Every company starts with spreadsheets.
Founder tracks metrics in Excel. Works great for a while.
Then you hire a team. Now 10 people are asking for the same numbers. Excel becomes a bottleneck.
Eventually you need to graduate to a metrics system.
The Evolution: Spreadsheets → Dashboards → Data Warehouse → Data Platform
Stage 1: Spreadsheets (0-20 people)
What: Manually pull data from your tools into a spreadsheet
Tools: Google Sheets, Excel
Metrics tracked: Revenue, users, churn (basics)
Frequency: Weekly or monthly
Pros:
- Free or cheap
- Requires no technical setup
- Good for starting out
Cons:
- Manual work and error-prone
- Data gets stale
- Hard to drill into data
- No historical tracking
- Doesn't scale
When to upgrade: When you spend more than 5 hours per week manually updating numbers.
Stage 2: BI Dashboards (20-100 people)
What: Connect your data sources to a dashboarding tool (Tableau, Looker, Metabase)
Tools: Tableau, Google Data Studio, Metabase, Looker, Power BI
Metrics tracked: All operational metrics, with drill-down capability
Frequency: Real-time or daily refresh
Pros:
- Automated data pulls (no manual work)
- Current data
- Can drill into segments and trends
- Good visualizations
- Team can build their own reports
Cons:
- Requires technical setup (data connectors)
- Data quality issues become visible
- Hard to do complex analysis
- Limited to pulling from existing systems
When to upgrade: When you need analytics beyond basic metrics, or your data sources are too fragmented.
Stage 3: Data Warehouse (100-500 people)
What: Build a centralized database combining data from all sources
Tools: Snowflake, BigQuery, Redshift, Postgres (self-hosted)
Metrics tracked: All metrics, at any granularity
Frequency: Real-time
Pros:
- Unified source of truth for all data
- Can combine data from multiple sources
- Enables complex analysis
- Historical data for trends and cohorts
- Supports growth at scale
Cons:
- Requires dedicated data engineering
- Complex setup and maintenance
- Expensive infrastructure
- Requires technical knowledge
When to upgrade: When your data gets fragmented (multiple sources, inconsistent definitions), or you need to do cohort analysis and complex joins.
Stage 4: Data Platform (500+ people)
What: Build self-service analytics where teams can answer their own questions
Tools: dbt, Looker, (custom data apps), data catalogs
Metrics tracked: Thousands of metrics, self-service
Frequency: Real-time
Pros:
- Teams can answer their own questions
- Metrics are discoverable and documented
- High data quality and consistency
- Enables scalable analytics
Cons:
- Highest complexity
- Requires skilled analytics engineering team
- Expensive to build and maintain
- Needs strong governance
When to upgrade: When you have 50+ analytics requests per week and your analytics team is a bottleneck.
How to Build a Metrics Infrastructure
Step 1: Audit Your Current State
- What metrics do you track?
- How do you currently track them?
- Who uses them?
- How much time is spent maintaining them?
- What's broken or missing?
Step 2: Choose Your Stack
For Stage 1-2 (spreadsheets → dashboards): Start with one data tool:
- Google Data Studio: Free, easy setup, connected to Google tools
- Metabase: Simple, good for small teams, self-hosted
- Looker Studio: Free but limited
- Tableau Public: Free public version
Then add connectors:
- Connect Salesforce
- Connect Stripe
- Connect Google Analytics
- Connect Mixpanel/Amplitude
Step 3: Prioritize Core Metrics
You can't ingest all data at once. Start with core metrics:
- Revenue metrics (MRR, ARR, churn)
- Operational metrics (users, engagement, NPS)
- Financial metrics (burn rate, runway, CAC)
Get those working perfectly. Then expand.
Step 4: Establish Data Ownership
Who owns each metric?
- Revenue metrics → Finance
- Product metrics → Product
- User metrics → Growth
- Operational metrics → CEO/Board
Make them responsible for accuracy and updates.
Step 5: Build Governance
Create simple rules:
- Data definitions are documented
- Data sources are documented
- Data freshness targets (when should data update?)
- Data quality checks (alerts if data looks wrong)
- Access control (who can see what data?)
Step 6: Regular Reviews and Refinement
Monthly meetings to discuss:
- Are metrics healthy?
- Do we trust the data?
- What new metrics do we need?
- What metrics can we retire?
Common Infrastructure Mistakes
❌ Mistake 1: Building Too Much Too Soon
You don't need a data warehouse on day 1. Start simple: spreadsheet → dashboard → warehouse → platform.
❌ Mistake 2: No Data Governance
Without definitions and ownership, every report tells a different story.
Establish governance early.
❌ Mistake 3: Building Instead of Buying
There are good products for each stage. Use them instead of building custom solutions (unless you have a dedicated team).
❌ Mistake 4: Ignoring Data Quality
Garbage in, garbage out. Invest in data quality from the start.
❌ Mistake 5: Poor Integration Planning
Your Salesforce has the truth. Your spreadsheet has the truth. Your warehouse has the truth. Now you have three truths.
Plan integrations carefully.
Quick Start: Building a Metrics System This Week
Day 1:
- List all metrics currently tracked
- List all data sources
- Identify pain points
Day 2:
- Choose a BI tool (Google Data Studio or Metabase if starting out)
- Connect your top 3 data sources
Day 3:
- Create core metrics dashboard
- Set up automated daily refresh
Day 4:
- Document metric definitions
- Share with team
Day 5:
- Weekly metrics review meeting with team
- Collect feedback and iterate
Result: Automated metrics that everyone can access.
The Bottom Line
Your metrics infrastructure should:
- Reduce manual work
- Improve data quality
- Enable self-service analytics
- Support team alignment
Start with what you have. Graduate to the next level when you outgrow your current system.
Don't over-build. But don't under-invest either.
Your metrics infrastructure is the foundation of data-driven decision-making.