Case Study 01

Makaliʻi Metrics

Soil Decision-Intelligence for Hawaiʻi's Regenerative Agriculture Community

The first decision wasn't what to build. It was what the platform would never do, and why that boundary was the product.

AgTechActive Development

Product Management · UX Research · Roadmap · Applied Product Lead · ʻāina-first Design

Role
Applied Product Manager · UX Hybrid
Engagement Type
Embedded · Ongoing
Backlog Progress
9 of 24 items shipped · Active Sprint
Roadmap Structure
3-phase · 90-day framework with gate criteria

Makaliʻi Metrics is an emerging soil testing and analytics lab built in Hawaiʻi, for Hawaiʻi. It pairs modern soil analysis with Hawaiʻi-specific interpretation to help farmers, ranchers, researchers, and ʻāina stewards make faster, more informed decisions about their land. My engagement is a hybrid Product Manager and UX role, embedded directly in the product operation, not advising from the outside.

This is what Applied Product Lead work looks like in practice: I'm in the build environment alongside the team, owning the product discovery layer, shaping the roadmap, and making scope decisions that the engineering team can execute against. The product is actively shipping against those decisions.

These numbers represent the current state of an active engagement. They are not outcomes. They are the operational reality of a product in progress, being built against decisions that are documented below.

24
Active Backlog Items
Tracked across 5 workflow stages in ClickUp
9
Items Shipped
2 Sprint Complete · 7 Dev Complete
10
Heuristic Findings
Across Nielsen's 10 usability principles
3
Roadmap Phases
Day 1–30 · 31–60 · 61–90 with gate criteria
Generic extraction versus ʻāina-first interpretation The same Hawaiian soil sample yields a generic, context-blind recommendation from a mainland lab, but a relevant land decision when run through an ʻāina-first interpretation layer. Soil sample Hawaiian volcanic EXTRACTION LOGIC · BUILT ELSEWHERE Mainland soil lab Generic model · weeks turnaround Generic recommendation Context-blind · one size fits all misses relevance ʻĀINA-FIRST INTERPRETATION Interpretation layer Volcanic soil chemistry Microclimate & regenerative practice ʻŌlelo Hawaiʻi & stewardship framing Relevant land decision Soil intelligence that knows where it comes from. relevance
The product problem wasn't speed. It was relevance. The same soil data becomes generic extraction or grounded intelligence depending on the interpretation layer. ʻĀina-first is the strategy, not a branding skin.

Hawaiʻi's farming and land stewardship communities are rich in traditional ecological knowledge, but they have been underserved by data infrastructure built elsewhere, for elsewhere. Mainland soil labs return generic recommendations that miss the specificity of Hawaiian volcanic soils, microclimates, and regenerative practice. Practitioners wait weeks for data that doesn't fully serve them, then make land decisions with incomplete intelligence.

The product problem wasn't just speed or accuracy. It was relevance. A soil analytics platform that doesn't understand the land it's analyzing isn't a better tool. It is the same extraction logic with a faster turnaround.

This is an embedded engagement. I'm not a consultant reviewing outputs. I'm in the product operation. My responsibilities span the full discovery-to-delivery arc:

  • Led product discovery and definition, establishing the product spine and the boundaries that prevent scope creep from fragmenting the platform's core value
  • Conducted a full heuristic evaluation of the existing product experience, surfacing friction points across all 10 of Nielsen's usability principles
  • Designed and facilitated the Kōkua User Testing Session (02/19/26) with real practitioners, not proxy users, to ensure findings reflect how stewards actually interact with soil data
  • Translated user testing findings into a severity-weighted backlog, structured so the engineering team works the highest-impact fixes first
  • Built and actively maintains a 3-phase, 90-day roadmap with explicit gate criteria at each phase transition
  • Maintained an ʻāina-first design perspective throughout, evaluating every product decision against cultural alignment: language choices, data presentation, and interpretation logic
From findings to a severity-weighted backlog to a gated roadmap Heuristic findings and user-testing themes feed a backlog ordered by user impact, which is delivered across a three-phase ninety-day roadmap with explicit gate criteria. RESEARCH FINDINGS Heuristic evaluation 2 High 5 Medium 1 strength kept Kōkua user testing · real practitioners Language accessibility ʻĀina-first framing Community sharing model SEVERITY-WEIGHTED BACKLOG Ordered by user impact not development convenience 9 / 24 shipped · 37.5% HIGH IMPACT FIRST → Urgent Normal ■ shipped 90-DAY ROADMAP · GATE CRITERIA 1–30 Context Absorption Product spine · discovery complete GATE 31–60 First Operational Win Ship highest-severity fixes · validate flow GATE 61–90 Product Signal Adoption: logins · repeat uploads
Scope ordered by impact, gated by phase. User-testing findings become a severity-weighted backlog, delivered across a 90-day roadmap with explicit gates, so the team has defined checkpoints before moving from discovery into execution.

The decisions that shaped this product aren't in the feature list. They're in the scope boundaries and the framing choices that came before the build:

  • ʻāina-first as a product strategy decision, not a branding layer. Language, terminology, and data presentation were all evaluated against whether they serve Hawaiian land stewardship practice or impose an outside framework onto it
  • Severity-weighted backlog structure, ensuring the engineering team's sprint cycles are ordered by user impact, not development convenience
  • Phased roadmap with gate criteria, so the team has explicit conditions for moving from discovery into execution, not an open-ended build with no defined checkpoints
  • User testing with actual practitioners. The Kōkua session was not a usability test with recruited participants. It was a structured session with a working farmer and an intern from the community the product serves

Heuristic Evaluation

2
High Severity findings
Visibility of System Status · Recognition Over Recall
5
Medium Severity findings
Prioritized for next sprint cycle
1
Identified Strength, preserved by design
Aesthetic and Minimalist Design: not changed

Kōkua User Testing Session, 02/19/26

2
Participants, real practitioners
1 primary tester (Mālama Maui Nui intern) · 1 observing farmer
3
Key insight themes surfaced
Language accessibility · ʻĀina-first framing · Community sharing model
1
Structured, task-oriented session
Moderated protocol with observation notes and severity tagging

Backlog and Sprint Status

9/24
Items shipped, 37.5% completion
2 Sprint Complete · 7 Dev Complete
4
Items On Hold
Pending payment flow, user registration, and CRM-database integration
2
Severity levels in active use
Urgent and Normal, triaged against sprint goals

90-Day Roadmap Gate Criteria

1–30
Context Absorption
Product spine · stakeholder alignment · discovery complete
31–60
First Operational Win
Ship highest-severity fixes · validate core dashboard flow
61–90
Product Signal Feedback
Adoption signals: weekly logins · repeat uploads · crop practice changes

The product is in active development and shipping against a structured backlog. User testing surfaced concrete direction for the dashboard experience. The ʻāina-first framing has shaped the product vocabulary and the interpretation layer. The platform is being built to serve Hawaiian land stewards on their own terms, not adapted from a generic agricultural data model.

The roadmap provides the team clear gates before moving from discovery into execution, and the backlog ensures engineering cycles are ordered by user impact. Nine items shipped. Twenty-four tracked. Work ongoing.

The goal isn't faster soil data. It's soil intelligence that knows where it comes from.