What is in the template?
Two sheets. The first is the working matrix, with one row per initiative and two calculated columns you do not touch: the numeric priority and the quadrant in words.
The second is the scoring scale. It is the part most underestimated: without written criteria, two people score the same initiative with two different numbers and the matrix stops meaning anything.
How do you score impact?
| Score | Criterion |
|---|---|
| 5 | Frees more than 500 hours a year or unlocks direct revenue |
| 4 | Frees between 200 and 500 hours a year |
| 3 | Frees between 50 and 200 hours a year |
| 2 | A noticeable improvement that is hard to quantify |
| 1 | Internal convenience with no measurable effect |
If an initiative does not fit any row, the useful question is not how it scores: it is why it is on the list.
How do you score effort?
| Score | Criterion |
|---|---|
| 1 | Weeks, no integrations, data available and clean |
| 2 | Weeks, one simple integration |
| 3 | One to two months, several integrations |
| 4 | Two to three months, or the data has to be cleaned first |
| 5 | More than three months, or it depends on systems you do not control |
The component most often forgotten is data quality. An initiative that needs clean data you do not have hides a whole project inside it.
What do you do with each quadrant?
Do it now (high impact, low effort): start here even if it is the least eye-catching. A visible delivery in weeks buys patience for the long projects.
Plan it in phases (high impact, high effort): it justifies the big investment, but after you have a win you can point to.
If there is room (low impact, low effort): never before a high-impact one.
Drop it (low impact, high effort): these are the projects done “because we can.” Dropping them explicitly is what frees up capacity.
Who should fill it in?
A small group with representation from the business and from whoever will execute. Scoring from the business alone underestimates the effort; scoring from technology alone underestimates the impact.
The conversation between the two sides is the real value of the exercise: the matrix is only the record of what was agreed. We use exactly this structure in AI Discovery sessions.