Data Analytics and Business Intelligence: Data-Driven SaaS Decision Making

Building a Data-Driven SaaS Organisation

Using analytics and BI to make better, faster decisions. **SaaS analytics architecture** Data sources: | Source | Data | Key metrics | Tool examples | |---|---|---|---| | Billing | Revenue, subscriptions | MRR, ARR, churn | Stripe, Chargebee, Recurly | | CRM | Pipeline, deals, contacts | Pipeline, conversion, ACV | Salesforce, HubSpot | | Product | Usage, features, events | DAU, feature adoption, engagement | Mixpanel, Amplitude, Heap | | Support | Tickets, satisfaction | CSAT, response time, volume | Zendesk, Intercom | | Marketing | Campaigns, attribution | CAC, lead quality, conversion | Google Analytics, HubSpot | | Finance | P&L, cash, expenses | Burn, runway, gross margin | Xero, QuickBooks, NetSuite | Data warehouse: Why centralise: - Single source of truth (no conflicting numbers) - Cross-functional analysis (combine product + billing data) - Historical analysis (track trends over time) - Self-serve reporting (team can build own queries) Options by stage: | Stage | Tool | Cost | Best for | |---|---|---|---| | Seed | Google Sheets | Free | <£1M ARR, simple needs | | Series A | BigQuery | £100-500/mo | Growing data needs | | Series B | Snowflake | £500-2K/mo | Complex analytics | | Series C+ | Redshift/Databricks | £2K+/mo | Enterprise scale | ETL (Extract, Transform, Load): Tools to connect data sources to warehouse: - Fivetran: £500-2K/month (premium, reliable) - Stitch: £100-500/month (mid-range) - Airbyte: Free/open-source (requires engineering) Example pipeline: Stripe → Fivetran → BigQuery → Metabase (dashboard) BI and dashboarding: | Tool | Cost | Complexity | Best for | |---|---|---|---| | Metabase | Free (self-hosted) | Low | Startups, simple dashboards | | Looker (Google) | £3K+/mo | High | Complex analytics, governed | | Tableau | £50-70/user/mo | Medium | Visual analytics | | Mode | £500+/mo | Medium | SQL-heavy teams | | ChartMogul | £100-500/mo | Low | SaaS-specific metrics | Recommendation by stage: - Seed: Google Sheets + ChartMogul - Series A: ChartMogul + Metabase - Series B: BigQuery + Looker or Metabase - Series C+: Snowflake + Looker **Executive dashboard design** The 7 metrics every SaaS CEO needs: 1. ARR and growth rate - Chart: Line chart, 12-month trend - Target: Stage-appropriate growth rate - Red flag: Growth decelerating faster than expected 2. Net new ARR (with waterfall) - Chart: Waterfall (new + expansion - churn - contraction) - Target: Monthly net new ARR target - Red flag: Churn exceeding new business 3. Monthly burn and runway - Chart: Bar chart (burn) + line (runway months) - Target: 12+ months runway - Red flag: Below 6 months 4. Net revenue retention - Chart: Line chart, trailing 12-month - Target: >110% (mid-market), >120% (enterprise) - Red flag: Below 100% (shrinking customer base) 5. Gross margin - Chart: Line chart, monthly - Target: >75% - Red flag: Declining trend 6. Pipeline coverage - Chart: Bar chart (weighted pipeline vs quota) - Target: 3-4x coverage - Red flag: Below 2.5x 7. Cash balance - Chart: Line chart, with forecast - Target: Above minimum threshold - Red flag: Declining faster than plan Dashboard rules: - Update weekly (automated if possible) - Show trend (not just current value) - Include target/benchmark - Traffic light status (green/amber/red) - Click-through to detail **Departmental dashboards** Sales dashboard: | Metric | Frequency | Source | |---|---|---| | Pipeline by stage | Daily | CRM | | Weighted pipeline vs quota | Weekly | CRM | | New deals created | Daily | CRM | | Deals closed (won/lost) | Daily | CRM | | Average deal size | Weekly | CRM | | Sales cycle length | Monthly | CRM | | Win rate