Track C · Who tests it

Pilot plan

MiFamilias has no value for a solo user — the product is a family network. So the pilot doesn't recruit individuals, it recruits families, one "connector" at a time, and starts with pure UI testing in a browser before anyone installs anything.

The metric that matters

Not signups — complete-family activation. One engaged family of 8 is worth more than 100 individual installs. If a tester can't get 3+ family members in during week one, treat it as a failed activation and learn why.

01Who we pilot to

Stage 0 · internal

Owner + 1–2 friendlies

Dogfood the web UI. Find the obvious breakage and rough copy before anyone outside sees it.

1–3 people

browser

Stage 1 · seed families

2–3 "Family Connectors"

The persona who organizes the family WhatsApp group and the shared photo album. Each brings their household in. This is the core network-effect test.

2–3 families

web → device

Stage 2 · closed beta

Extended cohort

Friends-of-seed families + recruited via a milestone moment (wedding, new baby, milestone birthday). TestFlight external / Play closed.

5–10 families

device

02The "Family Connector" recruit

Target one person per household who will pull the others in. Good signals: they already run a family group chat, organize gatherings, or keep the photos. Give them a reason to invite everyone — a shared capsule for an upcoming milestone is the natural hook.

Identify the connector — who in this family already does the coordinating?
Give a milestone hook — "record a capsule for [wedding / birthday / new baby]" that unlocks the family invite loop.
Target a complete unit — aim for 5–15 members, cross-generational.
Success = 3+ members active in week one, at least one capsule created, at least one Sage/feed interaction.

03UI-first testing (before app registration finishes)

Per the roadmap, the very first test is the web build behind Cloudflare Access — no install, no store account. Testers open a link and walk scripted flows. This is where copy, layout, and flow-confusion feedback comes from, and it starts on day one.

#	Flow to test	What we're watching
1	Sign up & onboarding	Where do people stall? Is age/consent clear?
2	Create a family & invite	Is the invite obvious? Would they actually send it?
3	Join a family (as invitee)	Does the invite link → join flow just work?
4	Post to the family feed	Do they understand it's private to the family?
5	Create a time capsule	Do they get "unlocks in the future"? Pricing clear?
6	Try Sage / Transcend (needs GPU box up)	First impression of the AI legacy idea
7	Communities & notifications	Discoverability, noise level

On-device only checks

Push notifications, camera capture, biometric/PIN unlock, and offline behavior can't be fully validated in a browser — those wait for the TestFlight / Play internal build (Phase 4).

04Feedback loop

Capture

Short per-flow form (clarity 1–5, friction, confusion, bugs)
Session notes / screen recordings where testers allow
A single triage list ranked by frequency × severity

Act

Fix the top-3 friction points each round
Re-deploy web build (instant) / re-cut device build
Log changes in the portal Updates log

05Pilot success criteria

≥ 3

members active per seed family, week 1

≥ 1

capsule created per family

≥ 50%

day-7 return rate

2–3

complete families activated

Hit these and the product has demonstrated its core loop — a family joining and staying. That's the green light for store submission and public launch (Roadmap P6).