Communicating with Data
Working with Data Teams
Data teams are your partners, not just service providers. Understanding how they work and speaking their language leads to better outcomes for everyone.
Understanding Data Team Roles
Modern organizations have specialized data roles. Knowing who does what helps you reach the right person.
| Role | What They Do | When to Contact |
|---|---|---|
| Data Analyst | Explores data, creates reports, answers business questions | Need insights, reports, or ad-hoc analysis |
| Data Scientist | Builds predictive models, ML algorithms | Need forecasting, recommendations, or advanced analytics |
| Data Engineer | Builds data pipelines, maintains infrastructure | Data quality issues, new data source integration |
| Business Intelligence (BI) | Creates dashboards, maintains reporting tools | Dashboard requests, visualization changes |
| Analytics Engineer | Transforms raw data into analysis-ready tables | Need new metrics defined, data model changes |
Pro tip: When in doubt, start with a Data Analyst or your BI contact—they'll route you correctly.
Speaking the Data Language
Essential Vocabulary
| Term | What It Means | Example Usage |
|---|---|---|
| Metric | A measurable value | "What's the definition of the 'active users' metric?" |
| Dimension | A category for grouping data | "Can we see revenue by region dimension?" |
| Aggregation | Combining data (sum, average, count) | "Is this a sum or average aggregation?" |
| Grain | The level of detail in data | "What grain is this data—daily or monthly?" |
| Filter | Limiting data to specific criteria | "We need to filter to enterprise customers only" |
| Segment | A group within your data | "How does this segment compare to others?" |
| Cohort | A group based on time of first action | "Show me the January 2024 signup cohort" |
| YoY / MoM | Year-over-year / Month-over-month | "What's the YoY growth rate?" |
Phrases That Help
| Instead of This | Say This | Why It's Better |
|---|---|---|
| "I need all the data" | "I need [specific metrics] by [dimensions] for [time period]" | Specific is actionable |
| "Can you make it look better?" | "Can we highlight [specific insight] more prominently?" | Clear direction |
| "Something's wrong" | "The data shows [specific issue] when I filter by [X]" | Reproducible problem |
| "When will it be done?" | "What information do you need from me to prioritize this?" | Collaborative approach |
Making Effective Data Requests
The Perfect Request Template
Every data request should include:
REQUEST TEMPLATE
━━━━━━━━━━━━━━━━
WHAT I NEED:
[Specific metrics, visualizations, or analysis]
WHY I NEED IT:
[Business decision this supports]
HOW I'LL USE IT:
[Presentation, report, dashboard, one-time decision]
TIMELINE:
[When you need it, and any flexibility]
CONSTRAINTS:
[Time period, segments, filters required]
DEFINITIONS:
[Your understanding of key terms - ask for confirmation]
Good Request vs. Bad Request
❌ Bad Request:
"Can you pull the sales data? I need it for a meeting."
Problems:
- What sales data? (Product? Region? Time period?)
- Which meeting? (When? Who's attending?)
- How will it be used? (Quick number? Detailed breakdown?)
✅ Good Request:
"I need Q3 2024 sales data broken down by product category and region, compared to Q3 2023. This is for our board meeting on Friday—I'll be presenting a slide showing YoY growth by category. I'm defining 'sales' as recognized revenue. Can you confirm that matches your definition?"
Why this works:
- Specific metrics and dimensions
- Clear time period and comparison
- Context for how it will be used
- Timeline provided
- Definitions clarified upfront
Asking the Right Questions
Questions About Data
| Question | What You Learn |
|---|---|
| "Where does this data come from?" | Source system, reliability |
| "How often is it updated?" | Freshness, when to check for updates |
| "What's the definition of [metric]?" | Calculation method, inclusions/exclusions |
| "Are there known data quality issues?" | Limitations, caveats to mention |
| "What time zone is this data in?" | Avoids comparison errors |
| "Is this data audited/verified?" | Level of confidence you should have |
Questions When You Get Unexpected Results
| Question | What You're Checking |
|---|---|
| "Are any filters applied that I might not see?" | Hidden filters affecting results |
| "Did anything change in how we track this?" | Definition or source changes |
| "Is this a complete dataset or sample?" | Data completeness |
| "What's the time period for this data?" | Correct comparison period |
Common Data Request Mistakes
Mistake 1: Not Providing Context
❌ "Give me the churn rate." ✅ "I'm presenting to the CEO about customer retention. I need our churn rate for Q3, compared to Q2 and Q3 last year, with a breakdown by customer segment."
Mistake 2: Asking for "All the Data"
❌ "Can I get a data dump of all our customer information?" ✅ "I need customer count, average order value, and purchase frequency by segment for the last 12 months."
Mistake 3: Unclear Timelines
❌ "I need this ASAP." ✅ "I have a meeting on Thursday at 2pm. Ideally, I'd have this by Wednesday EOD to review. If that's not possible, what's realistic?"
Mistake 4: Moving Target Requirements
❌ Asking for one thing, then continuously expanding scope ✅ Start with your full need; add-ons go in a separate request
Receiving Data: Best Practices
What to Check When You Get Data
DATA QUALITY CHECK
━━━━━━━━━━━━━━━━━
□ Does the total match what you expected?
□ Are there any obvious outliers or zeros where there shouldn't be?
□ Does the time period match your request?
□ Are all requested dimensions present?
□ Do the definitions match what you asked for?
□ Is the format usable for your purpose?
What to Do If Something Looks Wrong
-
Document the discrepancy: "The Q3 revenue shows $4.2M, but our finance report shows $4.5M."
-
Ask clarifying questions: "Could this difference be due to how we're defining 'recognized revenue'?"
-
Share your source: "I'm comparing this to the CFO's monthly report from October 3rd."
-
Be open to learning: The difference might be valid (different definitions, timing, adjustments).
Building Good Data Team Relationships
Do This:
| Practice | Why It Matters |
|---|---|
| Share context for your requests | Helps them prioritize and suggest better approaches |
| Say thank you | Recognition motivates |
| Give feedback on what was useful | Helps them improve |
| Ask if they need anything from you | Makes collaboration easier |
| Report issues constructively | Helps fix problems faster |
Avoid This:
| Behavior | Why It's Problematic |
|---|---|
| Treating requests as orders | Creates adversarial relationship |
| Escalating before asking | Damages trust |
| Blaming data teams for bad news | Data reveals reality; don't shoot the messenger |
| Last-minute "urgent" requests | Suggests poor planning on your part |
| Ignoring their input | Wastes their expertise |
The Data Request Lifecycle
1. IDENTIFY NEED
↓ What decision am I trying to make?
2. DEFINE REQUEST
↓ What specific data will inform that decision?
3. SUBMIT REQUEST
↓ Using template with full context
4. COLLABORATE
↓ Answer questions, clarify scope
5. RECEIVE & VERIFY
↓ Check data quality, confirm it meets need
6. APPLY & REPORT BACK
↓ Use the data, share outcome with data team
7. ITERATE (if needed)
↓ Refine request based on what you learned
Key Insight: The best data relationships are partnerships. Invest time in understanding your data team's work, and they'll invest in understanding yours.
Next: Create your personal data literacy roadmap and learn what to do next. :::