Test Track

Every track ever

Every published demo, in reverse-chronological order. Each one is an artifact of one team's week. Scroll to see how problems evolve.

  1. People returning from prison lose their job offers in week one

    Demo · Track 4 · published 1/1/1970

    Track 4: a re-entry buddy match that actually shows up

    What we built

    A lightweight SMS-based pairing flow that matches a returning citizen with a trained peer mentor within 48 hours of release, with a check-in scheduled before the pair meets.

    What we learned

    • SMS beats app installs for people without stable data plans
    • A scheduled first call beats 'reach out when you can' by 4x
    • Mentors stay engaged when they get a one-tap 'I checked in' button

    Decision experiments

    • Create· logged by Amara✓ Validated

      SMS-first with scheduled first call

      vs.Mobile app with in-app chat

      Hypothesis

      Higher first-week contact rate

      Outcome0

      78% of pairs connected in week one (up from 19%)

    • Brainstorm· logged by Naledi✓ Validated

      Match by ZIP code, not by job sector

      vs.Match by industry experience

      Hypothesis

      Mentors who can actually meet in person stay engaged longer

      Outcome0

      County-level matching held up across 3 cities.

    Next track: Carry forward. Pilot with two more re-entry orgs next week; add a Spanish-language flow.

    Team: Amara, Tomás, Naledi, Yusuf, Esperanza

  2. Transit deserts cut off entry-level workers from new jobs

    Demo · Track 1 · published 12/31/1969

    A neighbor-driver carpool with employer-paid gas cards

    Transcript

    Anika: Hey everyone — Team Alpha here. This week we tested whether a neighbor-driver carpool, paid for with a flat $25 weekly gas card from the employer, could keep new hires showing up past day 21. We built a one-page matchboard pinned at the front desk and a simple agreement riders and drivers both sign. Reza: Here's the headline — 11 of 14 new hires were still on the job at day 90. Up from 4 of 14. Riders trust a neighbor more than a stranger from an app. The flat gas card was way easier than tracking miles. And the employer paid for it the moment we showed them what week-3 turnover actually costs. We're carrying it forward to track 2 with a backup-driver list. Thanks.

    What we built

    A simple matchboard pairing new hires with neighbors who already drive that route, plus a $25 weekly gas card the employer reimburses. Drivers and riders sign a one-page agreement.

    What we learned

    • Riders trust a neighbor more than a stranger from an app
    • A flat weekly gas card is easier than per-mile reimbursement
    • Employers will fund retention if you show them the cost of week-3 turnover

    Decision experiments

    • Create· logged by Devon✓ Validated

      Neighbor carpool + flat gas card

      vs.Subsidized rideshare credits

      Hypothesis

      Higher 90-day retention

      Outcome· recorded 12/31/1969

      11 of 14 new hires still on the job at day 90 (up from 4 of 14)

    • Brainstorm· logged by Anika✗ Refuted

      Subsidized rideshare credits as a backup

      vs.No backup at all

      Hypothesis

      Catches the nights when the carpool falls through

      Outcome· recorded 12/31/1969

      Drivers cancelled most night rides — the credits sat unused. Replacing with a backup-driver list next track.

    Next track: Carry forward. Roll it out to two more employers next track; add a backup-driver list.

    Team: Reza, Anika, Sam, Jules, Devon

  3. People returning from prison lose their job offers in week one

    Demo · Track 3 · published 12/25/1969

    Track 3: a re-entry buddy match that actually shows up

    What we built

    A lightweight SMS-based pairing flow that matches a returning citizen with a trained peer mentor within 48 hours of release, with a check-in scheduled before the pair meets.

    What we learned

    • SMS beats app installs for people without stable data plans
    • A scheduled first call beats 'reach out when you can' by 4x
    • Mentors stay engaged when they get a one-tap 'I checked in' button

    Decision experiments

    • Create· logged by Amara◐ Mixed

      SMS-first with scheduled first call

      vs.Mobile app with in-app chat

      Hypothesis

      Higher first-week contact rate

      Outcome· recorded 12/25/1969

      78% of pairs connected in week one (up from 19%)

    • Brainstorm· logged by Naledi✓ Validated

      Match by ZIP code, not by job sector

      vs.Match by industry experience

      Hypothesis

      Mentors who can actually meet in person stay engaged longer

      Outcome· recorded 12/25/1969

      County-level matching held up across 3 cities.

    Next track: Carry forward. Pilot with two more re-entry orgs next week; add a Spanish-language flow.

    Team: Amara, Tomás, Naledi, Yusuf, Esperanza

  4. People returning from prison lose their job offers in week one

    Demo · Track 2 · published 12/18/1969

    Track 2: a re-entry buddy match that actually shows up

    What we built

    A lightweight SMS-based pairing flow that matches a returning citizen with a trained peer mentor within 48 hours of release, with a check-in scheduled before the pair meets.

    What we learned

    • SMS beats app installs for people without stable data plans
    • A scheduled first call beats 'reach out when you can' by 4x
    • Mentors stay engaged when they get a one-tap 'I checked in' button

    Decision experiments

    • Create· logged by Amara✓ Validated

      SMS-first with scheduled first call

      vs.Mobile app with in-app chat

      Hypothesis

      Higher first-week contact rate

      Outcome· recorded 12/18/1969

      78% of pairs connected in week one (up from 19%)

    • Brainstorm· logged by Naledi✓ Validated

      Match by ZIP code, not by job sector

      vs.Match by industry experience

      Hypothesis

      Mentors who can actually meet in person stay engaged longer

      Outcome· recorded 12/18/1969

      County-level matching held up across 3 cities.

    Next track: Carry forward. Pilot with two more re-entry orgs next week; add a Spanish-language flow.

    Team: Amara, Tomás, Naledi, Yusuf, Esperanza

  5. People returning from prison lose their job offers in week one

    Demo · Track 1 · published 12/11/1969

    Track 1: a re-entry buddy match that actually shows up

    What we built

    A lightweight SMS-based pairing flow that matches a returning citizen with a trained peer mentor within 48 hours of release, with a check-in scheduled before the pair meets.

    What we learned

    • SMS beats app installs for people without stable data plans
    • A scheduled first call beats 'reach out when you can' by 4x
    • Mentors stay engaged when they get a one-tap 'I checked in' button

    Decision experiments

    • Create· logged by Amara◐ Mixed

      SMS-first with scheduled first call

      vs.Mobile app with in-app chat

      Hypothesis

      Higher first-week contact rate

      Outcome· recorded 12/11/1969

      78% of pairs connected in week one (up from 19%)

    • Brainstorm· logged by Naledi✗ Refuted

      Match by ZIP code, not by job sector

      vs.Match by industry experience

      Hypothesis

      Mentors who can actually meet in person stay engaged longer

      Outcome· recorded 12/11/1969

      ZIP matches were rarer than expected; we relaxed to county-level.

    Next track: Carry forward. Pilot with two more re-entry orgs next week; add a Spanish-language flow.

    Team: Amara, Tomás, Naledi, Yusuf, Esperanza