# Stakeholder Interview Guide

> Generated in Step 2. Customise per role before each session.
> Run individual sessions before any group alignment workshop.

---

## Interview metadata

| Field | Value |
|---|---|
| Company | |
| Interviewee name | |
| Role | |
| Department | |
| Date | |
| Duration | |
| Interviewer | |
| Session type | Individual / Workshop |

---

## Role-specific focus

Select the appropriate section based on interviewee role.

---

### Executive / Business Owner

**Goal:** Understand KPIs they are accountable for, strategic decisions, and what "good" looks like.
**Do NOT:** Ask about data details, source systems, or report mechanics.
**Duration:** 30–45 minutes.

**Opening (5 min)**
> "Before we start — can you describe your role and what you're primarily measured on?"

**Core questions:**
1. What are the 3–5 numbers you look at every week to understand if the business is healthy?
2. What was the last important decision you made that relied on data? What data did you use?
3. If you could answer one question about the business that you currently can't, what would it be?
4. Where do you feel the biggest gaps are in your current reporting?
5. What does success look like for this data project in 12 months?

**KPI capture:**
| KPI name | How often reviewed | Current source | Confidence in accuracy |
|---|---|---|---|
| | | | |

---

### Business Analyst / Operations Manager

**Goal:** Understand the real definitions, the actual reports used, and the workarounds.
**This is your highest-value interview — go deep.**
**Duration:** 60–90 minutes.

**Opening (5 min)**
> "Can you walk me through a typical week — what reports do you run, what decisions do you make?"

**Process mapping (20 min)**
1. Walk me through what happens when [key transaction] occurs — step by step.
2. Where does the data for this process live? Which systems?
3. What's the highest-volume event in your area? How many happen per day?
4. Are there any processes that happen off-system — in spreadsheets or email?

**Definition probing (20 min)**
5. How do you define [customer / revenue / active user / churn]? *(use the key term for their domain)*
6. If two people pulled [key metric] independently, would they get the same number? Why not?
7. What workarounds do people use because the data isn't quite right?
8. What reports do people not trust? Why?

**Dimension discovery (15 min)**
9. How do you slice your analysis? By [region / product / channel / segment]?
10. What attributes do you always want to see alongside the numbers?
11. Are there any filters you apply as a matter of habit? Why?

**Fact discovery (15 min)**
12. What are your most important measures?
13. For [key metric] — can you sum it across all [channels / regions / products]? *(additivity probe)*
14. How is [key metric] calculated exactly? Is there a documented definition?
15. What questions do you get asked that you can't currently answer?

---

### Power User / Report Consumer

**Goal:** Understand the real grain of analysis, daily workflow, and Excel workarounds.
**Duration:** 45–60 minutes.

**Opening (5 min)**
> "Can you show me the reports or dashboards you open first thing every day?"

**Core questions:**
1. What's the most granular level you ever need to analyse data? One row per what?
2. What do you do when a report doesn't have the breakdown you need — do you export to Excel?
3. Can you show me an example of a spreadsheet you built that pulls from the current system?
4. What's the most frustrating thing about the current reporting setup?
5. If you could add one column to your most-used report, what would it be?
6. Are there numbers you calculate manually that you wish were automated?

**Grain confirmation:**
> Present a proposed grain statement and ask: "If I showed you a report with one row per [X], would that meet your needs? Is there anything you'd need that this wouldn't support?"

---

### Data / Systems Owner (DBA, ERP Admin, CRM Manager)

**Goal:** Confirm source system availability, data quality, and grain of raw data.
**Run this session AFTER the draft model is sketched — bring the fact and dimension candidates.**
**Duration:** 45–60 minutes.

**Opening:**
> "I'm going to walk you through the dimensions and facts we've identified. For each one I'd like to understand: does [system] capture this, at what grain, and what's the data quality like?"

**Per dimension:**
1. Is [dimension entity] tracked in [system]? As a single table or spread across multiple?
2. What's the primary key? Is it stable over time?
3. Are historical changes tracked, or does the record get overwritten?
4. What attributes are available vs. what we need?

**Per fact:**
5. Is [event] captured in [system]? At what grain?
6. Is [measure] stored directly or does it need to be calculated?
7. Are there any known data quality issues — gaps, duplicates, inconsistent values?
8. How far back does historical data go?

**Integration:**
9. What integrations exist today between systems?
10. Is there an existing ETL or data pipeline? How is it maintained?
11. Are there any systems we haven't mentioned that are relevant?

---

## Cross-cutting questions (ask in every session)

These surface conflicts and hidden requirements regardless of role:

- "How do you define [customer / revenue / active / churn]?" *(adapt to domain)*
- "What reports do you use today, and what's wrong with them?"
- "What questions can you not answer today that you wish you could?"
- "If two people pulled this number independently, would they get the same answer? Why not?"
- "Is there anything in a spreadsheet today that should be in a report?"

---

## Interview notes capture

Use this structure during the session:

### Business processes identified
> (Candidate fact tables)

| Process | Description | Volume / frequency | Systems involved |
|---|---|---|---|
| | | | |

### Dimensions mentioned
| Dimension | Attributes mentioned | Notes |
|---|---|---|
| | | |

### Facts / KPIs mentioned
| Fact / KPI | Calculation notes | Additive? | Source |
|---|---|---|---|
| | | | |

### Definition conflicts surfaced
| Term | Definition given | Conflicts with | Resolution needed |
|---|---|---|---|
| | | | |

### Open questions
1.
2.
3.

### Source systems mentioned
| System | Purpose | Data owner |
|---|---|---|
| | | |