Grokking the Forward Deployed Engineer Interview
Vote

0% completed

​

Writing a One Page Scope

What a One Page Scope Is

The Sections

The Insurer's Scope

Writing the Out of Scope Section

Reviewing and Signing

Using the Scope During the Engagement

Using This in an Interview

Key Takeaways

Practice Questions

Title image: one page that records the problem, what is in and out, the measures, the plan, and the sign-off.
Title image: one page that records the problem, what is in and out, the measures, the plan, and the sign-off.

Before a builder renovates a kitchen, the owner receives a written quote listing what will be done and what will not, with a price and dates. When a question arises in the third week, both sides read the quote instead of relying on memory.

An FDE engagement needs the same kind of document, written after only a few days with the customer and short enough to fit on one page.

What a One Page Scope Is

A one page scope is a short written agreement that states the problem, what will and will not be built, how success is measured, the plan, and who must do what. It is written at the end of discovery, before the prototype starts. A sponsor reads one page in five minutes.

The scope collects the answers from discovery: the user and the decision, the data map, the measures, and the smallest version. Every line repeats something the customer has already said or approved.

The Sections

The insurer's scope had nine sections, each answering one question that either side could later ask.

SectionThe question it answers
ProblemWhat is wrong today, in numbers?
User and decisionWhose work changes, and which choice does the system make?
In scopeWhat will be built in this engagement?
Out of scopeWhat will not be built, even if someone asks for it later?
Data and accessWhich sources are needed, and who grants access to each?
MeasuresWhich numbers decide success, with the baseline and the target?
PlanWhat happens in each week?
Risks and dependenciesWhat could delay the plan, and who must act?
Owners and sign-offWho approves the scope, and who runs the system afterward?
The layout of the one page scope: nine sections from the problem at the top to the sign-off at the bottom.
The layout of the one page scope: nine sections from the problem at the top to the sign-off at the bottom.

The Insurer's Scope

The FDE wrote the first draft at the end of the first week. This is the signed version in full.

SCOPE: Automatic claim routing, pilot and production (6 weeks) PROBLEM About 1,200 new claims a day are routed by hand into six queues, at about 5 minutes each: about 100 adjuster-hours a day. A claim waits 2 days on average to reach the right queue. An earlier demo misrouted 1 claim in 4. USER AND DECISION Users: the adjusters who route new claims. Decision: which of the six queues a new claim belongs in. IN SCOPE - Read new claims, photos, and police reports through the claims API. - Convert scanned police reports to text; extract a short list of facts. - Fixed rules, including: policy younger than seven days goes to possible fraud, where an adjuster reviews it. - Suggest a queue and a score for every claim; route automatically above each queue's threshold. - Decision log of every automated decision, as the regulator requires. - A switch that sends every claim back to adjusters. OUT OF SCOPE - Damage assessment and payment decisions. - A new screen for adjusters (suggestions appear in the existing notes field). - Changes to the model. - Automatic routing for possible fraud, at any time. DATA AND ACCESS - Claims API and document store: approved by security on day 4. - Policy API (read-only), owned by the policy team: approved on day 6. - Queue history: 2,000 past claims, final queue used as the correct answer. MEASURES (agreed) - Routing work: from about 100 to under 40 adjuster-hours a day. - Time to the right queue: from 2 days to under 1 hour for automatic claims. - Automation rate on the 2,000 past claims: at least 60 percent. - Agreement with the final queue: at least 95 percent. - Guardrails: no automatic fraud decisions; every decision logged. PLAN Week 1 Discovery: read 300 claims with adjusters (done) Weeks 2-3 Prototype on the 2,000 past claims; report the measures Weeks 4-5 Pilot: glass and minor damage only; adjusters check 1 in 10 Week 6 All six queues; handoff to the insurer's 4 engineers RISKS AND DEPENDENCIES - Policy API outages: those claims go to adjusters (policy team). - Scanned reports may read poorly; theft claims depend on them (FDE). - Compliance officer to confirm the decision log fields by end of week 2. OWNERS AND SIGN-OFF Sponsor: claims manager. Operators after week 6: insurer engineering (4). Go or no-go for the pilot: claims manager, after the prototype measures.

