SCUM Product Framework

SCUM is not an industry standard or a replacement for discovery or delivery, but a compact author’s framework for talking about a product. Use it as a list of questions: where the context is unclear, what assumptions are untested, and where the team loses logic.

SCUM: Signal → Constraints → Understanding → Momentum

** Idea: ** Productive work is the transformation of noise (opinions, desires, chaos of the backlog) into **signals, then into **limitations, then into **understanding, and only after that into **performance inertia, measured by effect.

The SCUM is specifically designed to:

• cut “feecherism,”
• Protect teams from random requests,
• accelerate learning,
• Focus on outcomes, not employment.

1) S - Signal: What is the truth important

The goal of the stage is to separate the **real signal ** from the noise of stakeholders.

** Entrances:**

• behavior data (analytics, funnels, retention),
• Customer voice (interviews, tickets, Sales call notes)
• Market context (competitors, prices, regulations)
• business restrictions (unit economy, channels, margin).

**Output: “Map of signals” **

• 5-15 signals per cycle (e.g., month/quarter)
• Each signal is formulated as an observation, not as a solution. “in activation, failure in step X”, “churn grows in segment Y”, “in sales, the objection of Z is repeated”.

SCUM Rule: If you can’t name a signal, you can’t discuss features yet.

(2) C — Constraints (Restrictions): rate frame

The goal of the stage is to turn “we want everything” into a “managed bet.”

** Types of restrictions:**

• Resources (people/time/budget)
• Technical (architecture, security, SLA)
• ** legal/regulatory**,
• contextual (market timing, seasonality, dependencies)
• guardrails (which cannot be degraded: churn, latency, quality).

**Output: “Bet Card” * *

• signal → desired outcome → guardrails
• appetite (how much are you ready to eat: 2 weeks / 6 weeks / quarter)
• Non-goals (what we don’t consciously do)
• risks/assumptions

The SCUM rule: Not yet defined the limitations - any idea looks good.

3) U - Understanding: prove that the bet is not fake

The goal of the stage is to quickly understand what decision has a chance to have an effect, and how we will know.

• Tools (not as dogma, but as a set):

• Mini-research (5-8 interviews), prototypes,

• Quick tests of channels/offers,

• “Paper” calculations of the unit economy,

• technical spikes,

• Preliminary model of metrics (what should shift and why).

Output: “Evidence Pack” (evidence package)

• 1-2 possible approaches (not 10 options)
• The expected mechanism of the effect (what causality)
• Success/failure criteria,
• measurement plan
• Confidence level (Low/Med/High) + why.

SCUM: If you don’t know how to measure, you don’t know what to do.

4) M - Momentum (Inertia): deliver and consolidate the effect

The goal of the stage is not just to “release”, but to “achieve a shift of the metrics” and consolidate it.

** Principles of implementation:**

• Small releases, early rollout,
• feature flags, staged rollout,
• mandatory post-release review,
• “Kill switch” in case of harm.

**Output: “Momentum Report” **

• What we delivered,
• What we measured,
• What’s changed?
• what they learned
• Continue/stop/scale.

The SCUM rule: A release without a measured effect is considered unfinished work.

Scum rhythms

** Weekly* *

• Signal Review (30-45 minutes): New signals + confirmation of old ones.
• Experiment/Evidence Sync (30 mins): What we learned next?

** Once every 2 weeks**

• Bet Triage: what bets we take / cut / postpone.

** Monthly* *

• Momentum Review: betting effect, lessons, updates to the “signal map”.

** Quarterly* *

• Portfolio SCUM: big bets, investments, strategy, system constraints.

Artifacts SCUM: minimum set

1. Signal Map - List of signals and their weight/frequency/cost of the problem.
2. *Bet Card – Restricted Bet and Guardrails
3. *Evidence Pack - Evidence and Measurement Plan
4. *Momentum Report - Outcome and Decision (scale/iterate/kill)
5. **Decision Log, so you don’t argue in circles.

Roles and responsibilities

• Signal Owner (usually PM/Analyst): The quality of signals and their sources.
• Constraint Keeper (EM/Tech Lead/Legal/Finance on the situation): the reality of constraints.
• Evidence Driver (PM + Design + Analyst): Quick checks and evidence.
• Momentum Captain (EM/Team Lead): Delivery, rollout, quality, post-analysis.
• Portfolio Steward (Head of Product): Selection of bets at the company level.

SCUM metrics

• Outcome metrics on rates (main).
• Guardrails (which is not degraded).
• **Learning Velocity: Time from Bet to First Proof
• **Time-to-signal: How quickly we see changes in data after release.
• *Kill rate (healthy): The percentage of stopped bets is a sign of honest learning.

What anti-patterns help to notice SCUM

• “We’ll make a feature because the CEO said” – Signal first.
• “Let’s just get started with the Saws” – Constraints and Understanding first.
• “Well done” – Momentum requires an effect.
• “We don’t have time for discovery” → Evidence Pack is minimal and mandatory.