Design Effective KPI Dashboards

This skill designs effective KPI dashboards that drive business decisions. It defines a three-level KPI framework - Strategic (long-term goals, monthly/quarterly, executives), Tactical (department goals, weekly/monthly, managers), and Operational (day-to-day, real-time/daily, teams) - and the SMART KPI criteria (Specific, Measurable, Achievable, Relevant, Time-bound). It lays out a dashboard hierarchy of a one-page Executive Summary (4-6 headline KPIs, trend indicators, key alerts), Department Views (Sales, Marketing, Operations, Finance dashboards), and Detailed Drilldowns. It catalogs common KPIs by department: Sales (MRR, ARR, ARPU, pipeline value, win rate, deal size, sales cycle length, activity metrics), Marketing (CPA, CAC, lead volume, MQLs, traffic, conversion rate, engagement, ROI, CAC payback period), Product (DAU/MAU, session duration, feature adoption, NPS, CSAT, growth and retention rates, churn), and Finance (gross margin, net profit margin, EBITDA, operating margin, liquidity ratios, cash flow, revenue per employee, DSO, inventory turnover). It provides three dashboard layout patterns - an Executive Summary with headline metric cards and revenue trend/breakdown charts, a SaaS Metrics dashboard (MRR/ARR growth, unit economics with CAC/LTV/LTV-CAC ratio/payback period, cohort retention by month, gross/net/logo/expansion churn), and a Real-time Operations Center (system health, per-service status, request throughput, error rate, recent alerts) - plus illustrative Streamlit example code (metric cards, a Plotly revenue trend line and product-revenue pie chart, a cohort-retention heatmap, and an alerts section) that is a synthetic layout illustration, not a connected or filterable dashboard. Its KPI-calculation SQL guidance covers using consistent cohort denominators, deduplicating activity before counting, aggregating spend before joining customers, returning NULL for undefined growth or acquisition cost, and normalizing subscription intervals, currency, refunds, and revenue recognition before computing monthly revenue.

Use it for designing executive dashboards, selecting meaningful KPIs, building real-time monitoring displays, creating department-specific metrics views, improving existing dashboard layouts, or establishing metric governance. It does not apply outside KPI dashboard design - a different domain or tool is out of scope.

Inputs: the target audience (executive, department, or operations), the department's relevant KPI set, and update-frequency requirements. Outputs: a dashboard layout following one of the three patterns, a limited (5-7) set of chosen KPIs with context and trends, and illustrative Streamlit/Plotly dashboard code as a starting layout, not production-connected data.

Streamlit for the dashboard framework, Plotly (plotly.express, plotly.graph_objects) for line, pie, and heatmap charts, and pandas for the underlying data structures; also references SQL for KPI calculation against users, events, spend, and monthly_revenue tables.

def metric_card(label, value, delta, prefix="", suffix=""):
    delta_color = "green" if delta >= 0 else "red"
    delta_arrow = "▲" if delta >= 0 else "▼"
    st.metric(
        label=label,
        value=f"{prefix}{value:,.0f}{suffix}",
        delta=f"{delta_arrow} {abs(delta):.1f}%"
    )

Product, analytics, and BI teams designing executive, departmental, or operational KPI dashboards who need a proven metric framework and layout pattern rather than an ad hoc chart collection. Its best practices: limit a dashboard to 5-7 KPIs, show context like comparisons and targets, use consistent red-bad/green-good coloring, enable drilldown from summary to detail, and update at a frequency matching the metric - versus showing vanity metrics, overcrowding the layout, using 3D charts that distort perception, hiding the calculation methodology, or ignoring mobile responsiveness. References: Stephen Few's and Edward Tufte's dashboard-design writing, and the Google Data Studio Gallery.