The page uses the customer's words, like "queues" and "adjusters", and states numbers rather than hopes. Every risk names who must act.

Writing the Out of Scope Section

The out of scope section prevents more disputes than any other section. Requests to add work arrive in every engagement, often in the third or fourth week. When the scope already lists the item, the answer is easy to give and easy to accept.

The first three items were ideas heard during discovery, from the claims manager, an adjuster, and an engineer. Writing them down showed that each was heard and postponed, not forgotten. The fourth, automatic routing for possible fraud, is not postponed but excluded for the whole engagement. A scope can hold both kinds of item, but it should make clear which is which.

Reviewing and Signing

Three short reviews came before the signature, each with a different reader.

  • Two adjusters read the user and decision sections and the plan. They confirmed that the notes field was where they would look for suggestions.
  • The insurer's engineers read the data and access section. They added the document store, which the first draft had left out.
  • The compliance officer read the measures and the in-scope list, confirmed the decision log in principle, and asked to check its fields by the end of week 2.

The claims manager signed the revised page on the first day of week 2.

Using the Scope During the Engagement

The FDE used the page in three ways. Each weekly update followed the measures section. Each new request was checked against the out of scope list before anyone discussed it. The go or no-go decision, whether to start the pilot, followed the page: the claims manager decided after the prototype measures.

When the plan must change, change the page and sign it again. In week 3, the compliance officer asked for one more log field, the version of the rules in use. The FDE added it to the in-scope list and noted about one day of extra work. The claims manager then signed the new version.

Using This in an Interview

Take-home projects and decomposition rounds often end with "what would you agree on with the customer before building?" Name the sections of a one page scope with one concrete line each, especially the out of scope list and the measures.

"I would write one page for the claims manager to sign. It would hold the problem in numbers, the user and decision, in and out of scope, data and access, measures with targets, a weekly plan, and owned risks. Out of scope would include payments and automatic fraud routing."

Expect two follow-up questions:

  • "The customer adds a new request in week 4. What do you do?" Check the out of scope list. If it is new and important, change the page, show its cost in time, and sign again.
  • "Why only one page?" A sponsor reads and remembers one page, and the limit forces you to state only what both sides must agree on.

Key Takeaways

  • A one page scope states the problem, the user and decision, what is in and out, the data, the measures, the plan, the risks, and the owners.
  • The out of scope section prevents the most disputes, because it records ideas that were heard and postponed.
  • Review the page with users, engineers, and compliance before the sponsor signs it.
  • When the plan changes, change the page and sign it again.

The next lesson, What Decomposition Means, explains how to break a customer problem into parts that can each be built and tested.

Practice Questions

1. Why does the scope include an out of scope section at all, when it already lists what will be built?

<details><summary>Show answer</summary>

Because requests to add work arrive during every engagement, and a written list makes the answer easy. Listing heard ideas as postponed means declining them later does not start a dispute.

</details>

2. The first draft of the insurer's scope left out the document store. How was that found?

<details><summary>Show answer</summary>

The insurer's engineers reviewed the data and access section before the sponsor signed. A review by the people who run the systems catches gaps the FDE cannot see.

</details>

3. Why must every item in the risks section name who must act?

<details><summary>Show answer</summary>

Because a risk without an owner is not acted on. Naming the policy team and the compliance officer told the sponsor whom to ask when a risk started to delay the plan.

</details>

Reading Progress

0%


Vote for new content

On This Page

What a One Page Scope Is

The Sections

The Insurer's Scope

Writing the Out of Scope Section

Reviewing and Signing

Using the Scope During the Engagement

Using This in an Interview

Key Takeaways

Practice Questions