A redesigned onboarding that took sign-up completion from 1% to 70%

Three screens of the new VIDI onboarding: welcome, “Let’s get started” with Continue with Google, and the sign-up success screen
Intro

VIDI pays people to work as mystery shoppers and store auditors: a side gig, like driving for Uber, but done inside real stores for Brazilian brands.

The app had no trouble getting downloaded. The trouble was what came next. Out of roughly 1,000 daily downloads, about 10 people finished signing up. The first screen alone lost 80% of them: a long, scrolling form with 10 fields, followed by terms, extra personal questions and a selfie.

But we couldn’t just delete the form. VIDI runs on trust: brands pay for real visits from real people, so we genuinely needed identity data to prevent fraud.

The challenge was asking for a lot, without it feeling like a lot.

01

Problems

Sign-up asked for everything upfront, in one long form. 99% of downloads never became accounts.

02

Solution

Phased the process: essential data first, broken into short steps that explain why each item is needed. Everything else moved to post-sign-up profile completion. Added Google sign-in.

03

Results

Daily sign-ups went from ~10 to ~700: completion grew from 1% to 70% of downloads.

Setting the scene

The drop-off wasn’t subtle.

Firebase showed the gap between downloads and completed sign-ups. Clarity showed where it happened: most people opened the first screen, scrolled through the form and left.

The old flow had four stages:

  1. 01A scrolling form with 10 required fields80% gave up here
  2. 02Terms of use
  3. 03Extra profile information (age, marital status and similar)skippable
  4. 04A selfieskippable

Stages 3 and 4 were skippable, but almost nobody got that far.

Firebase

daily downloads vs. completed sign-ups

Daily downloads vs. completed sign-ups (Firebase): people who reached each stage of the old sign-up
StagePeopleChange from previous stage
01 · A scrolling form with 10 required fields12,480—
02 · Terms of use4,720−62%
03 · Extra profile information1,040−78%
04 · A selfie300−71%
Design challenge

Who is downloading the app?

Many new users are just curious. They’ve heard they can earn extra money and want to see what it’s about. They haven’t decided to commit, and a 10-field form asks for commitment before giving anything back.

VIDI, on the other hand, can’t let people in with no identity checks. One fraudulent visit can compromise a major brand’s data.

Product decision

Not everything had to happen at sign-up.

I mapped every field in the old flow and split it into two groups with the team:

Essential to enter

Data that’s legally or operationally required before someone can register.

CPF (Brazil’s tax ID)Address

Can wait

Data that helps us match people to opportunities or allow them to take surveys.

AgeMarital statusOther demographicsProfile selfie

The second group moved out of sign-up entirely and became part of a new profile-completion experience.

See profile area case
Google sign-in

Some of the work was already done for us.

Adding Google sign-in did two things at once:

  • Returning users who had registered with their Gmail got straight in.
  • New users skipped name, email and password entirely and jumped to the steps Google couldn’t fill.
Screen recording of the Google sign-in: choosing an account on the login screen and landing on the home screen
Step by step

With a reason for each.

The essential data stayed, but the experience around it changed:

Sign-up split into short screens: e-mail and phone first, then “Where do you live?” with the ZIP code1

One topic per step instead of one long scroll.

CPF field with a “Why do we ask for this?” link right below it2

A short access point to the explanation of why we ask for things like CPF, so sensitive data feels justified, not invasive.

“One last step” screen with a three-part progress bar fully filled3

Clear progress, so users know how close they are.

What almost happened

I wanted to go further.

My original proposal was an Airbnb-style social-only sign-up: get in with Google, browse surveys, and complete your profile only when there’s something worth doing.

This addresses one of VIDI’s core pains. We can’t guarantee surveys in every town, so someone in a small rural city could complete the whole sign-up only to find nothing available.

It wasn’t technically possible at the time without reworking core parts of the backend, so we shipped the phased version. It remains the next step I’d push for.

Results
1%→70%

Sign-up completion, measured as completed sign-ups per daily download.

~10→~700

New accounts a day.

70×

More sign-ups, while still collecting the identity data the business depends on to operate.

Learnings

Shortening a form is not the only answer. Asking at the right moment, with a reason, works too.

We still ask for sensitive data. We just stopped asking for all of it at once and started explaining why.

Next project

A liveness check that eliminated identity fraud

Adding an extra layer of security without turning verification into a frustrating extra step.

Liveness check screens