Back to decisions

ONBOARDING EXPERIMENT · ELVA

Let users reach value before explaining everything.

Role at the company: Head of Product · Elva · 2026

Shorten an unfamiliar AI journey and measure activation without weakening downstream behaviour.

Selected result: +11% · progression to gallery access

Elva originally used three post-paywall tutorial screens to explain an unfamiliar AI creation flow. I suspected that the product was asking users to understand too much before letting them do anything useful.

We consolidated the three screens into one and tested the shorter journey, using progression to gallery access as the activation metric and retention, export, upload and cancellations as downstream guardrails.

A / Original onboarding
  1. Original onboarding · 1 / 5
    01
  2. Original onboarding · 2 / 5
    02
  3. Original onboarding · 3 / 5
    03
  4. Original onboarding · 4 / 5
    04
  5. Original onboarding · 5 / 5
    05
B / Shortened onboarding
Shortened onboarding
01
progression to gallery access
+11%
4–7 day retention
7.5% → 9.6%
retention uplift
+29% relative

No monitored guardrail deteriorated, and the winning flow was rolled out to all users.

OPEN TO PRODUCT ROLES & CONSULTING

Let’s build a product people choose to return to.