybody in the room has one.
Read the drop-off separately for mobile
Look for the step where completion falls off a cliff rather than stepping down. Do it on phone traffic on its own, because that is where most of your customers are and where most form problems live.
Write down what you truly need to quote
Four to six job questions that let you give a range or book a visit. If a field does not change the price or the scheduling, it is not a quoting question and it should not be on the form.
Fix the mobile mechanics and remeasure
Correct keyboards, working autofill, inline errors, no clearing on failure. Then leave it two weeks and read the same drop-off chart again so you can see whether the change did anything.