Rebuild the Flow: From a 12-Field Form to a Seeded Care Match
Objective: Given Care.com's current 12-field caregiver-search onboarding form, rebuild the first-run experience using a two-question intent picker and pre-seeded sample matches, applying the lesson's Step 4 and Step 5 directly to a real information architecture problem.
You're the growth PM at Care.com, the online marketplace for finding child, senior, and pet care. New users currently fill out 12 profile fields before seeing a single caregiver match.
Replace the 12-field intake form's blocking position with a 2-question intent picker that routes to a pre-seeded results screen, keeping the detailed fields but moving them after the first match is shown.
Before you start
What you'll need
Free path (everything below is enough to finish)
Enough to plan the rebuild before any engineering ticket is written
Paid upgrades (optional, faster/deeper)
Validates the rebuild plan against real behavior instead of assumption alone
The process
2 steps
Step 01 of 02
The lesson's Step 4 says a two-question intent picker lets you route users to completely different first experiences, since a solo user and an enterprise admin don't need the same first three screens.
Care.com serves parents needing childcare, adults needing senior care, and pet owners needing pet sitters, three audiences with almost no field overlap. What two questions replace the first 6 of the 12 fields?
Procedure
- Question 1: 'What kind of care are you looking for?' -> Child / Senior / Pet (replaces 3 category-specific fields collected identically today regardless of answer)
- Question 2: 'When do you need this care?' -> Right away / Within a month / Just researching (replaces urgency and scheduling fields, and determines whether to show live availability first)
- Map each of the remaining 10 original fields to whether it's still needed pre-match (schedule, location) or can move post-match (payment details, background-check preference, backup care needs)
- Result: 2 questions replace 6 fields pre-match, the other 6 move to a post-match profile step
BEFORE: 12 fields, 0 matches shown until form complete AFTER: 2 questions (care type, urgency) -> matches shown -> 4 remaining pre-match fields (zip code, schedule) -> matches refresh -> 6 fields deferred to post-match profile completion
Healthy
A user sees real caregiver cards within 2 questions and a zip code.
Unhealthy
A user fills 12 fields before seeing whether any caregivers are even available in their area.
What this means
Most of the original form was collecting profile-completeness data the business wants, not data the match algorithm needs before showing a first result.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Zero caregiver cards shown until form completion | Query available caregivers after question 2 + zip code only | dev ticket |
Step 02 of 02
The lesson's Step 5 says a pre-filled workspace demonstrates value while a blank canvas demands work; pre-populating with sample content shifts the user's job from 'figure out how to start' to 'see if this works for me.'
Once the intent picker routes a parent needing childcare in a given zip code, what should the very first results screen show if the real-time caregiver query is still loading or returns zero exact matches?
Procedure
- Loading state: show 3 anonymized sample caregiver cards from the nearest metro area, labeled 'Example matches while we search your area', never a blank spinner
- Zero-match state: show the same 3 sample cards with a 'expand your radius' prompt instead of a dead end
- Populated state: real matches replace the samples within 2 seconds in the common case
- Every sample card links to 'How matching works' instead of a real profile, so it can never be mistaken for an actual available caregiver
RESULTS SCREEN (loading, 0-2s) Example matches while we search your area: - Maria G. -- 4.9 stars, 6 yrs experience, CPR certified (sample) - James T. -- 4.8 stars, background-checked (sample) - Priya K. -- 4.9 stars, infant care specialist (sample) [Expanding your search...]
Healthy
New users see what a good match looks like within 2 seconds, real or sample.
Unhealthy
New users stare at a loading spinner or a 'no caregivers found' dead end as their first post-signup experience.
What this means
Seeding the empty state removes the single highest-risk moment in the new flow, the gap between finishing the intent picker and seeing real value.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Zero-match zip codes show a dead-end empty state today | Ship the 3-sample-card fallback before removing any of the 12 original fields | dev ticket |
Final deliverable
A field-reduction map (12 fields -> 2 questions + 4 pre-match fields + 6 deferred fields) plus a 3-state fallback spec for the results screen.
See a reference example
Duolingo, first-lesson-before-signup rebuild (excerpt) BEFORE: email + password required before lesson 1 AFTER: lesson 1 playable immediately, account creation deferred until user tries to save progress Result: 47% reduction in first-week churn (UserGuiding, 2024 case study)
Success criteria
You're done when you can:
- Reduces pre-match fields from 12 to 2 questions + zip code only
- Specifies a non-blank fallback for both loading and zero-match states
- Correctly defers the 6 non-essential fields to a post-match step rather than deleting them