"We need a form" is how most of these projects start, and about a third of the time a form is the wrong tool. The three formats — form, survey, funnel — look similar from the builder's side (screens, fields, a submit) but they are built for different jobs, and picking wrong costs you either conversions or data quality.

Here is the decision framework we use, followed by the concrete scenarios where teams most often pick wrong.

What's the actual difference?

Definitions first, because the marketing language blurs them:

  • A form is a single screen of fields with a submit button. The visitor already wants to do the thing — contact you, register, apply, order. The form's job is to not get in the way. Success metric: completion rate. Forms on Formspring →
  • A survey is a multi-question instrument for collecting data you will analyse in aggregate. The respondent is doing you a favour. The survey's job is honest, comparable answers at sufficient volume. Success metric: response quality × response rate. Surveys →
  • A funnel is a multi-step persuasion flow that qualifies and converts a visitor who is not yet committed. Each screen earns the next click; somewhere along the way the visitor becomes a lead. Success metric: qualified conversions per visitor. Funnels →

The one-line test: a form serves intent that exists, a survey extracts knowledge you lack, a funnel builds intent that doesn't exist yet.

Which three questions decide it?

When the format is not obvious, three questions resolve it:

1. Who wants this interaction more — you or the visitor? If the visitor does (they came to contact you, sign up, order): form. Minimise friction, ask the minimum, get out of the way. If you do (you want their opinion, their email, their attention): survey or funnel — and you must pay for the interaction with motivation, either a result worth getting (funnel) or goodwill and brevity (survey).

2. Will you read responses individually or analyse them in aggregate? Individually — each submission triggers a human action — points to form or funnel. In aggregate — charts, segments, trends — points to survey, and that changes the design rules entirely: now question wording, scales, and ordering matter more than conversion polish, because biased data at high volume is worse than no data.

3. Does the visitor need convincing or sorting along the way? If the visitor must be qualified, segmented, taught, or persuaded between arrival and conversion — funnel. The multi-step shape exists precisely to do work between the first click and the ask: self-qualification, branching by intent, scoring, and an email gate placed where motivation peaks.

What does each scenario actually call for?

The concrete cases, including the ones that trip teams up:

"Contact us" on a marketing site → form. Always. Visitors with intent want one screen and a fast confirmation. Multi-stepping a contact form adds drop-off and zero value. (Most intake should stay this boring — the forms guide covers the craft of boring done well.)

Demo requests from paid traffic → funnel. Cold visitors clicking an ad are not committed. A funnel that hooks ("see how teams like yours cut onboarding time"), qualifies (company size, urgency, current tool), and then asks for the email converts the uncommitted in a way a static form cannot — and the qualification answers route the lead before a human reads it.

Post-purchase feedback → survey. You want comparable, aggregate data ("how was checkout, 1–5?") across hundreds of buyers. Keep it under two minutes, fix the wording, trend the results. If you find yourself wanting to respond to individuals, you want a feedback form instead — or an NPS program with a closed-loop follow-up, which is a survey wearing operational shoes.

Job applications → form, even though they are long. The applicant brings maximal intent; what they need is clarity and file uploads, not persuasion screens. Long ≠ funnel.

"Which plan is right for me?" on a pricing page → funnel — specifically a quiz funnel. The visitor gets a personalised recommendation, you get segmentation and a lead. This is the scenario teams most often miss: it usually ships as a static comparison table when an interactive recommender would both help the visitor more and convert better.

Churn / cancellation feedback → survey, embedded in the flow. One or two questions, asked at the moment of cancellation, analysed in aggregate. Resist turning it into a retention funnel mid-cancellation; rescue offers belong after the honest answer is captured, or the answers become negotiating positions.

Event registration → form, unless tickets require qualification (an invite-only dinner with limited seats), in which case a short funnel that screens before it confirms.

Lead magnet download → form if the magnet is generic (one email field, instant delivery), funnel if the magnet can be personalised ("get the benchmark for your industry" — three questions, then the gate). The personalised version costs the visitor 30 more seconds and typically yields better leads and better engagement, because the asset matches them.

What about the hybrids?

Two patterns that legitimately sit between formats:

  • The scored survey — a survey with scoring and outcomes that shows the respondent a result. Structurally a survey (aggregate analysis, honest-wording rules apply), experientially a quiz. Use when research is the goal but a result raises response rates.
  • The qualifying form — a form with one or two branching questions that route the submission (sales vs support, region, plan). Still a form; the branching serves routing, not persuasion. The moment branches start persuading, you have grown a funnel and should design it as one.

The format can also change as stakes rise: many teams run a plain demo form for organic traffic (high intent) and a funnel for paid campaigns (cold traffic) — same offer, different commitment levels, different tools.

And one boundary case worth naming: the waitlist. Pre-launch signups look like a form (one email field), but if you want the signup to segment your launch list — role, use case, urgency — a three-question micro-funnel costs each visitor twenty seconds and hands you a launch email you can branch. The deciding factor is whether you will act on the segments; collecting them "because we might" is how forms grow barnacles.

The decision, on one card

  1. Visitor already wants it → form. Shortest path wins.
  2. You need aggregate knowledge → survey. Wording and sampling discipline win.
  3. Visitor needs convincing or sorting → funnel. Earned clicks and placed gates win.
  4. Reading responses one-by-one? Not a survey. Long but high-intent? Still a form. Persuading mid-flow? Already a funnel — design it as one.

On Formspring the three are separate products sharing one builder language — forms on every plan, surveys and funnels on Pro — so switching formats when you picked wrong is a rebuild of the screens, not the stack. The guides cover each format's craft in depth.

Related from this desk