Last verified May 2026 · 7 min read

Methodology

How we source signup drop-off numbers, how the per-field cost calculator turns them into a dollar figure, and the refresh cadence that keeps everything in lockstep.

BENCHMARKS VERIFIED MAY 2026

Primary sources

Twenty primary publishers across academic UX research, vendor-published case data, industry benchmark reports, and product-growth writing. We cite the original publisher in every numeric claim. Where two publishers disagree, we name both numbers.

Source	Cadence	What we take from it
Baymard Institute	Annual + quarterly large-sample studies	Per-field drop-off (8-10pp per extra field), checkout abandonment rates extended to signup forms, mobile vs desktop conversion deltas. Paywalled at roughly $200/month for full database; we cite the public summaries and the figures Baymard themselves publish in conference talks and free reports.
Nielsen Norman Group	Continuous; topic-led	Form usability, error messaging, mobile form input, password UX. Free, citation-friendly, peer-respected. Specific articles cited inline include Jakob Nielsen on form length, Kara Pernice on password rules, and the NN/g mobile form audits.
Luke Wroblewski (LukeW)	Continuous	Inline validation, single-column forms, the Web Form Design book. The reference for form best-practice prescriptions. Quoted for the foundational design patterns; we add the 2026-era drop-off numbers Wroblewski himself does not publish.
Mixpanel Product Benchmarks	Annual report	Cross-industry product conversion rates (web and mobile). Signup conversion ranges by industry vertical (SaaS, fintech, ecommerce, media, dev tools). Public report; segments are broad rather than narrow but the dataset (millions of monthly active users) is large enough to be authoritative.
Amplitude Product Report	Annual	Activation and signup conversion deltas. Amplitude's benchmarks track funnel-step completion; for signup the relevant slice is start-to-complete conversion. Used to triangulate Mixpanel and Pendo figures.
Pendo Product Benchmarks	Annual	B2B SaaS-focused signup and activation rates. Tilted toward enterprise and mid-market where Pendo's customer base concentrates. Helpful when triangulating B2B-specific signup conversion bands.
OpenView SaaS Benchmarks Report	Annual	Free-trial-to-paid and signup-to-trial conversion by ARR band. Where the audit benchmarks for SaaS signup come from when we stratify by company stage (seed, Series A, Series B, growth, public).
Reforge	Continuous	Product-growth and signup-funnel writing from Brian Balfour, Casey Winters, Andrew Chen. Cited for qualitative framing on B2C onboarding patterns and the upper-funnel signup decision (which we then back with cited percentages from Baymard / Mixpanel / Segment).
Lenny Rachitsky / Lenny's Newsletter	Continuous	B2C and B2B product benchmark posts (signup conversion 30-70% typical range; activation 5-30% typical). Lenny aggregates from his network of product leaders; we cite where numbers are attributable to a named source within his posts.
Andrew Chen / a16z	Continuous	User acquisition, growth-loop framing, paid-organic conversion deltas. Quoted for the qualitative argument on signup as a conversion event in a larger acquisition loop, not the per-field percentages.
ChartMogul SaaS Subscription Index	Annual	Trial-to-paid conversion by ARR band. ChartMogul concentrates on the post-signup conversion (trial conversion) but their dataset includes signup-to-trial counts that calibrate the upstream funnel.
FirstMark Capital SaaS Survey	Annual	Growth-stage SaaS funnel metrics (signup, trial, paid). Sample skews to FirstMark's portfolio plus respondents; helpful as a third triangulation point alongside OpenView and KeyBanc.
KISSmetrics archive (Neil Patel)	Archive (active 2010-2018, archived since)	Foundational conversion-rate-optimisation data including A/B test case studies of signup form changes. Historical reference rather than a live source; we use only the case studies still verifiable against the original test screenshots.
Baymard / Statista form abandonment data	Quarterly	Cross-industry form abandonment percentages. Used to calibrate the calculator's defaults for unknown-vertical inputs.
Userpilot	Continuous	Product-tour, onboarding-flow, and signup-to-activation benchmark blog posts. Vendor-published, so we cross-check Userpilot numbers against Mixpanel or Pendo before citing.
Heap	Continuous	Drop-off analysis methodology and case studies. Helpful for the funnel-diagnostic framing on the calculator and benchmarks pages.
Segment SaaS benchmark report	Annual (State of Personalization Report)	OAuth conversion lift (+15-25pp B2C / +8-15pp B2B from the 2023 report). The most-cited single number on social-login conversion across the literature.
Auth0 (Okta) case data	Continuous	Magic link conversion lift (+15-30pp), passkey rollout case studies, email verification gate impact. Vendor-flavoured; we use Auth0 numbers where they are corroborated by independent sources (Slack disclosed magic-link numbers, Notion blog posts).
Cloudflare Turnstile / Google reCAPTCHA published benchmarks	As-published per release	Captcha friction cost numbers (reCAPTCHA v2 checkbox 2-5pp, Turnstile sub-1pp). Vendor-published; we use the conservative end of the published range and cross-check against independent CRO blogs.
NIST SP 800-63B Digital Identity Guidelines	Periodic (current revision 3)	Authoritative US government guidance on password rules: length over complexity, no forced rotation, no forced special characters. The single most-ignored research in signup UX; we cite the actual paragraph numbers, not a vendor summary.

