When a Multistep Landing Page Form Beats One Short Form
When a multistep form beats one short form, per GOV.UK, NN/g and Baymard, what HubSpot and Unbounce field data measured, and a test that tracks lead quality.
On this page
- Field count moves completion more than step count
- How many fields can a form carry before conversions drop?
- What the HubSpot and Unbounce figures did and did not measure
- Do progress indicators lift conversion?
- Audit every field against the step that uses it
- A multistep layout for the same form (hypothetical)
- Test lead quality alongside completions
- Sources
The best practices that make multi-step forms increase completion rate start with an unglamorous rule: cut fields first, then decide whether the survivors belong on one screen or several. After reading the guidance from GOV.UK, Nielsen Norman Group and Baymard, our position is that steps earn their extra clicks when later questions depend on earlier answers. Splitting a form only so it looks shorter adds clicks and removes no work. We also think most landing page form tests are scored on the wrong number, because a count of submissions says nothing about whether sales wanted them.
Field count moves completion more than step count
The research on multi-step form UX, completion rates and progressive disclosure from Baymard, Nielsen Norman Group and GOV.UK agrees on more than the CRO folklore suggests. In its 2024 ecommerce benchmark, Baymard found that the average checkout had 11.3 form fields spread over 5.1 steps, and concluded that "the number of form fields in a checkout impacts overall usability far more than the number of steps." The same article says "most sites need only 8 form fields in total for a checkout flow," and that usability "does seem to suffer after 8+ checkout steps." A checkout is a different job from a lead form, but the lesson carries over well: count the fields before you count the steps.
Nielsen Norman Group makes the case for wizards in a 2017 article by Raluca Budiu. "By showing less information at a time, they allow users to focus better on the content pertinent to each step," and "When they include branching based on user input, wizards ensure that people will see only the steps applicable to their situation." The same piece lists the costs: wizards "can require a higher interaction cost (more clicks) than other input patterns," and "Wizards are not gracefully interruptible." Budiu's advice is to "Use wizards for novice users or infrequent processes (e.g., configuration or setup)." A first-time visitor arriving from a paid ad fits that description closely.
The GOV.UK Service Manual goes further and tells teams to "Start by splitting the form across multiple pages with each page containing just one thing." It pairs that with a question protocol, which "forces you (and your organisation) to question why you're asking users for each item of information." Then it orders whatever survives, starting "with questions that will let users know if they're not eligible for the service" and using "'branching' questions so people only have to answer questions that are relevant to them." It also leaves room to recombine pages, because "User research will tell you when you can merge pages together."
How many fields can a form carry before conversions drop?
How many form fields a form can carry before conversion rate drops depends more on the kind of field than on the count, at least in the HubSpot data people usually cite. Dan Zarrella analyzed "over 40,000 landing pages from our customers" for HubSpot and wrote that "as the number of form fields increases, conversion rates decrease slightly, but not as steeply as I expected." He found "very little decrease in conversion rate as the number of single line text fields increases," while "multiple textareas have a powerful depressing effect on conversion rates" and "the presence of multiple drop down fields on a landing page tends to be associated with lower conversion rates."
So the answer from that dataset is a gentle slope with no field count the post singles out as a cliff, and the clearest penalties attach to multiple open text boxes and multiple drop-downs.
What the HubSpot and Unbounce figures did and did not measure
The HubSpot lead gen landing page data on form fields and conversion rate is a correlation across many customers' pages, each with its own offer, audience and traffic source. The post describes no controlled test, the current page shows only a June 2025 "Updated" date and no original publish date, and the only outcome it reports is conversion rate. Whether the extra fields brought better leads sits outside what it measured.
The older form-length figures that circulate in CRO posts are thinner still. A 2013 Unbounce post by Oli Gardner passes on single-company results from an infographic it credits to QuickSprout, with no sample size, time period or method, so we leave those numbers out. Unbounce's own 2021 Conversion Benchmark Report is sturdier, drawing on more than 44,000 customer landing pages with conversion data from November 2019 to November 2020. On SaaS pages, most forms had four to six fields, yet Unbounce saw "higher conversion rates when forms are either super short (one to two fields) or kinda long (seven to nine)." It also reports conversion rate alone. Read all of these figures as reasons to test, and treat none as a benchmark for your own form.
Do progress indicators lift conversion?
On multi-step form UX design, progress indicators and conversion, GOV.UK's guidance runs against the usual advice. The GOV.UK Design System says: "Start by testing your form without a progress indicator to see if it's simple enough that users do not need one," and "Try improving the order, type or number of questions before adding a progress indicator." If people still struggle, it suggests "a simple step or question indicator," with the warning to "Only include the total number of questions if you can do so reliably." It warns against step lists that show every question at once and let users jump back, and reports that "A number of GOV.UK services have removed this style of progress indicator without any negative effects."
One of those removals is written up. In a 2014 post on the GOV.UK user research blog, Ben Holliday describes the Carer's Allowance team testing the hypothesis that "removing the progress indicator will not affect completion rates," after noticing that "Some people were intimidated by the number of sections they needed to complete." The result: "Completion rates stayed the same after we removed the progress indicator." Average time to completion held, and total online claims rose slightly, which the team said "wasn't a significant increase."
