Opportunity
An opportunity card is a customer problem written down, then interrogated until it earns its way onto the product backlog. Every piece of work in this playbook starts on one. The loop that does the interrogating is the Opportunity Backlog section. Here's the blank card, the order its fields fill in, and one card filled out end to end.
Copy the blank into whatever your team plans work in, and start every new problem on it.
The card, blank
Problem
When [situation],
I want [what they're after],
So that [the payoff],
But [the friction in the way today].
Users
[Who this is for. The specific people who feel it, not a market category.]
Expected Outcome
[We expect that... the change you're hunting, in the customer's world.]
Market Size
[How many customers this touches, and how hard. Your own numbers beat any industry report.]
Competitors
[What customers do about this today. Whatever they do now is the competition, including doing nothing.]
Our differentiator
[Why you're positioned to win this one. If there's no answer, say so on the card.]
How might we?
[How might we [solve this] without [the damage it could cause]? One line, framed to invite options, not pick one.]
Hypothesis
We believe that [the change you'll make]
For [who],
We can [the behavior you expect to see].
Measures of success
[The numbers that will move if the hypothesis is true. Named before the test runs, so nobody moves the goalposts after.]
Feasibility
[What you'd need to build or buy, sized honestly. Weeks or quarters, not "straightforward."]
MVE
[The smallest thing you can put in front of a real customer to learn. We can test this theory by...]
Number of users tested
[How many real customers the test actually reached.]
What did we expect?
[The prediction, written before the data lands. This field exists so success can't be redefined later.]
Experiment Success?
[Filled last. What happened, against the prediction, and which door. Build, pivot, or archive.]
The card fills as the loop runs
Don't sit down and fill all fourteen fields in one sitting. You can't. Half of them are answers you don't have yet.
The card fills step by step, and each step of the loop fills its own fields. Define the problem, and the card gains Problem and Users. Run the assessment, and it gains Expected Outcome, Market Size, Competitors, Our differentiator, and How might we. The hypothesis step adds Hypothesis and Measures of success. Concepts add Feasibility and MVE. The test adds Number of users tested and What did we expect, and that expectation gets written down before any data arrives, which is the whole reason it's worth anything. The last field, Experiment Success, waits until the room reads the results.
A half-filled card isn't a half-done card. It's a card mid-interrogation, and its blank fields are the honest record of what you don't know yet. Despite what some people may tell you, not knowing something is OK. It just means you have more work to do. This playbook was written to help you understand the steps, and the Opportunity Backlog section is where the getting-data steps live.
The card, filled
This is the invoicing case you can follow through the whole playbook. A company whose supplier customers invoice their own customers through the tool, and get paid weeks late. The full run, conversations and numbers and all, is told on the worked example page. The company and its numbers are invented to make the mechanics concrete; the method and the industry benchmarks are real. Here is the card that run produced.
Problem
When I invoice my customers,
I want to be paid by the due date,
So I can cover payroll and plan my cash without borrowing to bridge the gap,
But a big share of my invoices get paid weeks late, and I usually find out only after the date has already passed.
Users
Small and mid-size suppliers billing other businesses. The owner who feels it in her cash line, and the bookkeeper who fights it every Friday.
Expected Outcome
We expect that suppliers get paid sooner. Days-to-payment drops, and more of their cash sits in their account instead of somebody else's.
Market Size
Our own base, 4,000 supplier customers, with the small end both the majority and the most exposed. The industry says the problem isn't niche. 51% of businesses are paid late on average (Taulia, 2023), and 43% of credit-based B2B sales go overdue (Atradius, 2025). For a supplier doing $5M a year, that's roughly $13,700 booked per day, so every day of lateness parks another $13,700 in someone else's account.
Competitors
Factoring companies (cash now, painful slice of margin). Collections agencies (they chase never-paid debts, not late ones, and torch the relationship). Other invoicing tools (basic late reminders). And the real incumbent, the bookkeeper's Friday phone calls.
Our differentiator
The invoice already flows through us, so we sit at the exact moment payment happens.
How might we?
How might we help suppliers get paid on time without souring the relationships they depend on?
Hypothesis
We believe that a one-click payment link in the invoice email will get customers paid faster,
For suppliers and the payers who owe them,
We can watch days-to-payment against a control group and see the difference.
Measures of success
Baseline days-to-payment, measured before anything changes. Days-to-payment for invoices carrying the link, against a control group that doesn't have it. A short payer survey for the why behind the lateness. Link clicks are named as a vanity trap up front: they will go up and prove nothing.
Feasibility
The payment processor we already integrate with offers hosted payment pages, so a test needs almost nothing built. A real product adds reminder scheduling and reconciliation. That's weeks, not quarters.
MVE
We can test this theory by hand-sending a reminder email a few days before the due date, carrying a live pay link to the processor's hosted page. No software. Two people and a shared inbox.
Number of users tested
12 supplier customers opted in, roughly 600 invoices across two billing cycles, about eight weeks.
What did we expect?
Invoices with the reminder-plus-link get paid meaningfully faster than the control group. Written down before the data landed.
Experiment Success?
Yes. Median days-to-payment dropped from 58 to 49 while the control held at 57, and the payer survey confirmed the why. Paying was a chore, and the link removed the chore. Build door. The card left the opportunity backlog for the product backlog.
That card is the record of the whole run. Whoever picks this problem up next, you in six months or a teammate who never sat in the interviews, starts from what's written on it instead of from zero. It doesn't matter how the run ends. Build, pivot, or archive, the card keeps what was learned.
The full run, every conversation and number that filled this card, is told on the worked example page.
Copy it in
Copy the blank into your tracker today and put the next problem on it, the day you hear it, while the complaint is still specific. Fill the Problem and Users fields from real conversations, not from what you assume the complaint meant. And when a field fills, fill it that same day. Memory smooths over the interesting parts.
When a card finishes at the build door, it becomes the next artifact up the chain, the epic.