A "Get started" button promises the product. We wanted to know how often it delivers. On November 14, 2026 we clicked the main self-serve button on 169 B2B SaaS homepages and recorded where it led: 62% opened a real sign-up form, but 18% opened a demo or "contact sales" form instead. For buttons labelled just "Get started", that number was 40%.
Cueston helps people get through a product without getting stuck, so the first minutes after "sign up" are our home ground. The method is below and the full dataset is downloadable.
How we tested the sign-up flows
We started from the same random sample of Y Combinator B2B and SaaS companies as our after-hours study. On each homepage a headless Chrome browser found the first self-serve button ("Sign up", "Start for free", "Try for free", "Get started" and similar), opened its link and recorded the page it landed on. 194 homepages had such a button, and 169 of those pages loaded. We did not create any accounts. For each page we recorded:
- Where it landed: a sign-up form, a demo or contact form, or somewhere else.
- The visible form fields on the first screen, ignoring site navigation, footer and search boxes.
- Sign-in options: Google, Microsoft, GitHub, Apple, SSO/SAML, magic links and passwords. Social links in the header or footer were excluded.
- Whether a CAPTCHA loaded, and which sign-up page asked for a phone number or company name.
We counted a page as a demo or sales form when its address said demo, contact, request or get-started and it had no password field or social sign-in. We then checked each one by its page title ("Book a Demo", "Contact Us", "Request a Trial").
"Get started" often means "talk to sales"
The label decides a lot. Of 70 buttons that said just "Get started", 28 opened a demo or sales form. Of 99 buttons that said "Sign up", "Start for free", "Try for free" or similar, 3 did. A visitor who clicks "Get started" expecting a product and gets a 7-field sales form has been misled, and many will leave rather than fill it in.
If your product needs a sales call, that is fine. Just say so on the button: "Book a demo" sets the right expectation, and "Start free" should mean free.
How many fields do SaaS sign-up forms ask for?
On the 105 real sign-up pages, the median first screen has 2 fields. 36% ask for one field, almost always the email, and 11% have no fields at all, only "Continue with Google" and similar buttons. At the other end, 25% ask for 4 or more fields on the first screen.
- 43% ask for a password on the first screen, and 9% ask for it twice.
- 30% ask for a name, 10% for a company name and 6% for a phone number before the account exists.
- 31% load a CAPTCHA script such as reCAPTCHA, hCaptcha or Cloudflare Turnstile.
- 58% of sign-ups happen on a separate subdomain such as app. or auth., often run by an auth provider like Auth0 or Clerk.
Which sign-in options do SaaS products offer?
60% of the sign-up pages offer at least one social sign-in. Google is by far the most common at 54%, then GitHub (13%, mostly developer tools), SSO/SAML (12%), Microsoft (9%) and Apple (9%). 19% offer a magic link or one-time email code instead of a password.
Microsoft sign-in at 9% stands out as low for B2B products. Many companies run on Microsoft 365, and those buyers can't use "Continue with Google" with their work account.
How to fix your SaaS sign-up flow
- Click your own "Get started" button in a private window. If it opens a sales form, rename the button.
- Ask for the email (or a Google or Microsoft account) first, and everything else after the account exists.
- If you sell to companies, add Microsoft next to Google.
- Drop "confirm password" and offer a magic link instead.
- After sign-up, the hard part starts: an empty product. Show the next step inside the app, not in an email the next day.
Limitations of this study
The sample is YC companies tagged B2B or SaaS, so it leans towards younger, US-based startups. We followed only the first self-serve button on each homepage and only read the first screen; many flows ask more questions later. 11% of sign-up pages loaded their form in a way we could not read (inside a frame or after a delay), and they are left out of the field counts. Sign-in buttons were matched by their text in English. The full results are available as a CSV, free to reuse with a link back to this page.
Frequently asked questions
How many fields should a SaaS sign-up form have?▾
In our test of 105 B2B SaaS sign-up pages in November 2026, the median first screen asked for 2 fields, and 36% asked only for an email. Asking for one field (or a Google or Microsoft account) and collecting the rest later is the most common pattern.
What share of SaaS products offer Google sign-in?▾
54% of the 105 sign-up pages we tested offered Google sign-in. GitHub was offered by 13%, SSO/SAML by 12%, and Microsoft and Apple by 9% each.
Does "Get started" lead to a sign-up?▾
Often not. 40% of the 70 homepage buttons labelled "Get started" opened a demo or contact-sales form, against 3% of buttons labelled "Sign up" or "Start for free".
Getting someone to sign up is half the job. The other half is the first session, when a new user lands in an empty product and has to work out what to do. Cueston adds an AI guide inside your app that shows each new user the next step in your real interface, so fewer of them sign up and never come back.