The Product Management Field Card

Discovery

Find the real problem

Own the problem, not the solution. If you cannot state it in one sentence, you have not found it yet.

Problem statement

Segment struggles to job to be done because barrier. Today they workaround, which costs time / money / risk. If we solve it, measurable outcome.

Interview rules

Ask about past behaviour, never future intent. "Tell me about the last time" beats "would you use".
Five whys until you reach something a person can act on.
Never pitch your solution inside the interview.
Stop at saturation, usually five to eight people per segment.
Write the notes the same day or they are gone.

Evidence ladder

How much weight a claim earns. Most roadmap arguments are two people reasoning from different rungs.

Jobs to be done

When situation, I want to motivation, so I can expected outcome.

People do not buy the drill. They do not buy the hole either. They buy the shelf being up before their partner gets home.

Value prop

The value proposition canvas

Fill the circle before you touch the square. Every team that does it the other way ends up describing features nobody asked for.

Tap any block to see what belongs in it. The two halves are drawn apart on purpose: fit is something you have to prove, not something the diagram gives you.

Customer profile: fill this first

Customer jobsWhat they are trying to get done. Functional, social and emotional. In their words, not yours.

PainsBad outcomes, risks and obstacles. Rank each as severe or moderate.

GainsOutcomes they want. Split into required, expected, desired and unexpected.

Value map: design this second

Products and servicesA plain list of what you offer. Rank by how essential each one is.

Pain relieversHow the offer removes a specific named pain. Severe beats mild.

Gain creatorsHow the offer produces a specific named gain. No match on the right means cut it.

How to actually run it

Pick one segment. One canvas per segment, never a blended average customer.
Fill jobs, pains and gains from interview evidence, not from the room.
Rank every item. Severe pains and desired gains rise to the top.
Only now list your products and services.
Draw a line from each reliever and creator to the ranked item it addresses.
Anything with no line is a candidate to cut. Anything ranked high with no line is your roadmap.

Three fits, in order

Problem-solution	Evidence says the pains and gains you picked are real and ranked correctly. This is a canvas on paper.
Product-market	Customers actually use what you shipped and keep coming back. This is a retention curve.
Business model	They pay enough, often enough, to cover the cost of serving them. This is a P&L.

The one-line value proposition

Our offer helps segment who want to job by relieving pain and creating gain, unlike the current alternative.

Prioritize

Decide what wins

Scores do not make decisions. They make disagreements specific enough to resolve.

RICE, live

Reach times impact times confidence, divided by effort.

Reach: people affected per quarter

Impact per person

Confidence in the estimate

Effort: person-months

2133RICE score

Top of the backlog. Staff it this quarter.

Your item2133

Onboarding cut3200

CSV export900

Theme engine220

Kano model

Not every feature pays back the same way. Investment on the x-axis, satisfaction on the y-axis.

Basics earn nothing when present and cost you everything when missing. Compete on performance. Ship one or two delighters a year, not one a sprint.

Other scoring models worth knowing

WSJF	Cost of delay divided by job size. Best when every stakeholder claims urgency.
Opportunity	Importance plus the gap between importance and satisfaction. Finds underserved needs.
MoSCoW	Must, should, could, won't. Fast for fixed-date scope cuts, weak for ranking.
Shortest job	When capacity is the binding constraint, small and valuable beats large and valuable.

Tiebreakers when scores are close

Cheap and reversible? Ship it and learn. Do not run a committee on a two-way door.
Expensive or hard to undo? Slow down and buy more evidence first.
Dead heat? Pick the option that teaches you more about the customer.
Still stuck? Ask which one you would regret not doing in a year.

Strategy

From vision to backlog

If your strategy contains no tradeoff, it is a wish list with a nicer font.

VisionWhere we are going.

StrategyHow we win, and what we give up.

North starOne metric that proxies customer value.

OKRsQuarterly outcomes, two to four key results.

RoadmapSequenced bets against those outcomes.

BacklogThe work itself.

Each layer answers the one above it. When someone challenges a backlog item, walk up the stack until you reach the level where you actually disagree.

What a real strategy contains

Diagnosis. What is actually going on, stated plainly.
Guiding policy. The approach you are taking in response.
Coherent actions. What you will do, and what you are deliberately not doing.