In scope

The signup layer specifically: the form, the auth method, the verification step, the password rules, the captcha, the country / language defaults, and the mobile vs desktop adaptations.
Per-field drop-off math and the dollar cost of one extra field at named LTV and CAC bands.
OAuth (Google, Apple, GitHub, Microsoft) conversion lift over email plus password baselines.
Magic link and passkey conversion data and the speed-of-repeat-login trade-off.
Email verification gate strategies (hard gate, soft verify, verify-later, delay-until-action).
B2B vs B2C signup form-length norms, including the PLG B2B exception.
Industry-stratified benchmark bands: SaaS, ecommerce, fintech, healthtech, dev tools.
Mobile vs desktop signup conversion deltas, including the autocomplete and keyboard-type contribution.
International signup pitfalls: country dropdown defaults, phone formats, postcode validation.

Out of scope

Onboarding and activation after signup (covered briefly on the home page but not the site's focus).
Per-company contract pricing for auth vendors. The /auth-stack-comparison page covers public pricing only.
Checkout abandonment in ecommerce, which is Baymard's primary remit and is its own discipline.
SAML / SCIM enterprise SSO provisioning. We cover it briefly on /oauth-vs-email and /b2b-signup-form-fields; the deeper guidance lives at enterprise IAM publishers.
Anti-fraud and KYC verification beyond the conversion-cost framing on /b2b-vs-b2c-norms.
PCI compliance and other regulatory regimes that drive friction; we name them as legitimate reasons for added friction but do not write the compliance guidance.

Calculation framework

Six calculator rules, each tied to a named source band. The calculator at /calculator surfaces these explicitly in worked examples.

Per-field cost

annual_starts * 0.08 * LTV for each field beyond the second. The 0.08 is the conservative end of Baymard's aggregated checkout-form drop-off range (typically reported as 8-10pp per extra field). For B2B SaaS at $1500 LTV, two extra fields cost 16% * annual_starts * $1500 in lost LTV before any CAC waste is added back in.

OAuth lift attribution

+15-25pp baseline for B2C, +8-15pp for B2B per Segment 2023. We apply the midpoint when toggled on. The calculator does not double-count: if OAuth is toggled and email-verify is also toggled, OAuth's implicit verification absorbs the verify gate's drop.

Magic link / passkey adjustment

+15-30pp signup conversion lift over email plus password per Auth0 case data and Slack / Notion disclosed numbers. The calculator's default uses the midpoint (+22pp). Trade-off note: magic links slow repeat login, which the activation funnel pays for, not the signup funnel.

Email verify gate adjustment

Hard gate before product access: 8-20pp drop (typically modelled as 12pp). Verify-later or delay-until-action gates: 2-5pp drop. Soft verify (banner inside the product): roughly zero. Auth0 and Segment case data converge here.

Captcha friction adjustment

reCAPTCHA v2 checkbox: 2-5pp drop on legitimate signups. reCAPTCHA v3 invisible: 0.5-1.5pp drop. Cloudflare Turnstile: under 1pp drop. hCaptcha: 1.5-3pp drop. Numbers from Cloudflare and Google published benchmarks plus independent CRO blog cross-checks.

Mobile vs desktop adjustment

Mobile signup conversion runs 5-15pp below desktop per Baymard and NN/g mobile form research. The calculator does not split the result by device by default; the /mobile-vs-desktop sub-page covers the device-stratified math.

Refresh cadence

First business week of each month, every dated claim is re-checked against its source. The label changes in one place (src/lib/schema.ts: LAST_VERIFIED_DATE and LAST_VERIFIED_LABEL), which propagates to every visible Updated stamp, every Article schema dateModified, every WebSite schema dateModified, and every per-page Last verified line across 24 routes. No drift between visible date and structured-data date.

Out-of-cycle refresh triggers:

Annual benchmark release from Mixpanel, Amplitude, Pendo, OpenView, ChartMogul, FirstMark, or Segment.
Material auth vendor pricing or product change (Auth0, Clerk, Stytch, Supabase Auth).
NIST SP 800-63B revision update.
New case study publication from a named company on signup flow numbers (Slack, Notion, Dropbox, Stripe, Shopify, Figma, Airbnb).
Reader-submitted correction with verifiable source.

Limitations

Benchmark ranges are central tendencies, not predictions. A specific product may run higher or lower than the cited bands depending on category, market maturity, and user intent.
Vendor-published numbers (Auth0, Cloudflare, Userpilot, Heap) are vendor-flavoured. We use them only where cross-checkable against independent sources, and we use the conservative end of the published range.
Calculator outputs are estimates, not guarantees. The right way to use the dollar figure is to support the decision to A/B test a change, not to forecast the lift you will see.
OAuth lift, magic link lift, and email verify gate drops do not stack linearly. The calculator caps the combined adjustment to keep results inside realistic ranges; the explicit math is shown on /calculator.
Some primary sources are paywalled (Baymard Institute). Where we cite a paywalled source, the underlying number is one Baymard or the relevant publisher has shared publicly in conference talks, free reports, or interview transcripts.

Corrections

Email [email protected] with the page, the cited number, and the verifiable source. We aim for a five-business-day turnaround and acknowledge the fix in-page with the date.