by stage | Monthly | CRM | | Rep quota attainment | Weekly | CRM | | Activity metrics (calls, demos) | Daily | CRM | Engineering dashboard: | Metric | Frequency | Source | |---|---|---| | Sprint velocity | Bi-weekly | Jira/Linear | | Deploy frequency | Daily | CI/CD | | Uptime / availability | Real-time | Monitoring | | Incident count and severity | Daily | PagerDuty | | Bug backlog | Weekly | Jira/Linear | | Feature ship rate | Monthly | Jira/Linear | | P95 response time | Real-time | APM | Customer success dashboard: | Metric | Frequency | Source | |---|---|---| | Health scores distribution | Daily | CS platform | | At-risk accounts | Daily | CS platform | | NPS/CSAT scores | Monthly | Survey tool | | Onboarding completion rate | Weekly | Product | | Expansion pipeline | Weekly | CRM | | Churn forecast | Monthly | CS platform | | Support ticket volume | Daily | Zendesk | | Time to first value | Monthly | Product | Marketing dashboard: | Metric | Frequency | Source | |---|---|---| | Website visitors | Daily | Google Analytics | | Lead volume (MQL, SQL) | Daily | CRM/HubSpot | | Conversion rates (visitor→lead→MQL→SQL) | Weekly | CRM | | CAC by channel | Monthly | Finance + CRM | | Content engagement | Weekly | CMS | | Email metrics (open, click) | Weekly | Email platform | | Paid ad performance (CPC, CAC) | Daily | Ad platforms | **Leading indicator framework** Leading indicators for key outcomes: Revenue growth: - Leading: Pipeline coverage, lead volume, conversion rate - Lagging: ARR, MRR, new ARR Churn: - Leading: Health score changes, login decline, NPS drop, support spikes - Lagging: Churn rate, churned ARR Profitability: - Leading: Gross margin trend, headcount-to-revenue ratio, cloud cost per customer - Lagging: EBITDA, operating margin Cash: - Leading: DSO trend, burn rate trend, pipeline coverage - Lagging: Cash balance, runway How to act on leading indicators: Example: Product engagement declining Signal: Average weekly logins dropped from 4.2 to 3.1 over 3 months Investigation: - Which customer segment? (Enterprise stable, SMB declining) - Which features? (Core features stable, new feature low adoption) - Correlation with churn? (Customers with <2 logins/week churn at 3x rate) Action: - SMB onboarding improvement (drive feature adoption) - In-app engagement campaign (drive login frequency) - CSM outreach to declining accounts Expected impact: - Prevent 5% of at-risk SMB churn - Estimated ARR saved: £50K **Data governance** Metric definitions: Create and maintain a data dictionary: | Metric | Definition | Calculation | Owner | Source | |---|---|---|---|---| | ARR | Annual recurring subscription revenue | Sum of active subscription values × 12 | Finance | Stripe | | MRR | Monthly recurring revenue | ARR ÷ 12 | Finance | Stripe | | New ARR | ARR from new customers | Sum of first subscription values (annualised) | Sales | CRM + Stripe | | Churn rate | Monthly customer churn | Customers cancelled ÷ starting customers | CS | Stripe | Data quality rules: - Single source of truth per metric (no conflicting reports) - Automated data pipelines (reduce manual error) - Regular audits (monthly spot-check) - Access controls (who can modify base data) - Documentation (how each metric is calculated) Common data pitfalls: Pitfall 1: Multiple ARR numbers - CRM shows £1.2M, billing shows £1.15M, spreadsheet shows £1.25M - Fix: Define one source of truth (usually billing system) Pitfall 2: Inconsistent time periods - Marketing reports monthly, sales reports quarterly - Fix: Standardise on monthly with quarterly roll-ups Pitfall 3: Vanity metrics - Tracking metrics that don't drive decisions - Fix: For each metric, ask "What decision does this inform?" - If no answer, remove from dashboard