Start from the decision
🎯You'll be able to
Turn a request into the decisions a dashboard supports, choose a few metrics for them, and define each metric so everyone means the same thing.
🏠The everyday picture
A car's dashboard shows speed, fuel and warning lights, because those are what a driver acts on. It doesn't show the engine's oil pressure history, though the car measures it.
📘What it really is
-
Ask who reads it, what decision they make, and how often. "A sales dashboard" isn't a brief; "store managers deciding each morning which products to restock and push" is.
-
Choose a few metrics, usually three to five headline numbers, each tied to a decision.
- Define every metric in words on the page and in the warehouse docs (DT7 Lesson 3), with what's included, what's left out, and the unit. SabziSathi's definitions:
- Net revenue: money from delivered order lines; cancelled and refunded lines are left out.
- Delivered orders: orders with at least one delivered line.
-
Average order value: net revenue divided by delivered orders.
-
A metric tree links a goal to its parts: net revenue = delivered orders × average order value. A change in revenue can then be traced to one of its parts.
-
Give context: compared with what? Last week, the same day last week, a target. A number alone rarely tells anyone what to do.
📝A worked example
Input: compare the starter's and the solution's dashboards (python dashboard/build.py in each, then open dashboard/out/index.html).
Expected output: the starter says "SabziSathi is growing fast!" and shows "Revenue ₹2,366,966" and "3,476 orders" with no definitions. That figure includes cancelled and refunded orders, repeated rows and staff test orders. The solution shows net revenue ₹19,65,764, 2,918 delivered orders and an average of ₹673.67, each defined under "What the numbers mean".
🌍Where it's used
- Retail: a morning dashboard for store managers with three or four numbers.
- Healthcare: bed occupancy, waiting time and discharges for hospital managers.
- Public services: scheme coverage by district, with definitions agreed between departments.
📝Case scenario
Two departments of a state government each reported "beneficiaries reached" for the same scheme, 4.1 lakh and 3.2 lakh. One counted payments, the other counted people. A shared definition, printed on both dashboards, ended a month of meetings about which figure was right.
🖱️Try it
Write a one-paragraph brief for SabziSathi's store-manager dashboard: who, what decision, how often, and three metrics with definitions. Success: each metric is tied to a decision a manager makes.
⚠️Common mistakes
- Building before asking. Sign: a dashboard nobody opens after the first week. Fix: name the reader and the decision first.
- Undefined metrics. Sign: two teams quote different numbers for the same name. Fix: definitions on the page.
- Numbers with no comparison. Sign: readers ask "is that good?". Fix: show last period or a target.
🛠️For your project
Project 3 tests 2 and 3 check the headline numbers and their definitions.
✅Quick check
What should a dashboard brief say?
Show the answer
Who reads it, what decision they make, and how often.
What does a metric definition include?
Show the answer
What's counted, what's left out, and the unit.