---
title: KPI Dashboard Design
category: product
entity_type: skill
price: $15
canonical: https://forgehouse.ai/skills/kpi-dashboard-design/
lang: en
hreflang_alt: https://forgehouse.ai/tr/skiller/kpi-dashboard-design/
last_updated: 2026-06-20
---

# KPI Dashboard Design

> Design effective KPI dashboards with metrics selection, visualization best practices, and…

A complete methodology for designing KPI dashboards that drive decisions, not just display numbers. It covers metric selection, a strategic-to-operational hierarchy, visualization best practices, and ready layouts for executive, SaaS, and real-time operations views, backed by SQL and Streamlit implementation code.

## Use cases
- Designing an executive summary dashboard with headline KPIs and trends
- Building a SaaS metrics view with MRR, unit economics, and cohort retention
- Creating real-time operations dashboards with system health and service status
- Selecting the right 5-to-7 KPIs from dozens of candidate metrics
- Writing SQL for MRR, cohort retention, and customer acquisition cost
- Improving cluttered existing dashboards for clarity and mobile use

## Benefits
- Drive faster decisions by leading with the result, then the supporting detail
- Cut visual noise by focusing on the highest-impact metrics, not every number
- Give every metric meaning by showing it against targets, trends, and context
- Build dashboards people actually read with fast load times and clear hierarchy

## What’s included
- KPI hierarchy framework spanning strategic, tactical, and operational levels
- Department KPI catalogs for sales, marketing, product, and finance
- Layout patterns for executive summary, SaaS metrics, and real-time operations
- SQL templates for MRR, cohort retention, and customer acquisition cost
- Streamlit dashboard implementation code with KPI cards and charts
- Dos and don'ts covering vanity metrics, 3D charts, and mobile responsiveness

## Who it’s for
Data analysts, product managers, and operators who design business dashboards and want metrics that lead to action instead of clutter.

## How it runs
Out of dozens of available metrics, only 5 to 7 survive. Each one gets an audience, a target and a pyramid layer before any chart is drawn, and color always means position against the SLO.
1. Maps every candidate metric into the KPI framework first: strategic (executives, monthly), tactical (managers, weekly) or operational (teams, real-time). Audience and update frequency are decided here, not after the fact.
2. Applies Pareto selection: out of dozens of available metrics, 5 to 7 headline KPIs survive. Vanity metrics are rejected, and every surviving KPI gets an SLI to SLO to SLA chain so the number is judged against a target, never shown bare.
3. Lays out per the Minto pyramid: 4 to 6 headline cards with trend arrow and percent delta on top, trend charts and breakdowns as the second layer, drilldowns capped at three clicks deep.
4. Writes and documents the calculation layer: SQL for MRR with month-over-month growth, cohort retention heatmaps keyed by signup month, CAC from spend over acquired customers. Methodology is never hidden behind the widget.
5. Builds operational sections on USE and RED: system health shows utilization plus saturation plus errors for each resource (utilization alone says nothing), service rows show rate, error percent and p50/p95/p99 latency with explicit thresholds.
6. Wires alerting and color coding to the SLO targets: green, yellow and red are positions against the objective, an alerts section surfaces breaches, and each widget refreshes at the cadence its metric actually changes.

## FAQ
### We're on Tableau, not Streamlit. Is the implementation code wasted on us?
No. The core of the methodology is tool-agnostic: the KPI hierarchy, metric selection, and layout patterns apply identically in Tableau or Looker. The SQL templates for MRR, cohort retention, and acquisition cost run on whatever database sits underneath, and the Streamlit code is just one reference implementation.

### I have 40 candidate metrics. How do I get to a dashboard people actually read?
Through the hierarchy: metrics split into strategic, tactical, and operational levels, and each view commits to one level with 5 to 7 KPIs. Every metric is shown against targets, trends, and context, with the result leading and the detail following. The department catalogs help you cut vanity metrics early.

### Does it connect to my data sources and build the pipeline?
No. It's a design methodology with templates, not an ETL tool. The SQL runs on data you already have in a queryable store; moving the data there is separate engineering work.

## Price
$15, one-time, no subscription. VAT included.

Related guide: [AI for data analytics](https://forgehouse.ai/guides/ai-data-analytics/)