Onbrd Docs — Checks
Check Categories
Signup friction
Identifies barriers in the signup process that prevent users from completing registration.
- Primary CTA clarity — Why: reduces hesitation. Fix: action verbs, one dominant color.
- Form bloat — Why: each field adds drop-off. Fix: defer non-essential fields.
- Password friction — Why: complexity blocks. Fix: passkeys/social as options.
Activation path
Evaluates how quickly and effectively users can reach the core value of your product.
- Empty-state guidance — Why: first-run stalls. Fix: checklist or sample data.
- First value time — Why: early win retains. Fix: prefill, templates.
- Progress feedback — Why: reduces uncertainty. Fix: visible steps or tips.
Copy & clarity
Assesses whether your messaging clearly communicates value and guides users through the experience.
- Headline specificity — Why: benefits > slogans. Fix: verb + outcome + time.
- Jargon scan — Why: confuses new users. Fix: plain language, tooltips.
- Error messaging — Why: dead-ends drop users. Fix: actionable next steps.
Momentum
Measures how well your onboarding maintains user engagement and reduces drop-off.
- Distraction audit — Why: leaks attention. Fix: remove non-critical links.
- Celebration moments — Why: reinforce progress. Fix: subtle success states.
- Return cues — Why: drive day-2 use. Fix: "Continue where you left off."
Technical
Checks for technical issues that could impact user experience during onboarding.
- Core Web Vitals baseline — Why: perf affects completion. Fix: image sizing, lazy-load.
- Focus order — Why: keyboard users. Fix: logical tab sequence, visible focus.
- Mobile tap targets — Why: fat-finger errors. Fix: 44px min target.
Integrity & Robustness
Validates that our scoring system is reliable and not overfitting to benchmark data through comprehensive falsification testing.
- Ground truth leakage detection — Why: prevents scoring system from accessing fixture data. Fix: automated validation of probe implementations.
- Label shuffle falsification — Why: tests robustness to random label permutations. Fix: expected score degradation confirms no overfitting.
- DOM perturbation testing — Why: evaluates stability under layout changes. Fix: maintains semantic meaning while varying presentation.
- Holdout validation — Why: ensures unbiased performance estimation. Fix: separate test sets for final validation.