Real-Time and Predictive Reporting

Before opening any tool, document each metric with: the exact name, how it's calculated, which system it comes from, which API endpoint provides the data, the normal range (based on 12 months of historical data), the threshold that triggers an alert, who receives the alert, and what action they're expected to take. For most professional services firms, the core metrics fall into four categories: **Revenue metrics**: Monthly billed vs. target (daily tracking), pipeline value by stage, average engagement size, close rate on proposals. **Operations metrics**: Time-to-first-response on new leads (from Play Two), CRM field completeness (from Play One), invoice aging by bucket (from Play Six), new client onboarding cycle time. **Team metrics**: Billable hours by person vs. target, utilization rate, capacity for new work. **Client health metrics**: Key accounts with no logged activity in 45+ days, open service issues by age, accounts where sentiment has trended negative for 2+ consecutive interactions. Start with one metric from each category - four metrics total. Get those working before adding more.

Create a separate n8n workflow for each data category, running on appropriate schedules: - Hourly: new lead response time, new unread VIP messages (from Play Eight) - Every 4 hours: billing system invoice status, CRM activity count - Daily (6 AM): all other metrics - pipeline values, billing actuals, utilization rates, account health signals Each workflow queries the source system API, calculates the metric value, and writes it to a central data store. A Supabase database works well for most firms - it can store time-series metric data and serve it to your dashboard tool in real-time. A Google Sheet works as an alternative for firms not ready to manage a database. The data store structure: metric_name, value, timestamp, period (daily/weekly/monthly), source_system. Each workflow adds a new row for each metric on each run.

Example: Billing metric workflow (daily)

1. Schedule trigger: 6:00 AM daily
2. Query billing system API for all invoices issued in current month
3. Calculate: sum of paid invoices = billed_to_date, target = monthly_target field
4. Calculate: billed_to_date / monthly_target * 100 = percent_of_target
5. Calculate: (billed_to_date / day_of_month) * days_in_month = projected_month_end
6. Write to data store: {metric: "monthly_billing_pace", value: percent_of_target, projection: projected_month_end, timestamp: now}
7. Check: if percent_of_target < alert_threshold → trigger alert workflow

Connect your data store to your dashboard tool. Google Looker Studio is free and connects to Google Sheets, Supabase, and most major databases. Metabase is open-source and self-hosted. Grafana works well if you're comfortable with more technical setup. Build one view per metric category - not one view for everything. Cluttered dashboards get ignored. Each view shows: - Current value vs. target (large, prominent) - Trend for the last 30 days (line chart) - Alert threshold line (horizontal reference line) - Date of last update (to confirm data is current) Set the dashboard as the default view for leadership to check every morning. The goal is for checking the dashboard to take 2 minutes, not 20.

In each data collection workflow, after writing to the data store, add a threshold check. If the metric value crosses the defined threshold, trigger an alert workflow. The alert workflow: generates a brief context message (current value, threshold, recent trend, who owns the response), and delivers it via the configured channel (Slack message, email, SMS via Twilio). Alert message structure: what metric crossed, what it moved to, what the recent trend looks like, who's responsible for acting, and a direct link to the relevant dashboard view. Keep alerts to 3 - 5 sentences - if they're too long, people stop reading them. Route each alert to the person responsible for acting on that specific metric. Billing alerts go to the CFO or billing manager. Pipeline alerts go to the managing partner. Utilization alerts go to the practice leader. Routing everything to one person creates alert fatigue that causes the alerts to be ignored.

Generate a brief, actionable alert message for a performance metric that has crossed its threshold.

Metric: {{metric_name}}
Current value: {{current_value}}
Target/threshold: {{target_value}}
Direction of concern: {{direction}} (above or below threshold)
Recent trend (last 7 days): {{trend_description}}
Responsible owner: {{owner_name}}
Dashboard link: {{dashboard_url}}

Write a 3-5 sentence alert message that:
1. States what metric moved and by how much (specific numbers)
2. Provides brief context - has this been trending this direction, or is this a sudden change?
3. States what the owner should look at or do
4. Includes the dashboard link

Write in direct, business language. No filler. The owner receiving this is busy - they need to understand the situation and know what to do in under 30 seconds.