Ask¶
Define the problem. Understand stakeholder expectations. Ask before doing.
Understand the problem before jumping into analysis. Solve the right problem with the right answer and right data.
Why this phase matters¶
The Ask phase determines whether the rest of the project is worth doing. A vague question yields a vague answer. The skill is converting fuzzy stakeholder requests into specific, answerable problems.
"If I had an hour to solve a problem I'd spend 55 minutes thinking about the problem and 5 minutes thinking about solutions." — attributed to Einstein
Effective questions¶
- Define the problem you're trying to solve
- Confirm stakeholder expectations
- Focus on the actual problem; avoid distractions
- Collaborate; keep open communication
- See the whole situation in context
Ask yourself:
- What are stakeholders saying their problems are?
- How can I help resolve their questions?
- What decision will be made with this analysis?
- What does success look like?
- What's the cost of being wrong?
SMART Questions¶
| Letter | Meaning | Example |
|---|---|---|
| Specific | Single, narrow topic | "What features drive churn?" |
| Measurable | Quantifiable | "Reduce churn by 5%" |
| Action-oriented | Drives a decision | "Should we launch onboarding emails?" |
| Relevant | Matters to the goal | Aligned with business OKR |
| Time-bound | Has deadline | "By Q3" |
Common problem types¶
| Type | What it asks | Example |
|---|---|---|
| Making predictions | What will happen next? | Will this customer churn? |
| Categorizing things | What group does X belong to? | Is this email spam? |
| Spotting unusual things | What's the outlier? | Which transactions are fraud? |
| Identifying themes | What patterns appear? | What do support tickets cluster around? |
| Discovering connections | What variables relate? | Does ad spend correlate with revenue? |
| Finding patterns | What recurs over time? | When do users drop off? |
Structured Thinking¶
- Problem domain — area of analysis (industry, product, geo)
- Scope of work (SOW) — project doesn't start until SOW is approved
- Deliverables — what work; what artifacts
- Milestones — major progress markers
- Timeline — duration estimates per step
- Reports — how/when to update stakeholders
- Constraints — time, data access, headcount, tooling
- Success criteria — how you'll know you're done
SOW template — make a copy and get approval before starting.
Five Whys¶
When stakeholders state a request, peel back to the root:
- "We need a churn dashboard." — Why?
- "Churn went up last month." — Why does that matter?
- "The board is asking about retention." — What decision will the board make?
- "Whether to fund the customer success team." — What evidence supports that decision?
- "Cohort retention curves and segment-level churn drivers."
→ Build the curves and driver analysis, not just a generic dashboard.
Take good notes¶
- Facts — concrete info: dates, names, specifics
- Context — relevant details to interpret facts
- Unknowns — questions to follow up on
- Decisions — what was agreed and by whom
Example notes:
- Project: collect customer flavor preference data
- Goal: offer/create more popular flavors
- Sources: cash register receipts + customer surveys (email)
- Target: Q2
- TODO: call manager about survey data location
Example questions by audience¶
- Specific: Do you currently use data to drive decisions? What kind?
- Measurable: What percentage of sales is from your top-selling products?
- Action-oriented: What decisions would you make with the right info?
- Relevant: How often do you review business data?
- Time-bound: How did data drive decisions this past year?
- Specific: What data do you use to build lessons?
- Measurable: How well do benchmark scores correlate with grades?
- Action-oriented: Do you share data with other teachers?
- Relevant: Have you shared grading data with a class?
- Time-bound: In the last five years, how often did you review prior-year data?
- Specific: What data drives purchasing/inventory?
- Measurable: Rank these factors on sales: price, flavor, season.
- Action-oriented: Single factor where more data could increase sales?
- Relevant: How do you communicate with customers?
- Time-bound: YoY sales growth for the last 3 years?
Red flags during Ask¶
- Stakeholder can't articulate the decision
- "Just send me everything" requests
- Conflicting goals across stakeholders
- Deadline before scope is clear
- "Make the data say X" — that's confirmation bias, push back
Checklist¶
- [ ] Identified the business task in one sentence
- [ ] Confirmed key stakeholders and decision-maker
- [ ] Agreed on success criteria
- [ ] Documented constraints and assumptions
- [ ] SOW signed off
- [ ] Deliverable: clear statement of business task