The Power User Myth
Users who create a Goal retain six times better, so growth wants to make it a required onboarding step. Work out whether the Goal is what keeps them — or whether staying is what lets them set one — then make the call.
The brief
You have joined Halden, a personal-finance app, as a product analyst.
Your manager forwards a thread from the quarterly planning review:
"Users who create a Goal retain at about 58% at day 60. Users who never create one retain at about 9%. That is six times better, and it is the strongest signal in the whole retention deck. The proposal is to make Goal creation a required step in onboarding and run a push campaign at everyone who has not set one. That is a sprint of work and it makes signup longer. Can someone check the number before we commit?"
Nobody has looked at this comparison any way other than Goal users against everyone else. You have the full session and goal history for the January signup cohort.
What you are asked to do: establish whether creating a Goal is what keeps people, and give a straight answer on making it mandatory.
All timestamps are UTC, and there is no other timezone to reconcile. Everyone in the data signed up in January 2026 and the session log runs to 31 May 2026, so every user has at least 120 days of history — nobody's retention is cut short by the end of the file.
Three notes from the team, worth reading before you start:
The onboarding flow used to create a starter goal for new users automatically, before they had chosen anything. Those rows are flagged in the data.
Someone who sets one goal usually sets more later, and each goal is its own row.
Nobody has checked whether Goal users were already different from everyone else before they set a goal.
Treat this as you would any claim that one behaviour causes another: work out who could have ended up in each group, and when, before you compare them — and be ready to show how you got every figure you report.
What to submit
- 1Your five verified answers, entered on the submission form.
- 2A one-page recommendation in your repo README — whether Goals keep people, how you know, and whether to make it mandatory. One page, not five.
- 3The SQL, Python or R you used to get your five answers, in the same repo.
- 4Optional: a chart, a dashboard link, or a three-minute Loom. None of these are required and none of them earn marks on their own.
How it is scored
| Area | Weight |
|---|---|
Data accuracyauto-graded The five verified answers. Graded automatically against the dataset, so this half of your score is arithmetic, not opinion. | 40 |
Problem framingreviewed Did you ask who could have landed in each group in the first place, rather than accepting the two groups as given? Restating the brief scores low. | 15 |
Analysis and methodreviewed Did you fix the comparison at a point in time instead of on an outcome that can only happen to survivors, and did you exclude the onboarding-created goals — and show how? Correct numbers with no visible method score low. | 20 |
Insight qualityreviewed Naming the bias is not the finding. Showing how much of the six-times claim it accounts for, and being honest that what survives the correction is still confounded rather than proven, is. | 15 |
Recommendationreviewed A clear answer on the mandatory onboarding step, argued from your own numbers, plus what you would run instead to answer the question properly. Hedging scores low. | 10 |
The data
Three files, about 4 MB total.
users.csv — one row per user. Columns: user_id, signup_date, acquisition_channel, platform, country. Everyone signed up in January 2026. platform is ios or android, written with inconsistent capitalisation, and some country values are blank. Neither is a trap; neither changes an answer.
sessions.csv — one row per app open. Columns: user_id, session_ts. Runs to 31 May 2026.
goals.csv — one row per goal created. Columns: user_id, created_at, goal_type, source. source is either user, meaning the person created the goal themselves, or onboarding_default, meaning the onboarding flow created it for them.
The data is synthetic, generated for this challenge, and contains the kind of mess real product tables contain. No user in it is a real person.
Results
All 3 submissions, scored on diagnosis, rigour, judgment and communication.
- 1katendeWinnerPerfect Objective Score96
- 2Misbha A.Top 3Perfect Objective Score88
- 3HaleemaTop 3Perfect Objective Score87
Scores are 40% verified answers, graded automatically, and 60% written analysis against the published rubric. All-time leaderboard
Recognition
Free to enter — everyone is scored and ranked
Every entrant is graded against the same rubric, ranked on the public leaderboard, and gets written feedback on their analysis — no fee, no prize, just the credential.
- The full brief and dataset
- Your score on the verified answers
- Written feedback on your analysis
- Ranked on the public leaderboard
- A certificate and badges you can add to your CV
Rules
- Scoring follows the published rubric below: 40% verified answers, graded automatically, and 60% your recommendation, scored against the rubric
- Scoring is automated. The top 10 submissions and any disputed score are reviewed by a human before results are finalized
- Your repository must be public and its commit history must show the work happening before the deadline
- Work individually. You may use any tool, including AI assistants — but you must be able to explain every number you submit
- One submission per person. You can update it any time before the deadline
- Everyone who submits is ranked by name on the public leaderboard when results are published
- Your repo has to be public for us to read it — but do not promote your solution on LinkedIn, X or Discord until judging closes, so others can still attempt it
