# Propose a north star the team can trust

Cartwhistle · Product-management work sample
All company, customer and data in this exercise are original synthetic material.

## Situation and outcome
Cartwhistle is a fictional marketplace where independent restaurants and cafés reorder from local food and drink suppliers. It earns a 9% commission on delivered GMV. You are the product manager for Buyer Ordering: the ordering app, order guide, reorder lists, substitution preferences and auto-substitute. Each product area proposes one north star, inputs and guardrails to the product leadership review on 16 October, and the Head of Product adopts the set (R001, R002). An analytics agent drafted a first proposal overnight. Nobody has checked it.

Your outcome is a proposed indicator set for that review, marked proposed, with a feasibility review of every metric in it. It holds one north star, three to five inputs with owners and three to five guardrails, each with a definition, a source and a baseline or the date one will exist (R002). You cannot query data or commit other teams, so where no baseline exists write "baseline unavailable, date pending agreement with Data"; that is an accepted answer. A well-argued "the north star cannot yet be validated, so here is a staged proposal" is a valid result, and so is an honest evidence gap. The set does not adopt itself, set targets, specify or build instrumentation, or cover Supplier Tools and Payments metrics.

The period is 2026-08-24 through 2026-10-02 inclusive. Work with 56 records (strategy and finance notes, data definitions and analyses, interviews, support tickets, churn notes, supplier notes, team messages and outside articles) alongside an AI first draft. The AI draft is a draft, not an authority. Use the supplied record IDs and source context; do not contact the fictional teams or seek private data or live queries.

## Your role
You are joining the team as Product Manager.

You own the proposed north star, inputs and guardrails for Buyer Ordering, marked proposed for the 16 October leadership review, with owners proposed and other teams' agreement stated as pending. You may not adopt the set, set targets, commit Supplier Success or Data to work, or change finance's definitions.

## About the team
A marketplace product team proposing a north star and indicator set for its area to a leadership review.

An analytics agent queries the warehouse and posts to #metrics, the research tool files interviews to accounts, and an agent drafts a first proposal overnight; people own definitions, owner agreements and adoption.

The Head of Product adopts the set; Finance owns revenue definitions, Data owns the warehouse and events, and Supplier Success owns supplier fill rate.

## Work to complete
1. Frame the value: what a buyer gets from Buyer Ordering that keeps them, and what this area does and does not own. Read sources/coverage-and-access.md before you make any claim about what the records cover.
2. Judge the candidates for a north star against the analyses and the rest of the evidence. Say what each result shows and what it cannot show, and keep evidence that points the other way visible.
3. Define the north star so it can be computed only one way: the event it counts, the unit, the week, what is included and excluded. Label it a hypothesis or evidence-backed, and give a plan to validate it.
4. Derive three to five inputs the team can move and three to five guardrails that would detect harm from pushing the north star. Name an owner for each, and mark owner agreement as pending where it is.
5. Audit the AI first draft against the original records. Keep, correct or reject each line and say why. Carry every correction through the indicator set, the feasibility review and the evidence table.
6. Check feasibility and baselines: for each metric, where the data comes from, whether it can be computed now or is a gap, and its baseline or the date one will exist.
7. Apply the correction note (U001) in your download like the other sources. Carry each justified change through every file, and keep the original reading of any record or figure it changes visible next to the revised one (a before-and-after entry in the revision log is enough; you do not need to keep whole earlier files). The revision log is your own account of how you worked.

## Time
You have 90 minutes of work time. The clock starts when you press Start.

After that comes a 10-minute break, then a written interview: up to eight typed questions about your own files, for about 15 minutes. If you submit early, your break starts straight away.

## Mid-task correction
There is no mid-task correction in this take-home. Everything you need is in the files you download at Start, including the correction note **updates/U001-event-audit-v1.md**. Read it with the other sources, apply it where it changes your work, and cite it as U001.

## Tools and help
Use any tools and help you like: AI tools and agents, search, your own notes, friends or colleagues. You don't need a paid tool, an API key or an account.

Nothing you do in your own tools is watched. Only your files and your interview answers are reviewed. When you submit, you can say which tools or help you used. That is optional, and leaving it empty is never held against you.

## Submission
Hand in one ZIP with these files at its top level. Start from the copies in the templates folder.
- README.md: list every file you hand in and say how you used the correction note (U001).
- indicator-set.md: about 1,000 words at most. The value moment and scope, the candidates you considered and why, the north star definition with its confidence and validation plan, three to five inputs with owners, three to five guardrails with the harm each detects, and the open questions and decisions for leadership. Mark the set proposed.
- feasibility.md: one row per metric in the set: source, computable now or a gap, and baseline or date.
- evidence.csv: already has a row for every record, R001–R056, with the record ID and account ID filled in. Do not retype them. For each row, fill in metric (name your own metrics; write none for a record that bears on no metric) and use (what the record does: supports, contradicts, feasibility, context or excluded are suggested and free text is accepted). If a record bears on more than one metric, add a row for it. You may add a column such as rationale. If the correction note changes a record, revise its row in place and cite it as U001 in a rationale column or the revision log.
- revision-log.md: material changes, verification checks and remaining gaps. No private prompts or unrelated files.

Words are counted in the prose of each document you write; headings, tables, the evidence file and the README do not count.

Use plain text files only: .md, .txt, .csv or .json. Put your work in the files themselves; a link to a document somewhere else does not count. Your ZIP can be up to 5 MB, or 20 MB unzipped, with at most 39 files of up to 2 MB each. Programs, macros, password-protected ZIPs and ZIPs inside your ZIP are not accepted.

Upload your ZIP on your take-home page and wait for **Received and checked**. It normally appears within a few seconds of the upload finishing. If it has not appeared after a minute, upload the ZIP again; if it still does not appear, use **Help and support**. You can replace your ZIP until you submit. Submitting is final and starts your break.

If time runs out before you submit, the last ZIP marked **Received and checked** is handed in for you. If none was received, the team is told that you need support.

If something goes wrong, use **Help and support** on your take-home page. A technical problem is never counted against you. Writing to us does not pause the clock.

## What happens next
The written interview covers these parts of the job:
- Checks sources and counts
- Judges what the evidence supports
- Checks and repairs the AI draft
- Hands off clearly
- Explains their own work
- Defines metrics the team can trust

An AI then drafts a review from your files and your answers. Each point quotes what it rests on. A person at the company reads the review and makes any decision. No one is hired, rejected or ranked automatically.
