X-Matrix orientation: vision on top, or at the bottom?
You will meet two drawings of the X-Matrix. In one, the 3-5 year vision sits on top, annual objectives on the right, initiatives at the bottom and KPIs on the left. In the other, the vision sits at the bottom, annual objectives on the left, initiatives on top and KPIs on the right. They are the same matrix rotated half a turn: same four roles, same order, same links.
What matters is the role of each quadrant and the order you read them in, never the compass point. This page shows where each role sits in both layouts, how the corners follow, and how to move a matrix from one template to the other without losing a link.
Four roles, read in one order
Every X-Matrix holds the same four roles: breakthrough objectives (where the company must be in 3 to 5 years), the annual objectives that get it there, the initiatives that deliver the annual objectives, and the indicators that prove the initiatives work. Each role only links to its two neighbours, which is why the matrix is drawn as a ring around the X.
In both layouts the ring is read clockwise: vision, then annual objectives, then initiatives, then KPIs, back to the vision. Only the starting point changes: the top in one drawing, the bottom in the other.
- Vision on top: vision top, annual objectives right, initiatives bottom, KPIs left.
- Vision at the bottom: vision bottom, annual objectives left, initiatives top, KPIs right.
The corners follow the roles
The correlation marks sit in the corner between two neighbouring roles. Rotate the matrix and the corners rotate with it: the vision × annual objectives grid is the top-right corner when the vision is on top, the bottom-left corner when it is at the bottom. The marks inside do not change, only where the grid is printed.
That is the one trap when you copy a matrix from one template to another: copy the marks corner by corner and you pair the wrong roles. Copy them role by role (this objective serves that vision, this initiative moves that KPI) and nothing is lost.
Which one to choose
Neither is more correct. Pick the one your committee already reads, and keep it: a team that has learned to look for the KPIs on the right should find them on the right at every review. If you deploy through a cascade (group, business unit, site), keep the same orientation at every level, so a site manager reading the group matrix finds the same roles in the same places.
If you work with a consultant, ask which layout their templates use before the first workshop. Switching mid-cycle costs more attention than either layout is worth.
How Hoshin handles it
In Hoshin, a quadrant is stored as a role, not a position. The score, the cascade, the API and the MCP server reason on roles, so an AI agent reading or writing your matrix never needs to know how it is drawn. The Excel template carries its links in columns, one row per link, so a workbook filled by someone who thinks vision-at-the-bottom imports the same.
Hoshin draws the vision on top today. Drawing it at the bottom is a display setting we have not built yet: if your committee needs it, say so from the roadmap on the home page, it is one of the capabilities we build on demand.
Frequently asked questions
Is one X-Matrix orientation the original one?
Both appear in practice and they hold the same content. What the method fixes is the four roles, their order and the links between neighbours, not which edge of the page the vision sits on.
Which way do you read an X-Matrix?
Clockwise from the vision in both layouts: vision, annual objectives, initiatives, KPIs, and back to the vision. With the vision on top you start at the top; with the vision at the bottom you start at the bottom.
How do I convert a matrix from one layout to the other?
Move each element by role, not by position, then rebuild each correlation as a pair of roles (this annual objective serves that vision). Copying the corner grids as they are pairs the wrong quadrants, because the corners rotate with the matrix.
Does the orientation change the completeness score?
No. The score measures coverage, density, orphans, chains from vision to KPI and indicator balance, all on roles and links. The same matrix scores the same whichever way it is drawn.