Playbook · 17 guides

The product fundamentals, written down.

Not theory. The specific artefacts you will be asked to produce in your first ninety days, with a worked example and a template for each.

Foundations · 3

What a PM actually does

Discovery, delivery, and the third job nobody lists: being the reason a decision gets made this week rather than next month.

Vision → strategy → roadmap

Where each one stops. The most common failure is a roadmap doing a strategy's job.

Working with engineering without becoming a ticket queue

Context over instruction. What to bring to refinement, and what to stop bringing.

Measurement · 4

Writing OKRs that survive Q2

Objective, key result, and the difference between a KR and a task. Ten real examples and why six are wrong.

KPIs, north star and guardrails

How to pick one number without letting the team game it. Counter-metrics as a design constraint.

Instrumenting before you launch

An event-tracking plan written at spec time, not after the exec asks how it's performing.

The metrics review that isn't a status update

An agenda that produces decisions: what moved, what we believe caused it, what we're changing.

Writing · 4

User stories with real acceptance criteria

Given/When/Then that engineers don't rewrite. Before-and-after on a live example.

The one-page PRD

Problem, evidence, non-goals, success measure, open questions. Anything longer is a document nobody finishes.

Storytelling for product people

Structure a narrative an executive can retell without you in the room: situation, tension, choice, ask.

Writing the update nobody has to chase you for

Weekly, short, same shape every time. Status, risk, decision needed.

Discovery · 3

Customer interviews without leading

Question scripts, the five-why trap, and how to spot a polite lie.

Prioritisation that holds up

RICE, WSJF and opportunity scoring — when each is honest and when it's theatre.

Turning research into a decision

Synthesis that ends in a choice with a date, not a deck with themes.

AI PM · 3

Specifying an AI feature

Acceptance criteria for a probabilistic system: eval sets, failure budgets, fallback behaviour, human review.

Responsible AI as a product requirement

Governance, privacy-by-design and model cards written as tickets, not policy PDFs.

Pricing and unit economics for AI features

Cost per request, margin per plan, and the caps that stop a power user costing you money.

Pathway · 6 steps

Breaking into product management.

The honest version. Most advice assumes you can afford a bootcamp and a career gap; this assumes you can't.

1 · Name the door you're already standing at
Week 1
Engineers, analysts, support leads, consultants and QA all have a shorter path than they think. Your domain is the asset — pick roles where it counts double.
2 · Do the PM job before you have the title
Months 1–3
Write the spec nobody wrote. Run the interviews nobody ran. This is what turns into your portfolio.
3 · Build three artefacts, not a certificate
Months 2–4
One teardown, one PRD with metrics, one launch retro. Hiring managers read artefacts; ATS reads keywords.
4 · Fix the CV for the ATS and the human
Month 4
Outcomes with numbers in the first three bullets. The rest is context.
5 · Practise the four interview types
Months 4–6
Product sense, execution/metrics, technical, behavioural. Each has a structure, and each is failed for a different reason.
6 · Negotiate the level, not just the salary
Offer
Coming in one level low costs more over three years than any signing bonus makes back.