Claim
The default growth instinct is to remove friction. The right rule is to remove friction that doesn't add to the user's understanding of fit, and add friction that does. Mercury's onboarding rebuild added explicit qualifying questions and became the highest-impact growth project of the quarter despite increasing the number of steps.
Mechanism
Frictionless onboarding maximizes signups but doesn't maximize activation. Users who could not have succeeded on the product enter the funnel anyway, lowering activation rates and burning support time. Qualifying friction at the front filters in the right users and trains them on what the product is for. Activation per signup rises faster than total signups fall, net activation lifts.
Conditions
Holds when:
- The product has a clear ICP and friction can be designed to identify it.
- The added friction connects to the user's own decision (vertical, use case, scale), not gating revenue (paywall, sales call).
Fails when:
- The product is broad enough that any friction excludes valid users.
- The friction is bureaucratic (forms with no learning value). Users sense the difference.
Evidence
"Cut friction when it doesn't add to a user's experience of why the product is for them. Add friction when it helps users understand whether the product is for them."
Mercury's onboarding rebuild — a whole quarter of fixing onboarding quality (not optimization) — became their highest-impact growth project. Anthropic's onboarding asks who-you-are / what-you're-here-for to recommend the right products and features.
— Amole Naik on Lenny's Podcast, 2026-04-27
Signals
- Activation-per-signup rises despite total signups holding flat or falling.
- Support tickets shift from "how do I use this" to "advanced workflows."
- Users self-describe their use case in the funnel and the system routes them differently.
Counter-evidence
Bret Taylor and others argue outcomes-based pricing and zero-friction trials are the dominant pattern in commodity SaaS. There, any front-end friction loses to a competitor offering instant access. Apply Amole's rule when the product genuinely serves narrow ICPs; don't apply it when the product is generic and the value is speed-to-first-action.
Cross-references
- Brand and quality are growth levers, not constraints on growth — the broader thesis Amole makes that brand and quality drive growth, not slow it