The Discovery Hub is a Notion workspace I built to hold real product work — a PRD, notes from 20+ discovery interviews, and a RICE table scoring five decisions I actually made on the Distributor Management System. A live Zapier automation keeps its feedback database filling itself.
Not a template exercise. Every entry is grounded in real product decisions from my work on the DMS at Salescode.ai — the same interviews, the same trade-offs, scored and structured after the fact. The Zapier automation is a working example, not a claim: submit the form and a Notion row appears.
Discovery generates a particular kind of mess: interview notes scattered across documents, decisions made in meetings and never written down, priorities argued rather than scored. The work happens, and then it evaporates. Six months later nobody can reconstruct why a call went the way it did.
I wanted a working answer to that in a tool remote teams actually document in. So I built the Discovery Hub in Notion and filled it with my own real material rather than sample data — the interviews I ran and the decisions I made building the Distributor Management System.
That constraint is the point. Anyone can populate a Notion template. Structuring your own messy history into something referenceable is the part that's actually hard, and it's the part that proves the skill.
The Hub is one workspace with four linked parts, arranged in the order discovery actually moves — interviews, then prioritisation, then a spec, then a loop back for new input:
Written-up entries from 20+ real discovery interviews with distributors, sales heads and finance teams. Structured so a note can be filtered and cited later, not just read once.
Five real DMS decisions, scored retrospectively on reach, impact, confidence and effort. Scoring calls I'd already made was the honest test — the framework had to agree with what shipped, or explain why it shouldn't have.
A full spec for the finance decision I'd defend hardest on the DMS: invoices are immutable, with a corrections log, never free-form edits. Slightly less convenient, dramatically more auditable.
A live intake table that fills itself from a Google Form. This is the piece that turns the Hub from a document into a system that keeps running.
RICE is easy to perform and hard to use honestly. Running it forward on hypothetical features tells you very little. Running it backward on decisions whose outcomes you already know is a real test of whether your inputs were any good.
So the table scores five DMS calls I'd already lived with, including the density-versus-polish trade-off that a pilot client pushed back on:
| Decision | Why it scored where it did | Rank |
|---|---|---|
| Classic / density mode | High reach across every power user, low effort once the data layer existed | 01 |
| Invoice immutability | Narrow reach, but impact and confidence both maximal — finance can't sign off without it | 02 |
| Role-based access | Structural. Cheap on day one, expensive as retrofitted security debt | 03 |
| Orders-processed as the adoption metric | No user-facing reach at all, but it governs every later prioritisation call | 04 |
| Phone-first field, desktop-first admin | Highest effort of the five; justified by two genuinely different jobs | 05 |
The interesting result was the density decision landing first. That was the call we got wrong initially — we shipped clean and got pushed back on. Scored properly against reach and effort, it should have been obvious from the start. That's the value of doing this retrospectively: it shows you which instinct to trust next time.
The Feedback database is connected to a Google Form through a live Zapier automation. A submission arrives, Zapier fires, and a structured Notion entry appears. There is no copy-paste step and no batching — the loop runs whether I'm looking at it or not.
submit it and the row is just… there ✿
The tooling here is deliberately unglamorous. What it demonstrates is trigger-and-action thinking: noticing that two tools don't talk to each other natively, and treating the manual step between them as a defect rather than as work.
Remote product roles increasingly assume that instinct. This is a working example of it rather than a line on a skills list.
There's one idea underneath most of my product work, and it came out of watching distributors use the DMS:
Distributor operators process high volumes of orders, returns and accounting entries every day, so a minimal UI that hides information behind extra clicks and scrolling actually slows them down — density beats polish when speed is the job.
The Discovery Hub is that instinct at a much smaller scale. A prettier system that still needed me to copy form responses into Notion by hand would have been the wrong build. Structure that removes manual friction beats structure that looks tidy — the same trade-off, one tool down.