Skip to main content

Playbook Extreme

This document defines an extreme lightweight version of "Product Speak" — just enough shared language and structure to align teams, avoid feature-first thinking, and support fast, confident decisions.

The goal is not completeness. The goal is clarity with minimal ceremony.

The Minimal Stack (3 artifacts, 1 habit)​

1. Lean Persona (½ page)​

Purpose: Make the primary user concrete without turning personas into marketing fiction.

Minimum fields:

  • Role / context (not demographics)
  • Primary goal
  • One key constraint
  • How success is judged

Explicitly exclude:

  • Names, photos, quotes
  • Backstories
  • Demographics unless they directly affect behavior

Example:

- Role: Service owner responsible for a customer-facing production service
- Goal: Keep the service reliable and recover quickly when things go wrong
- Constraint: Limited time during incidents; not an observability specialist
- Success looks like: Issues are detected and understood quickly, without jumping between tools

If the persona isn’t referenced in real discussions or decisions, it isn’t doing work.

2. JTBD Lite (Problem Statement)​

Purpose: Anchor conversations on outcomes without requiring full JTBD rigor.

Required format (one sentence):

When situation, I want to make progress, so I can outcome.

Optional (only if useful):

  • Why now
  • Competing alternative

Example:

When my service begins to degrade or fail,
I want to understand what’s happening and why using correlated signals,
so I can restore service health quickly and reduce customer impact.

If it takes more than two minutes to write, it’s too heavy.

3. Problem Backlog (JTBD as a list)​

Purpose: Replace feature backlogs with prioritized problems.

Each item includes:

  • JTBD sentence
  • Pain score: High / Medium / Low
  • Confidence level: High / Medium / Low

No solutions. No acceptance criteria.

Problem Backlog — Owner Olivia​

[H / H] Understand incidents quickly using correlated logs, metrics, and traces
[H / M] Detect service degradation before customers notice
[M / M] Reduce time spent maintaining dashboards and alerts

Simple Discovery (Signal, Not Research)​

Discovery is not a phase and not a ceremony in this model.

The only rule​

Every meaningful problem must be backed by at least one real signal.

Valid signals include:

  • User conversation
  • Support ticket cluster
  • Incident review
  • Observed usage pattern
  • Internal dogfooding pain

Discovery Signal — Owner Olivia​

Signal: During an incident review, Olivia reported spending the first 15 minutes switching between dashboards, logs, and traces to understand what failed (incident review, Jan 12).

No transcripts. No decks. No synthesis documents.

How the Pieces Fit Together​

Persona → frames relevance
JTBD → defines the problem
Discovery signal → proves it exists
Problem backlog → forces prioritization

This loop is the entire Product Speak core.

What This Model Intentionally Excludes​

To stay lightweight, this model does not require:

  • Full persona sets
  • Emotional / social JTBD breakdowns
  • Formal discovery plans

These tools may be used selectively, but are never defaults.

Guardrail Heuristic​

If an artifact:

  • is not referenced in weekly discussions, or
  • does not change a decision,

delete it.

Product Speak only works when it is spoken, not documented.

One-Sentence Definition​

We know who we’re building for, the job they're trying to get done, and we’ve seen evidence that it matters.