Prioritization Model
A scoring rubric I use when roadmap debates get too subjective. It compares business impact, technical complexity, and regulatory pressure so the tradeoffs are visible before a decision is made.
The full scoring rubric
Three dimensions, ten weighted signals. Every roadmap candidate gets one number; the matrix at the bottom turns those numbers into a go / no-go.
Roadmap conversations can get political quickly
In a lot of roadmap meetings, the loudest stakeholder wins, the engineering estimate becomes the whole conversation, and the bet that actually moves the P&L gets pushed to next quarter because it's “hard.”
This rubric exists to take that conversation off the rails. It forces every candidate on the roadmap through the same ten signals, produces a single comparable score, and puts the result on a 2×2 matrix that finance and CEO can read in three seconds. It doesn't make the decision for you — but it makes the decision defensible.
Dimension 1 — Business Impact (4 to 20)
Conversion Rate
1: <1 bp · 2: 1–2 bps · 3: 2–3 bps · 4: 3–4 bps · 5: 4 bps and above. Modeled against the surface's baseline traffic.Revenue Per Transaction
1: <$1 · 2: $1–$2 · 3: $2–$3 · 4: $3–$4 · 5: $4 and above. Net of refund and chargeback risk.Sign Ups
1: <100/day · 2: 100–200/day · 3: 200–300/day · 4: 300–400/day · 5: 400–500/day. Use rolling 30-day average.Call Volume
1: <100/day · 2: 100–200/day · 3: 200–300/day · 4: 300–400/day · 5: 400–500/day. Reductions count as positive impact (deflection cost).
Dimension 2 — Technical Complexity (4 to 20, inverted)
Technical Innovation
1: New tech, no experience · 2: New areas, limited experience · 3: Some new, some before · 4: Minimal, done infrequently · 5: None, done frequently. Higher score = lower risk.Architectural Complexity
1: Impacts 5+ systems · 2: 4 systems · 3: 3 systems · 4: 2 systems · 5: 1 system. Forces honest blast-radius accounting.Dependencies
1: More than 2 external parties · 2: 1–2 external parties · 3: Other projects and teams · 4: Some but manageable · 5: None. External dependencies always slip — penalize accordingly.Implementation Cost
1: >$40K/month · 2: $30K–$40K · 3: $20K–$30K · 4: $10K–$20K · 5: <$10K/month. Include design, eng, QA, and PM time.
Dimension 3 — Regulatory & Consistency (2 to 10)
Is this a regulatory change?
1: No · 5: Yes. Regulatory work jumps the queue regardless of business impact — non-negotiable.Is this required to maintain parity with desktop?
1: No · 5: Yes. Cross-platform consistency is a multiplier on lifetime value; debt here compounds fast.
Score, plot, defend
Every roadmap candidate gets a Business Impact score (4–20), a Technical Complexity score (4–20, where higher means easier), and a Regulatory score (2–10). The first two dimensions plot the candidate on a 2×2 — high-value / low-complexity goes first, the opposite quadrant gets challenged or killed. Regulatory acts as an override.
The single most useful artifact this produces isn't the ranking — it's the shared vocabulary. When engineering says “this is a 2 on architectural complexity” instead of “this is hard,” the conversation gets measurably better. Finance learns to push back on Conversion Rate scores. PMs stop inflating their own bets because the numbers are visible to everyone.
The rubric gives product, engineering, finance, and the CEO the same facts to react to. That is what makes the decision easier to defend.
What this model is not
This is not a substitute for product judgment. It does not capture strategic optionality (a low-score bet that opens a future market), brand effects, or talent-development value. Use it as the floor for the conversation, not the ceiling.
Recalibrate the score thresholds yearly. A “5” on Conversion Rate at $300M ARR is not the same as a “5” at $5B. Drift in your score distribution is a leading indicator that the bands are stale.
Want to talk through this work?
If this maps to a problem you’re hiring for, I’m happy to walk through the decisions behind it — including the ones I’d make differently now.