Skip to content

Lesson 1 of 6

What Scoping Is: The Ask, the Problem and the Build

Learn to keep three things apart (what the customer asks for, the problem behind it and what you will build) and see what a finished piece of scoping looks like.

4 minForward Deployed Engineering

After this lesson you will be able to tell a customer's ask from the problem behind it, describe what scoping has to produce, and set up the one-page plan that the rest of this course fills in.

Why asks arrive vague

A customer rarely arrives with a specification. They arrive with a sentence: "we want a dashboard", "we want to automate onboarding", "we want an AI agent for our emails". The sentence is short because the person saying it is busy, because they are describing a solution they have seen somewhere else, and because nobody has yet looked closely at what goes wrong.

That is not a fault in the customer. Working out what to build is part of what a Forward Deployed Engineer (FDE) is paid for. An engineer who waits for a clear specification will wait a long time, and an engineer who builds exactly what the first sentence said will often build the wrong thing.

Three things to keep apart

Scoping is the work of separating three things that usually arrive stuck together.

  • The ask is what the customer says they want. It is nearly always a solution: an agent, a dashboard, an integration.
  • The problem is what is going wrong today, for whom, and what it costs. It exists whether or not anyone builds anything.
  • The build is what you agree to make first. It should be the smallest thing that removes a useful part of the problem.

One ask can hide several problems. "We want a dashboard" may mean that managers cannot see late orders, or that two reports disagree, or that somebody spends every Monday copying numbers between spreadsheets. Each of those leads to a different build. You find out which one is real by asking and by looking at evidence, and that is what the next two lessons cover.

What scoping produces

Scoping is finished when there is a short written plan that the customer has read and agreed to. A conversation that felt good is not a result. One page is enough, and one page gets read. It answers nine questions:

  • Problem: what is going wrong, for whom?
  • Evidence: what did you see or measure that shows it?
  • Success measure: what number will change, from what, to what?
  • In scope: what will be built?
  • Out of scope: what will not be built, even though somebody asked?
  • First piece of work: what will the customer be able to see first, and when?
  • Milestones: what follows, in steps that each end in something visible?
  • What we need from the customer: access, data and people's time.
  • Risks: what could make the plan wrong, and how will you find out early?

Every lesson in this course fills in part of that page.

What goes wrong without it

Skipped scoping shows up later, in recognisable ways:

  • The build works and nobody uses it, because it solved a problem the users did not have.
  • The work never ends, because nobody wrote down what "done" means.
  • Requests keep arriving and each one sounds small, because nothing says what is out of scope.
  • The customer is disappointed by a build that does what was asked, because the ask was never the problem.
  • The project stalls in its second week waiting for access that nobody asked for at the start.

None of these is a coding failure. Each one was decided, or left undecided, before the first line of code.

The example this course uses

The rest of this course follows one made-up customer. Harrow Building Supplies sells building materials to trade customers such as builders, roofers and kitchen fitters, and delivers them by truck. Its operations director opens your first call with this:

"We want an AI agent that handles our order emails. How long would that take?"

Notice what the sentence contains. There is a solution (an AI agent), a place to put it (order emails) and a request for a date. There is nothing about what is going wrong. By lesson 5 you will have a plan for Harrow that the operations director can agree to, and it will not be the thing that sentence asked for.

Create a folder called scoping. Inside it, create a file called plan.md with these nine headings and nothing under them yet:

# Build plan: Harrow Building Supplies

## Problem

## Evidence

## Success measure

## In scope

## Out of scope

## First piece of work

## Milestones

## What we need from the customer

## Risks

An empty page with headings is a useful thing to bring to a first meeting. It tells you what you still do not know, and every gap is a question to ask.

Key takeaways

  • The ask is the customer's idea of a solution; the problem is what is going wrong today; the build is what you agree to make.
  • A vague ask is normal and is not a fault in the customer.
  • Scoping ends with one page that the customer has agreed to, not with a conversation.
  • Most failed customer projects built the wrong thing well, and the cause was skipped scoping.

Try it

Open plan.md and, under the heading "Problem", write three different problems that could be behind the ask "We want an AI agent that handles our order emails". For each one, write the single question you would ask to find out whether it is the real problem. A good result has three problems that would lead to three different builds, and questions that a person on the customer's side could answer from memory.

Official documentation

Ready to use this in a job?

Rolebound applies to matching roles for you.

Start now

Written with AI and checked automatically. Nobody has run these steps by hand, and tools change, so a command may behave differently in the version you install. Spotted a mistake? Tell us.

    Cookies on Rolebound

    We use necessary cookies to sign you in and keep the site working — those cannot be turned off. We would also like to set analytics and advertising cookies so we can see which pages help people get interviews and which adverts brought them here. Nothing optional is set unless you say yes, and you can change your mind at any time from “Cookie settings” in the footer. Read our Privacy Policy.