# Problem & Research

This document applies IBM Enterprise Design Thinking and UX research frameworks to define why Elevra exists and who it's for. It reflects the state of research at the end of a ~24-hour hackathon: hypotheses framed for future validation, not claims of completed field research.

## 1. From business problem to human problem (the 5 Whys)

Enterprise Design Thinking argues a problem well-stated is half-solved, and gets there by asking "why" until a business problem becomes a human one.

1. **The problem, as framed by the challenge brief:** Girls leave the STEM pipeline at a disproportionate rate, and the gap is worse for intersectional groups (First Nations, regional/remote, CALD, disabled, LGBTQIA+ students).
2. **Why?** They don't see a version of themselves persisting in a STEM career.
3. **Why?** The role models outreach programs show them — posters, panels, "women in STEM" campaigns — don't reflect their specific background.
4. **Why?** Programs default to the same small pool of well-known women, because manually matching a role model to one student's specific background doesn't scale.
5. **Why?** No tool matches role models to students by *lived-experience overlap* rather than gender alone — and nothing tells organisers which combinations of identity are going unrepresented.
6. **So what?** Build the matching tool, and make every unmatched search visible as a signal instead of a dead end.

## 2. Problem statement

> Girls with intersecting, underrepresented identities struggle to find STEM role models who reflect their specific background, because outreach relies on a small, generic pool of stories with no mechanism to notice — or close — the representation gap.

## 3. Hills (Who / What / Wow)

Enterprise Design Thinking "Hills" state intent as a measurable user outcome, not a feature list.

- **Who:** A high-school-aged girl from an underrepresented background considering a STEM path.
- **What:** Can find and connect with a woman in STEM who shares her specific lived experience — culture, location, disability, class, identity — in under 60 seconds, without flattening herself into a generic "women in STEM" profile.
- **Wow:** Unlike a static directory, Elevra's matching improves the more it's used, and treats every zero-match search as a recruitment brief for organisers — turning an invisible gap into an actionable one.

## 4. Personas

Three personas anchor design decisions, built from the challenge brief's target demographics.

### Anika, 15 — "I don't want to disappoint anyone"
- **Background:** South Asian, first-generation Australian, low-SES, outer-suburban Melbourne.
- **Goals:** Wants to study medicine; needs to bring her parents along, who are pushing commerce.
- **Pain points:** Every STEM role model she's shown is white, metropolitan, or already has parents who "get it."
- **Behaviour:** Researches quietly before committing to anything in front of family.
- **Quote (hypothesised):** *"I need proof it's actually possible for someone like me before I bring it up at home."*

### Jessie, 17 — "I've never met one"
- **Background:** First Nations, regional NSW, interested in environmental science.
- **Goals:** Wants to know a STEM career is reachable from where she lives.
- **Pain points:** Has never met a First Nations scientist; assumes the path requires leaving everything familiar behind.
- **Behaviour:** Trusts stories over statistics; disengages from anything that feels like a corporate campaign.
- **Quote (hypothesised):** *"Nobody who looks like me is on any poster at my school."*

### Ms. Patel — STEM teacher, multicultural high school
- **Background:** Runs STEM electives; wants better representation than "the same ten famous women."
- **Goals:** Wants a tool she can point a whole class at, and evidence of where representation is thin.
- **Pain points:** No visibility into which of her students see themselves in STEM and which don't.
- **Behaviour:** Adopts tools that save her prep time and give her something concrete to report on.

## 5. Success metrics — applying the HEART framework

| HEART dimension | Goal | Signal |
|---|---|---|
| **Happiness** | A user feels "someone like me exists in STEM" | Post-match sentiment / qualitative feedback |
| **Engagement** | Users explore beyond the first match | Cards swiped or stories opened per session |
| **Adoption** | Workshop attendees convert into profiles | % of QR scans that complete onboarding |
| **Retention** | Users come back between workshops | Return-visit rate at 2 weeks |
| **Task success** | Users reach a meaningful match and act on it | % reaching 3+ identity-overlap match; click-through to a scholarship/program |

The **"identity combinations with zero matches"** count is tracked as a separate accountability metric — it's the number Elevra wants to see fall over time, and it's the one shown to program coordinators, not just users.

## 6. Why not just another directory (positioning)

Most existing role-model resources treat diversity as a filter applied after the fact to a fixed list of well-known women. Elevra inverts that: identity overlap *is* the primary sort key, the matching logic is a transparent weighted-tag score (not a black-box model, so it's auditable and explainable to a 15-year-old), and a missing match is logged and routed to organisers as a story-recruitment target rather than silently returning nothing.

## 7. Risks & mitigations

| Risk | Mitigation |
|---|---|
| Identity data feels invasive to collect | Every identity question is optional with a "prefer not to say" path; data is never sold or shared |
| Representation becomes tokenistic | Rare identity combinations are weighted *higher*, not treated as edge cases; women submit their own stories rather than being curated top-down |
| Matching algorithm feels like a black box | Deliberately simple weighted-tag scoring instead of ML — the "why you matched" reasoning is shown in plain language |
| Cold start — new/rare profiles get no match | The "Missing Story" screen reframes a zero-match as a data point, not a dead end, and feeds the coordinator dashboard |