Which signup fields should I never require at signup?

Never require at signup: SSN or national ID (unless legally required), address (unless delivering to it), required marketing opt-in, confirm-password (replace with visibility toggle). Ask later via progressive profiling: phone, company, role, country (unless needed for tax/compliance), postcode.

Last verified April 2026 · 6 min read

The field-by-field drop-off table

Q: How much does each form field reduce signup conversion?

Varies by field. The Baymard rule of thumb is approximately 8-10pp per additional field beyond the minimum. But the impact varies: required phone costs 3-12pp (Formisimo/Zuko), SSN costs 20-50pp (NN/g), confirm password costs 3-7pp (NN/g, LukeW). The table on this page gives per-field figures with sources for 15 common fields.

Q: What is the minimum viable signup form?

For B2C: email only (with OAuth as primary option) or email plus password. For B2B PLG: email plus password plus optional role/company. For B2B sales-led: email plus company plus role (lead qualification is a different form than product signup). Never more than 3-4 fields without an explicit justification for each additional one.

The rule of thumb that each field costs ten percentage points is a useful shorthand, but it is a simplification. Different fields carry very different drop-off costs. Email is the irreducible baseline. An SSN field in a non-regulated context is catastrophic. A first-name field is barely measurable.

This is the only open-web version of the field-by-field drop-off table with every row sourced. Baymard has the deeper data behind a paywall. This aggregates the publicly available figures from Baymard, HubSpot, Formisimo/Zuko, and NN/g in one reference.

FIELD	DROP-OFF	SOURCE	NOTES	ALTERNATIVE
Email	baseline	Required baseline field	The irreducible minimum. OAuth removes it entirely.	OAuth / magic link
Password	5-10%	Baymard Institute	Higher if complexity rules are enforced (NIST-violating rules add friction).	Magic link, passkey
Confirm password	3-7%	NN/g, Luke Wroblewski	Zero-value field. Users retype the same typo. Replace with visibility toggle.	Password visibility toggle
First name	1-3%	HubSpot form benchmarks	Low friction, low value at signup. Personalisation rarely justified.	Ask later (progressive)
Last name	1-3%	HubSpot form benchmarks	Often paired with first. Same low-friction, low-value trade-off.	Ask later (progressive)
Phone number (required)	3-12%	Formisimo/Zuko field analytics	High perceived privacy risk. Wide range depends on product context (delivery vs SaaS).	Ask later, or SMS optional
Company name	2-5%	Formisimo/Zuko	B2B-justifiable. Can often be enriched from email domain automatically.	Enrich from email domain
Role / job title	2-4%	HubSpot form benchmarks	B2B-justifiable. Use a dropdown with an escape hatch for unusual roles.	Dropdown with 'Other'
Country dropdown	1-3%	Baymard Institute	Higher if default is USA for non-US users. IP-based defaulting fixes most of this.	IP-defaulted with override
Postcode / zip code	2-6%	Baymard Institute	High friction on mobile. Alphanumeric (UK) breaks USA-zip validators.	Ask later (address context only)
Full address	5-15%	Baymard Institute	Signup-stopper for B2C where delivery is not the core product.	Progressive / checkout only
SSN / national ID	20-50%	NN/g, Baymard Institute	Catastrophic in non-regulated contexts. Triggers identity-theft fear.	Only if legally required
Date of birth	3-8%	Formisimo/Zuko	Age-gate required contexts only. Age checkbox plus later verify is lighter.	Age checkbox + verify later
Captcha (reCAPTCHA v2)	2-5%	HubSpot, Baymard Institute	reCAPTCHA v2 checkbox. Invisible variants (v3, Turnstile) are near-zero.	Turnstile, reCAPTCHA v3
Terms checkbox (required)	1-3%	NN/g signup research	Dense text amplifies it. Summary plus link to full reduces it.	Summary + link to full terms
Marketing opt-in (required)	5-15%	HubSpot form benchmarks	Never make it required. Unchecked optional with clear copy is the standard.	Optional, unchecked by default

Note: Drop-off figures are research-aggregated ranges. Your funnel will vary by product type, audience, and context. Use these as priors for A/B tests, not as absolute predictions.

How to use this table in a design review

Bring the table. Point at the row. Ask: what happens if we do not include this field? The field's advocate will give you their reason. That reason may be legitimate. But now it is explicit, costed, and on the record. Most fields cannot survive that scrutiny.

The question to ask about every required field: what legitimate product or legal purpose does this field serve, and when is the earliest we could collect it inside the product rather than at signup? For most fields, the honest answer is: later, in context, once the user has seen value.

The minimum viable signup form for B2C is email only (or OAuth). For B2B PLG products, add a role dropdown if personalisation genuinely uses it. For B2B sales-led, the form is a lead qualification questionnaire and the rules are different entirely - see B2B vs B2C norms.

Frequently asked questions

How much does each form field reduce signup conversion?+

Varies by field. Required phone costs 3-12pp (Formisimo/Zuko), SSN costs 20-50pp (NN/g), confirm-password 3-7pp (NN/g, LukeW). The Baymard rule of thumb is 8-10pp per extra field on average, but field type matters significantly.

Which fields should never be required at signup?+

Never require: SSN/national ID (unless legally mandated), full address (unless core to the product), required marketing opt-in, confirm-password field. Ask later: phone, company, role, country, postcode.

What is the minimum viable signup form?+

B2C: email only, or OAuth. B2B PLG: email + password + optional role/company. B2B sales-led: email + company + role for qualification. Never more than 3-4 fields without an explicit justification.

DIGITAL SIGNET CONSULTING

Need an outside eye on your signup funnel?

Digital Signet runs two-week signup-funnel audits. We map the drop-off at every step, attach the dollar cost to each field, and deliver the memo your growth team will sign.

See the engagement format