Form Design and Friction: Every Field Costs You, Which Ones Earn Their Place

B2B forms grow the way committees grow. Sales wants phone number, marketing wants company size for segmentation, ops wants a dropdown that maps to CRM picklists, and each request is individually reasonable, so the form swells one defensible field at a time. Nobody owns the total, and the total is what the visitor experiences. A form audit that asks each field to justify itself against a simple question, what decision do we actually make differently with this answer, will typically find several fields that exist for a report nobody reads.

The cost of each field is not uniform either. Email and name are expected and cheap. Phone number carries the implicit threat of a cold call and often costs disproportionately more completions than it is worth at the first touch. Open text fields asking visitors to describe their needs put real work on the person with the least obligation to do it. Rank your fields by what they cost in abandonment against what they change in your handling of the lead, and cut from the bottom.

A large share of what B2B forms traditionally ask, company size, industry, revenue band, tech stack, can be derived from a work email via enrichment rather than extracted from the visitor. Every enrichable field you remove is conversion you get back at no cost in data quality, and often at a gain, because self-reported firmographics are frequently wrong anyway, filled with whatever gets the visitor through the gate fastest.

The practical rule: ask for what only the visitor knows. Their specific problem, their timeline, their role in the decision if it matters to routing, these are legitimately human answers. Their employee count is not. Progressive profiling extends the same logic across time, collect the minimum at first conversion, then fill in the picture at later, higher-commitment touchpoints when the relationship has earned deeper questions.

Not all friction is waste. If your sales team is drowning in unqualified demo requests, a slightly longer form can act as a filter that trades volume for fit, and that trade is sometimes correct. The mistake is applying one friction level to every conversion point. A gated whitepaper deserves near-zero friction because the intent is shallow and the asset is cheap. A demo request or pricing conversation can carry more fields because the visitor's intent is high enough to survive them, and the answers genuinely shape the first conversation.

The test for a deliberate-friction field is whether a human uses the answer within the first touch. If the rep opens the call by referencing what the prospect wrote, the field earned its place. If the answer goes into a CRM field that nobody reads before the call, you charged the visitor a toll and threw away what they paid. In practice, most forms fail this test on at least a third of their fields.

Beyond field count, the implementation details often cost more than the fields themselves. Validation that fires while the visitor is still typing, error messages that fail to say what to fix, dropdowns with dozens of unsorted options, forms that wipe all input on one error, each of these is a small cruelty that compounds. On mobile these problems roughly double, so test the form on a real phone, not a resized browser window.

Measure abandonment at the field level, not just form-level conversion, because the aggregate number hides which specific field is the cliff. It is common to find one field, often phone number or a mandatory dropdown, accounting for the bulk of drop-off. And remember that the form is not the end of the flow: a submission that leads to a dead thank-you page with no expectation set about what happens next leaks the intent you just paid to capture. Say what happens, when, and from whom, right on the confirmation.