Copywriting and CTAs for SaaS flows·

Nobody Signs Up to Answer "What Best Describes Your Role?" — A Copy Teardown of Onboarding Step One

Nobody Signs Up to Answer "What Best Describes Your Role?" — A Copy Teardown of Onboarding Step One

You paid for the click. They paid with their email. And on the very first screen after the account exists, you ask them to fill in an HR field.

Nobody signed up for that. They signed up for the thing your landing page promised. The role picker is the most common onboarding step one and the most common place to lose people who were already sold. Here's why its copy fails and exactly how to rewrite it.

Why the first onboarding question costs more than any other screen

  • It lands right after commitment. Sunk cost is high, delivered value is zero. Any confusion here reads as "this was a mistake."
  • It's a fork. Every option writes a different path through your product, so the labels decide what people see next.
  • It's the one screen they'll actually read. They will not read your help docs. They will read this.
  • It converts intent into identity. "I want a weekly churn report" becomes "I am a Marketer." Identity admin is the least motivating work in software.

A role picker isn't inherently wrong. Asking for it in the language of your org chart, with a shrug for a button, is.

Find the leak before you rewrite a word

Don't take my word for where your drop-off is — your own funnel data will tell you, usually within an afternoon.

  • Step-level funnel, not top-level signup. Compare step-one views → step-one submits → step-two views. High views with low submits is a comprehension problem, not a traffic problem.
  • Time-on-step. Long dwell plus high abandonment means people are stuck deciding. That's a labeling failure, not a motivation failure.
  • Session replays. Watch twenty sessions of people who bailed on this screen. Look for the cursor hovering between two options and leaving. That's a taxonomy problem you can see with your eyes.
  • Option distribution. If the catch-all ("Other," "Just exploring") is one of your top two picks, your labels don't map to how users describe themselves.

If instrumenting this by hand is more work than the fix, run a free audit on your signup flow and get a prioritized list of where the copy is costing you.

The three copy failures in almost every role picker

1. The question is about your database, not their job (clarity). "What best describes your role?" answers a question the user never asked. Replace it with the job: "What do you want to set up first?" Now the question has an obvious reason to exist.

2. The options mix taxonomies (hierarchy). Founder / Marketing / Product / Engineering / Enterprise / Agency is three different axes in one list — seniority, function, company size, business model. Pick one axis and make every option answer the same question. Five options max.

3. The button is a shrug (guidance). "Continue" and "Next" tell the user nothing about what they get. Name the next screen or the payoff. A button is a promise, not a page turn.

Bonus failure: the hidden skip link (trust and friction). If "Skip for now" is 11px grey text in the bottom-left corner, you've told the user the honest exit is shameful. Promote it to a peer option.

Before/after: rewriting the whole screen

Before

Tell us about yourself What best describes your role? [ Founder / CEO ] [ Marketing ] [ Product ] [ Engineering ] [ Other ] [ Continue ] skip

After

What should we set up first? Pick one — we'll preload the template so you're not staring at a blank screen. [ Send a weekly report to my team ] [ Track churn on new signups ] [ Clean up my CRM before launch ] [ Just looking around — show me a demo workspace ] [ Build my first report ]

What changed and why:

  • Clarity: the headline now states the payoff of answering, so the question isn't administrative.
  • Hierarchy: one axis, four options, one of which is an honest exit. No org-chart labels.
  • Guidance: the button names the next screen. Users know what happens when they click.
  • Trust: the skip path is a peer option, not a whispered link. People who take it still get a working product instead of a dead end.

If you genuinely need role data, ask for it later — progressive profiling after the user has had one win is a well-worn pattern and far cheaper than blocking activation.

The P0/P1/P2 fix list

P0 — this week

  • Rewrite the question in outcome language ("What do you want to set up first?").
  • Rewrite the CTA to name the next screen, not "Continue."
  • Make the skip path visible and non-punitive.

P1 — next sprint

  • Collapse to one question per screen. Move role, company size, and team count to a later, optional step.
  • Fire an event per option so you can see which label produces which downstream behavior.

P2 — after you have data

  • Personalize the post-answer headline by option (feedback: confirm what they chose and what's happening next).
  • A/B the option set, not the button color. The labels are the variable that matters.

How to know it worked

  • Step-one submit rate, not overall signup rate. That's the number this screen owns.
  • Option distribution. Fewer catch-all picks means your labels finally fit how users describe themselves.
  • Time-on-step. Dropping dwell time with rising submits means people are deciding, not guessing.
  • Downstream activation by option. If the choice doesn't predict who succeeds, the question isn't earning its place. Cut it. A deleted question is the highest-converting question.

We've torn down several other flow leaks on the blog — same method, different screen.

Start a free FlowAudit at /signup and you'll get a prioritized P0/P1/P2 fix list for your own onboarding flow in minutes — no consultant, no workshop, no forty-slide deck.