OKRs that survive contact

Objective: qualitative, ambitious, time-boxed Key result: measurable outcome, not output

Weak Launch the new checkout.
Strong Cut checkout abandonment from 34% to 22%.

Metrics

Know if it worked

Every target metric gets a guardrail. Conversion up and refunds up is not a win.

Retention is the honest chart

Percent of a signup cohort still active, day zero to day ninety.

Fix retention before you spend a dollar on acquisition. Filling a bucket with a hole in it is an expensive way to measure the hole.

The funnel, with drop-off

Acquisition, activation, retention, revenue, referral.

Work the widest leak first. Ten points of activation is usually worth more than doubling the top of the funnel.

Formulas you will be asked for

Activation rate: users reaching the aha moment / new signupsDefine the aha moment first, or this number means nothing.
Day-N retention: cohort active on day N / cohort sizeAlways cohort-based. Blended retention hides the truth.
Net revenue retention: (start ARR + expansion − churn − downgrade) / start ARRAbove 100% means you grow without new logos.
Lifetime value: ARPU x gross margin % / churn rateUse gross margin, not revenue, or you will overstate it.
CAC payback: CAC / (ARPU x gross margin %)Target under twelve months for most SaaS.
LTV to CAC: target 3:1 or betterBelow 1:1 you are buying revenue at a loss.

North star, by business model

Model	North star	Guardrail to watch
Marketplace	Successful transactions per week	Supply-side fill rate, dispute rate
SaaS	Weekly active accounts doing the core action	Seat utilisation, support ticket volume
Content	Time spent on content rated good	Session quality, complaint rate
Enterprise tool	Tasks completed per seat per week	Admin overhead, time to first value

Roadmap

Sequence the bets

Confidence horizons, not a dated calendar. A roadmap is a set of intentions, not a set of promises.

This quarter

Now

Scoped, staffed, committed. Outcomes named, metrics instrumented, engineers already in the code.

One to two quarters

Problem is defined and evidenced. The solution stays loose on purpose. Discovery is running now.

Beyond

Later

Themes and bets only. No dates, no estimates, no promises to sales. This column will change.

Write outcomes, not features

Weak Add Apple Pay to checkout.
Strong Cut mobile checkout abandonment. Apple Pay is the first hypothesis.

Feature-shaped roadmaps commit you to the answer before you have tested it. Outcome-shaped roadmaps let the team find a cheaper route to the same result.

Protect capacity every sprint

65%roadmap20%debt15%unplanned

Roadmap: the committed outcomes Debt: reliability, refactors, upgrades Unplanned: escalations, incidents

Tune the split to your defect load. Teams that book 100% roadmap deliver about 60% of it and lose the trust either way.

Documents

Write it down

If you cannot make it compelling on one page, do not put nine engineers on it.

One-pager, first

Problem and evidence
Target user and segment
Proposed approach
Success metrics
Non-goals
Risks and unknowns
Rough size, order of magnitude

PRD, when it is funded

Context and problem
Goals and non-goals
Users and scenarios
Requirements with acceptance criteria
Flows and edge cases
Dependencies and open questions
Instrumentation plan
Rollout and rollback

Story and criteria

As a role I want capability so that benefit. Given context when action then result.

INVEST: independent, negotiable, valuable, estimable, small, testable.

Working backwards

Write the launch announcement and the customer FAQ before the spec. It forces you to say what changes for the customer, in language they would recognise. If the announcement is boring to write, the feature will be boring to receive.

Delivery

Working with engineering

Bring the problem and the constraints. Let the team own the how.

The four levers

Scope, time, people, quality. Quality is not a real lever. Adding people to a late project makes it later. That leaves scope and time, and only one of those is usually yours.

Never negotiate an estimate. Negotiate the scope behind it.
Pay tech debt down continuously, not in heroic quarters.
Definition of ready before a story enters a sprint. Definition of done before it leaves.
If you are in the standup, you are there to unblock, not to collect status.

