How Do You Design a Form That Asks for Less?

How do you design a form that asks for less?

Start from the next action rather than the database. Write down what happens immediately after someone submits, list only the information that action genuinely requires, and delete everything else. Most B2B forms shrink by half under that test and nothing downstream breaks.

Forms grow the way spreadsheets grow. Somebody adds a field because it might be useful, nobody ever removes one, and three years later a demo request asks for company size, industry, number of employees, and a dropdown labelled how did you hear about us that nobody has read a single answer from.

I have rebuilt a lot of these across more than seventy projects, and the pattern is consistent. The team defends every field in the abstract and cannot name who used any of them last month.

What is each extra field actually costing you?

Each field is a separate decision the visitor has to make, and each decision is a chance to stop. The cost is not the seconds of typing. It is the moment where someone looks at a field, does not have a confident answer, and decides to come back later, which usually means never.

Fields with dropdowns are worse than they look because they force a categorisation. Asking someone to pick their industry from twelve options when they sit awkwardly between two of them creates a small unresolved problem, and small unresolved problems are exactly what makes a person close a tab.

There is a second, quieter cost. A long form signals what kind of company you are. It says that the process from here will also be long, that you will qualify before you help, and that you are organised around your own internal handoffs. Some buyers are fine with that. Founders and small teams usually are not.

Which fields can you delete outright?

Anything nobody has acted on in the last quarter. How did you hear about us, company size, industry, job title, and phone number are the usual candidates. If you cannot point to a decision somebody made differently because of a field, it is collecting data for the feeling of completeness.

Phone number deserves a specific mention because it is the most defended and the least honest field on most B2B forms. It is marked optional or required depending on how aggressive the team is feeling, and in both cases a meaningful share of what arrives is invented. You have not gained a channel. You have added friction and acquired fiction.

The how did you hear about us field is the other one I remove almost every time. The intent behind it is attribution, which is a real need, but a free text box filled in by a distracted person is the least reliable way to meet it. Attribution belongs in your analytics, not in your visitor's homework.

Which fields can you infer instead of asking?

Several. Company and company size can usually be derived from the email domain. Country and time zone come from the request. Which page they submitted from, what campaign brought them, and what they read before converting are all already in your hands if you capture them as hidden fields.

Inference is better than asking on two counts. It is more accurate, because a domain is a fact and a self reported dropdown is a guess made in three seconds. And it is invisible, so it costs the visitor nothing. Every field you move from asked to inferred is a pure gain.

The one thing to be careful about is honesty. Inferring context to route a request well is ordinary practice. Presenting inferred information back to the person as though they told you it is unsettling, and it damages trust at exactly the moment you were trying to build it.

How should the remaining fields be ordered?

Easiest first, most personal last. Email before name, name before anything about their company, and the free text box at the end. Each completed field increases the chance the next one gets completed, so the order is not cosmetic, it is a sequence of small commitments.

Put the free text field last and make it genuinely optional, because it is the field that produces your most useful information and the one most likely to stall someone. A person who has already filled in three things will often write you a paragraph. The same person facing a blank box first will usually leave.

Label the free text box with a specific question rather than a generic invitation. Asking what you are trying to fix gets you a real answer. Asking for a message gets you the word hi, and then you are back to a discovery call you could have skipped.

What should a good error state do?

Appear next to the field it concerns, in plain language, and explain what would be acceptable. Not red text at the top of the form. Not a generic failure notice. The visitor should never have to scroll or guess which of six fields the complaint is about.

The worst version of this, which I still see regularly, is a form that clears itself on a validation failure. Somebody has typed four fields, one is wrong, and the reset wipes all four. That is not a form problem, it is a decision to make a person do the work twice, and almost nobody does it twice.

Validate as the person leaves a field rather than only on submit, and never block typing. Inline validation that fights the user while they are mid entry is more irritating than late validation. The goal is to catch the mistake before submit, quietly, once they have finished the thought.

What should happen after submit?

Something should visibly change, immediately, and the next step should be named. A button that shows it is working, then a confirmation that says what will happen and roughly when. Silence after a click is the point where people submit a second time and assume you are broken.

The confirmation itself is a design surface that most sites waste. It is the one moment where you have someone's full attention and a completed action behind you, and most teams spend it on the word thanks. It is a better place for a relevant next step than almost anywhere else on the site, which I went into in thank you page design and wasted real estate.

Set the expectation honestly. If you reply within a day, say a day. If you reply within three, say three. An accurate promise that is slower than the visitor hoped is far better for your reputation than a fast one you miss, and they will remember which one you were.

How do you get a team to agree to fewer fields?

Do not argue about the fields. Ask who used each one, for what decision, in the last ninety days, and write the answers down next to each field. The conversation resolves itself, because the fields nobody can answer for become obviously indefensible in front of everyone.

When someone genuinely does use a field, keep it, and you have learned something real. My experience is that one or two survive and the rest were added by people who have since left or for a process that no longer runs. That is not incompetence, it is just what happens to forms over time.

If the team is still nervous, offer the reversible version. Remove the fields, watch for a month, and put back anything whose absence caused an actual problem. Almost nothing gets put back, and framing it as reversible gets you the decision today instead of in a quarter. The same framing works for most conversion changes, including where you place a secondary call to action, which I covered in where to put the second call to action.

What should you do next?

Open your main form and write, next to each field, the name of the person who used it and the decision they made with it this quarter. Delete every field where that line stays blank. That is usually the entire redesign and it takes half an hour.

Then move whatever you can into hidden fields and inference, order what remains from easiest to most personal, and check that the error state points at the right field without clearing the rest. Those four moves cover most of what makes a form feel heavy.

If you have a form that gets traffic and few submissions and you want a second opinion on which fields are doing the damage, send me the link. I am happy to tell you what I would cut first.