Form analytics: which field is losing your leads?

Short answer: form drop-off analysis compares the people who started your form with the people who finished it, field by field. The field with the biggest gap, or the one people reach before giving up, is your first fix. Your analytics says where the gap is; session replays and heatmaps show the cause.

What form drop-off analysis actually measures

A form has three moments: the visitor lands on the page, they start filling the form in, and they submit it. Drop-off analysis is the comparison between those moments, then the same comparison repeated field by field.

Most small sites only have the first and the last number: the form tool reports submissions, the analytics reports page views. That is enough to know something is wrong, not enough to know what to change. That is where the field-level view earns its place.

Metric	What it counts	What a bad value tells you
Views to starts	Visitors who saw the page, then touched the form at all	The page is the problem, not the form: placement, contrast, or a promise that does not fit the form
Starts to submissions	Visitors who began the form, then sent it	The form itself is losing people: a field, a format, a bug, or a cost revealed late
Field abandonment	People who reached a field and never reached the next one	The question to fix first, and the only number that names a guilty field
Validation errors per field	Failed submits caused by each field	The format is fighting the answer: a number field that rejects a landline, a postcode that rejects a foreign address
Time on field	Seconds between arriving on a field and leaving it	The visitor had to go and look something up. That question belongs later in the process, or nowhere

One warning: with form fields masked, as they should be, you never see what people typed. You do not need to. What matters is where each session stopped, and how many stopped in the same place.

The five questions I answer before touching the form

The mistake I see most often is jumping from "the form converts badly" to "let us shorten the form". Five questions first, in this order.

Where is the loss, exactly? Before the form, inside it, or on the submit button. The three numbers in the table above answer this, and the rest of the work depends on the answer.
Is it one field or all of them? One field is a wording problem. A steady drip across every field is a length, order or trust problem, and shortening the form is then a fair answer.
Does it depend on the device? Open the form on a real phone, not in a resized desktop window. A field that needs a dropdown, a date picker or several taps fails on mobile long before it fails on your laptop.
Does it depend on the source? If visitors arriving from an ad abandon at a field your returning visitors fill in happily, the page promised something the form did not deliver.
Did they try and fail, or decide not to? A failed attempt leaves errors and repeated taps. A decision leaves a quiet exit. The first is a bug to repair today, the second is a question to rewrite.

That last distinction changes the fix: bugs are cheap to correct and cost leads every hour they run, while wording is a judgement call where watching a handful of sessions beats any amount of internal debate.

How to find the field that loses the lead

This takes an afternoon on a site with modest traffic. The order is what makes it reliable.

Define completion once. Put a goal on the submit itself, not on the click of the button: a click is an attempt, a submit is a lead. If your goals currently count something vaguer than that, start with goals that answer a real question.
Write down a baseline before you change anything. Same period, same audience: page views, starts, submissions, then the same three filtered on mobile, then on your one biggest source.
Compare the two ratios. The two ratios decide where the problem lives: views to starts low is a page problem (placement, a form that looks long, a cost shown late), starts to submissions low is a form problem. When I skip this step, I end up rewriting a form that nobody ever reached.
Get the field view, or reconstruct it. If your form tool reports abandonment per field, take the field at the top of the list. If it does not, filter the sessions that reached the form page and did not convert: the last field they touched, or the page position where the sessions end, points to the culprit. Ten to twenty sessions are usually enough to see the pattern repeat, and it is the same leak-first logic as the wider exercise in why visitors leave your website.
Check device and source before blaming the field. A field that is innocent on desktop and guilty on mobile is a keyboard and touch problem. A field that only fails for one source is a promise problem.
Write the change as a hypothesis, and ship one field at a time. "People do not know their customer number, so we will ask for it by email after the first reply, and we expect completion to rise on mobile within two weeks." One field, one date, one expectation.

If your traffic is too thin to see a difference afterwards, that is a known problem with a known method: what to do instead of A/B testing.

The symptoms and what they usually mean

Once you know where the sessions end, the symptom usually names the cause. In my experience these five cover most of what I find on small business sites.