Every ceremony, one line

Ceremony	The actual point
Refinement	Shared understanding, then sizing
Planning	Commit to a goal, not a list
Standup	Surface blockers within a day
Review	Show working software, collect feedback
Retro	Change exactly one thing that matters

Experiments

Test before you commit

A flat result is a real result. Ship the simpler option and move on.

Hypothesis format

We believe change for segment will cause metric movement. We will know we are right when threshold within timeframe.

Rules that save you

One primary metric. Everything else is a guardrail or a secondary.
Run a full business cycle, usually one to two weeks. Never stop early on a peek.
Small effects need big samples. Detecting a 1% lift takes roughly a hundred times the traffic of a 10% lift.
Not everything needs a test. Bug fixes, tiny traffic and genuine strategic bets can just ship.

Rollout ladder

Every rung has a kill switch and a named owner watching the guardrails.

1%Canary

10%Early read

50%Powered test

100%Ship

Move up a rung when the guardrails are clean, not when the calendar says so. Internal dogfood sits below the first rung and costs you nothing.

Pricing

Packaging basics

Price on the metric that scales with customer benefit, not with your cost.

Pick the value metric

Seats, usage, transactions or outcomes. The right one grows when the customer succeeds, is easy to forecast, and does not punish them for adopting the product more deeply.

Three tiers beat one

Good, better, best. Most buyers pick the middle, which is why the middle is where you put the margin. A single plan leaves money on the table at both ends.

Discounts are permanent

Credits and trials expire. Discounts do not. Grandfather existing customers through price changes and announce early, or you will spend the quarter on retention calls.

Stakeholders

Manage the room

Escalate early, with a recommendation attached. Never bring a problem with no options.

When someone says "just add this"

What problem does this solve, for whom?
How many people have it, and how often?
What happens if we do nothing?
What would you like me to trade out for it?

Question four does most of the work. Saying no is hard. Making the tradeoff visible and letting them choose is not.

RACI, one accountable name

Responsible	Does the work
Accountable	Owns the outcome. Exactly one person.
Consulted	Input before the decision
Informed	Told after the decision

Communication cadence

Weekly written update. Shipped, in flight, blocked, metrics, decisions needed.
Monthly. Metrics against OKRs, honestly, including the ones going sideways.
Quarterly. Roadmap and strategy refresh with the tradeoffs shown.

Traps

Failure modes and fixes

Most of these feel like productivity while they are happening. That is what makes them expensive.

Trap	What it looks like	Fix
Feature factory	Ship count is up, no metric moved	Tie every roadmap item to an outcome
Loudest customer	One account drives the quarter	Segment the request and quantify demand
Promise calendar	Dated roadmap shared with sales	Confidence horizons: now, next, later
Solutioning early	The spec arrives before the problem	Separate discovery from delivery
Vanity metrics	Signups and page views on the wall	Move the wall to retention and revenue
Analysis paralysis	Weeks of debate on a reversible call	Two-way doors get decided in a day
Yes to keep peace	Backlog grows, nothing gets cut	Make the tradeoff explicit and public
No non-goals	Scope creeps in through the edges	Write what you are not doing, first

Glossary

Terms, defined properly

Half of these get used loosely in meetings. Being the person who uses them correctly is cheap credibility.

MVP The smallest thing that tests the riskiest assumption. Not a small version of the finished product.

Two-way door A reversible decision. Decide it fast and cheaply.

Value metric The unit customers happily pay more for as they get more value.

Aha moment The first action that predicts long-term retention.

Product-market fit Retention flattens and demand outpaces your ability to serve it.

Time to value Signup to first meaningful outcome. Shorten this before anything else.

Dogfooding Your team using the product for real work, not a scripted demo.

Sunk cost Already spent. Irrelevant to the next decision.

Tech debt Shortcuts, deliberate or not, that raise the cost of every future change.

Guardrail metric The thing you are not allowed to break while chasing the target.

A week that works

Mon

Metrics review, sprint check, priorities set for the week.

Tue

Customer contact. Calls, tickets, session recordings.

Wed

Deep work. Specs, analysis, strategy. Guard this block.

Thu

Stakeholder syncs and cross-team unblocking.

Fri

Demo, written update, backlog refinement.

Two habits separate strong product managers from busy ones: talk to a customer every week, and write your decisions down.