Lesson 15 of 15 · Lab
How to write a decision memo from a dashboard
A decision memo states what to do, the evidence for it, what the data doesn't show and what would change your mind. It's how a BI project ends, because a dashboard shows numbers and leaves the decision to the reader. Before you write one, try to break the conclusion.
| Weak | Tested |
|---|---|
| Sales are up 8 points | Sales are up 8 points, and all of it came from one customer |
It's Friday at fictional Vandelay Coffee Supply, and founder Max Thrust wants one sentence on where the hire goes, North or South. After this course's fixes, Q3 net value is €12,098, or 100.8% of target. Your page has one filter.
South leads. Take out each region's biggest customer in turn. Does the lead look sturdier than the headline, or weaker?
Weaker on both sides. South without Initech is €3,520, behind North's €5,578. North without Starhopper is €1,690: Starhopper is 69.7% of North. Each region leans on one customer, but South survives the loss better, and that's the case for it.
Test the conclusion both ways
- Read this live page in Vista, Oplema's browser app for dashboards: total, target, regions, customers.
- Select Initech. Read South.
- Select Initech again, then Starhopper. Read North.
- Select Everyone. Now write your memo.
Write the memo
Five lines: the decision, the evidence, the limits, the review trigger, the open question. Write yours before reading the model.
Compare with a model memo
Decision. Place the hire in South, which has two accounts of €3,000 or more against North's one, and review at Q4 close.
Evidence. Q3 net value is €12,098, and reaching target (100.8%) depends on Sydney's ruling on part-paid. South €6,520, North €5,578, South by €942. South has led only since September 23.
Limits. South without Initech is €3,520, behind North, and Initech signed only in September. It ordered again on October 12 (INV-1017, outside Q3). Starhopper is 69.7% of North, so North is concentrated too. Undated pending INV-1016 is out, and regions follow the current register.
Review trigger. If Initech stops ordering by Q4 close, the case reverses.
Open. Excalibur has no Q3 invoices, and nobody yet knows why.
Compare this with the report you chose in the first lesson. Don Tremendous's report named the right region for the wrong reasons, and the closer report named the wrong one. The numbers narrowed the decision. They don't prove the hire will work, and the memo says so. Max reads the five lines, approves South and asks who's calling Excalibur.
Common memo mistakes
- Hiding the limits. A reader who finds one later stops trusting the rest.
- No review trigger. Without one, a decision stays in place long after its evidence has gone.
- Claiming cause. These invoices show who bought, and not why, or what a new hire would add. They say nothing about profit or cash collected either.
- Leaving the decision to the reader. A page of findings with no recommendation hands back the work.
Write a memo when someone has to commit money or people, and skip it for a routine weekly number. Any word processor will do, and the template below fits on a page.
Template: decision memo
| Line | Your answer |
|---|---|
| Decision: what happens, and when you review it | |
| Evidence: the figures it rests on | |
| Limits: what the data does not show | |
| Review trigger: what would change it | |
| Open: what nobody knows yet |
Try your own file
Course complete Now use your own file.
A dashboard and a one-page decision memo for fictional Vandelay Coffee Supply's hire, with contract and memo templates for your own work.