What you see	What it usually means	The check that settles it
Drop on mobile only, at one field	The input type fights the phone: a full keyboard where a keypad belongs, a picker nobody can hit	Fill the form on a real phone, one thumb, in daylight
Drop right after the submit button	Silent validation failure: an error message out of view, or a button that looks dead	Submit the form empty, then with a wrong format, on purpose, and watch what happens
A steady drip across every field	Length, order, or a question people cannot answer from memory	Count the fields, and mark the two you have never used in a reply
Abandonment concentrated at one label	The question is ambiguous, or asks for a format many visitors do not have	Ask someone outside your business to fill it in while you stay silent
Long time on a field, then exit	The answer needs looking up: an invoice number, a policy number, an account reference	Move that field to a follow-up step, or make it optional

None of these need a redesign. All of them change the completion rate, because they act on the exact point where the visitor stopped.

Which fix to make first

I order these by cost of delay, not by elegance.

A form that rejects valid input. Every hour it runs, it loses leads and damages trust. This is not a test, it is a repair, and it comes first.
The one field with the biggest drop, if you can delete it, defer it or make it optional. Free to ship, easy to revert, and it removes the exact obstacle the data pointed at.
Input types and formats on mobile. The right keyboard, autofill that works, no fight with the address or the calendar. Boring work, immediate effect on a group that is often half your traffic.
Whatever hides the cost or the commitment. The price, the delivery time, the sales call the visitor is agreeing to. In my experience, saying it on the page converts better than revealing it on a later step: the people who leave at the reveal rarely come back to read it afterwards.
Shortening the form, last. Cutting fields people fill in without hesitating buys you nothing and costs you information. Cut the field that is losing people, then look at the total count with fresh eyes.

Whatever you change, mark the date on the chart and read the trend afterwards. You are looking for a step on a line rather than a feeling; as an illustration, a mobile segment moving from 40 to 60 submissions per hundred starts is the kind of before and after that shows up.

What to measure it with

Three things, and no more:

The form's own numbers: starts, submissions, and per-field abandonment if your form tool reports it.
One goal on the submit, so views and completions line up in the same report instead of two dashboards you compare by hand.
The sessions behind the gap: normally, to see the session end, and a click and scroll view of the form page to see what got ignored on the way there.

On tooling, the honest boundary first: SessionInsight does not rebuild your form, and it does not read what visitors type into it. What it does is keep the numbers and the sessions in one place: a goal on the submit, the page journey of each visit, click and scroll heatmaps (clicks and scroll positions only, not field content), sources, realtime, email reports and shared dashboards, plus a cookie-less lightweight mode for basic analytics. If you record visitors in the EU, sort the consent question before you record anything: a short checklist is in are session replays GDPR compliant?

You can see how it works or look at the free plan, which covers one website and 50,000 pageviews a month. Pair it with whatever your form tool already reports: the field-level numbers and the session that ended are two halves of the same answer.

FAQ

What is form drop-off analysis?

It is the comparison between the people who started your form and the people who submitted it, repeated field by field. The first number tells you the form is losing leads, the field-by-field view tells you which question to change first.

How do I find the field that loses the most leads?

Take the per-field abandonment from your form tool if it reports it. If it does not, filter the sessions that reached the form page and did not convert, and look for the field or the screen position where those sessions end. Ten to twenty sessions are usually enough to see which field repeats.

Is a shorter form always better?

No. Length only hurts when it delays something the visitor wants, or asks for an answer they do not have at hand. Cutting fields nobody abandoned buys you nothing and costs you information. Fix the field people actually leave, then look at the total count.

Should I use a multi-step form?

It helps when the first step is easy and the commitment grows step by step. It hurts when each step asks for something personal before the visitor has seen any value. Keep the number of fields identical, and compare your completion rate before and after.

Can analytics show form field drop-off on its own?

Analytics alone usually shows two numbers: views of the page holding the form, and submissions. Everything between them comes from your form tool or from watching the sessions. That gap is the part worth measuring, because it is where the lead is lost.