Nielsen Norman Group leans the other way for wizards, advising designers to explain "how many steps are involved" and noting that, especially with complicated processes, "a clear indicator of the current step can help keep users oriented." We read the two together. A two or three step landing page form rarely needs a numbered bar, and a longer flow with a fixed number of steps benefits from a plain "Step 3 of 5." When branching changes the total, name the current step and leave the count off.
Audit every field against the step that uses it
The optimal number of form fields for a lead gen landing page is the number the next step needs, and the quickest way to find it is to ask, for each field, who uses the answer and when. The table below is an example audit of a typical nine-field demo request form; your rows will differ.
| Field | Who uses it next | Needed before the first reply? | Decision |
|---|---|---|---|
| Work email | Sales and the confirmation email | Yes | Keep, step 1 |
| Company size | Routing between sales and self-serve | Yes | Keep, step 2, as buttons |
| Main use case | Preparing the first call | Helpful | Keep, step 2, single choice |
| Phone | A same-day call | Only if you call the same day | Optional, step 3 |
| Budget range | Qualification | Only if you turn leads away below a floor | Step 3, as ranges |
| Job title | Lead scoring | Rarely | Ask on the call |
| Full name | Personalizing the reply | Helpful | Keep, step 3 |
| "Tell us about your project" | Discovery | No | Cut, since HubSpot's data flags textareas |
| "How did you hear about us?" | Marketing attribution | No | Cut and rely on UTM parameters |
A multistep layout for the same form (hypothetical)
This layout is a hypothetical example built from that audit for a B2B demo request page that receives paid search traffic.
- Step 1: work email and a "Continue" button. GOV.UK labels its button "Continue" and aligns it left so users do not miss it.
- Step 2: company size as four buttons, plus main use case as a single choice. A visitor who picks the smallest size band goes to a self-serve booking page and skips step 3.
- Step 3, larger companies only: full name, optional phone, budget range.
Every step after the first gets a Back link at the top, because the Design System notes that "Some users do not trust browser back buttons when they're entering data," and it adds that the browser back button must still take people to the previous page they were on, in the state they last saw it. The header names the current step and shows no total, since branching changes how many steps a visitor sees. The heaviest path shows six fields, the self-serve path shows three, and three fields left the form entirely. If you plan to follow up on someone who stopped after step 1, confirm first that your privacy notice and consent wording cover it.
Test lead quality alongside completions
Google's own lead forms make the trade-off explicit. In Google Ads Help, the "More volume" option "uses a lead form design with fewer steps to submit," which "may result in more leads. However, they may include people who are less interested in your product or service." The "More qualified" option "uses a lead form design that contains more steps," which "may result in fewer leads, but those leads generally have a greater interest in your business, product or service." For both options, "the fields in your lead form won't be changed," so the two options differ in steps while the fields stay the same. We read that as Google treating the number of steps as a quality lever in its own right.
That is why a form test needs a quality column. The worked example below uses invented numbers, labeled as assumptions, to show how the verdict can flip.
| Hypothetical test, same traffic | Single form, 9 fields | Multistep, up to 6 fields |
|---|---|---|
| Visitors (assumed) | 2,000 | 2,000 |
| Submissions (assumed) | 160 (8%) | 140 (7%) |
| Sales-accepted leads (assumed) | 40 (25%) | 49 (35%) |
| Ad spend (assumed) | $10,000 | $10,000 |
| Cost per lead | $62.50 | $71.43 |
| Cost per sales-accepted lead | $250.00 | $204.08 |
Scored on completions, the single form wins. Scored on cost per sales-accepted lead, the multistep version wins by about $46 a lead. Agree on the deciding metric before launch, and keep the test open until sales has graded the leads, because the grading lags the submissions.
Cost per sales-accepted lead feeds straight into customer acquisition cost, and closed revenue from those leads is what your ROAS should be built on; the ROAS calculator shows how a change in lead quality moves the break-even. A randomized split on the landing page also answers a causal question for the form itself, which platform attribution cannot, and our piece on incrementality covers the same logic at campaign level. The quality column matters more as budgets grow, since lead quality can shift as spend reaches new audiences, a point we cover in scaling paid ads without killing ROAS.
If you run paid traffic into forms, our CRO playbook and paid media team work from the same rule set out here. Cut every field the next step will not use, add steps only where an answer changes what comes next, and judge the result on the leads sales accepts. We would rather lose a little completion rate than buy cheaper leads nobody calls back, and we are happy to hear the case for the opposite.
Sources
- Baymard Institute, Checkout Optimization: 5 Ways to Minimize Form Fields in Checkout
- Nielsen Norman Group, Wizards: Definition and Design Recommendations
- GOV.UK Service Manual, Structuring forms
- GOV.UK Design System, Question pages
- GOV.UK User research blog, 6 case studies: using research and data to improve a live service
- HubSpot, Which Types of Form Fields Lower Landing Page Conversions?
- Unbounce, How To Optimize Contact Forms For Conversions
- Unbounce, Conversion Benchmark Report: SaaS
- Unbounce, Conversion Benchmark Report: Methodology
- Google Ads Help, How to use lead forms in responsive search ads and campaigns
