Hoshin Kanri

Cascade map

Cascade map is licensed CC BY 4.0. Attribution: Katafacts (katafacts.com).

1 · What it is

What it is

A cascade map records the one hand-off Hoshin Kanri depends on and nothing else in this catalogue tracks: whether a parent plan's annual objectives actually became real, owned work at the level below. An X-Matrix sets one level's plan; catchball is the negotiation that produces it; a cascade map is the separate, later question of what happened to each objective after that — did a real unit take it on, and is that commitment real or just a name on a slide. It is not a KPI tree: a KPI tree breaks one goal into the driver metrics that move it, inside a single function; a cascade map tracks which unit owns which objective, across multiple units, and says nothing about the metrics that prove any of them are moving.

2 · When to use it

When to use it — and when not to

Use it when

You have a real X-Matrix (or a prior year's cascade) whose annual objectives need to become owned work at the level below it.
You want an honest record of which objectives have a real, committed owner and which are still just proposed — not a plan that looks complete because every cell got filled in.
You're checking whether the load is actually balanced across units, not quietly concentrated on the same team that always says yes.

Not when

You don't have a parent plan yet — build the X-Matrix first; a cascade map has nothing to cascade from without one.
You want a single function's own goal-to-metric breakdown — that's a KPI tree, not a cascade map. A cascade map is about which unit owns which objective; a KPI tree is what that unit tracks once it does.
You're tempted to mark a link 'committed' just to make coverage look complete — an objective assigned on paper with no real agreement behind it is exactly the failure catchball itself exists to prevent, one level up.

3 · How to fill it in

How to fill it in

Parent plan objectives: Name the parent X-Matrix and describe its real annual objectives — a first-pass list gets drafted from that, for you to correct against the plan's actual objectives and targets.
Cascade units: The teams, divisions, or sites this plan needs to cascade to — drafted from your description.
Cascade links: Only enter a link where a unit has actually proposed or committed to carrying a parent objective forward as its own, with a local objective that's genuinely more specific than the parent one — a real hand-off, never the parent text copied down.
Coverage: Computed for you from your parent objectives, units, and links — nothing to enter here.
Narrative: Drafted from the cascade actually built.
Next steps: What will you do about a dark objective or an unloaded unit, ranked by impact and effort, owned and dated.

4 · What good looks like

What good looks like

The example below cascades Northwind Fabrication's own FY2027 X-Matrix (the Hoshin Kanri worked example) to its shop-floor teams: two of the three annual objectives have a real, committed owner — the CNC Shop Floor and Scheduling teams — while the late-shipment objective is still only proposed to Quality, honestly left uncommitted rather than marked done to round the numbers up.

Same example, as a downloadable xlsx workbook.

Download .xlsx

Cascade map · Northwind Fabrication — FY2027 Hoshin Kanri

Northwind Fabrication — FY2027 cascade map

Oksana Petrova, Plant Manager · 2027-01-26

Team: Priya Nair, Ops Manager · Sponsor: Board of Directors

Parent plan objectives

Reduce average order lead time to 12 days
12-day average by Q4 FY2027
1 unit(s) committed
Cut the late-shipment rate from 14% to 6%
6% late-shipment rate by Q4 FY2027
Dark
0 unit(s) committed
Launch a repeat-customer priority queue
Live by Q2 FY2027; 20% of repeat orders routed through it by Q4
1 unit(s) committed

Cascade units

CNC Shop Floor
Team · 1 objective(s) owned
Scheduling Team
Team · 1 objective(s) owned
Quality Team
Team · 0 objective(s) owned

Cascade links

CNC Shop Floor — Complete second-shift hiring and training, and hit a 12-day average lead time on CNC-routed orders by Q4 (Marcus Webb, Shop Supervisor, committed)
Scheduling Team — Build and launch the repeat-customer priority queue in the scheduling system by Q2, with 20% of repeat orders routed through it by Q4 (Dana Ruiz, Scheduling Lead, committed)
Quality Team — Cut the late-shipment rate to 6% — proposed at the Q1 cascade review, owner not yet named (TBD, proposed)

Coverage

Parent objectives

Cascaded

Cascade coverage

66.7%

Dark objectives

Units

CNC Shop Floor

1 owned

Scheduling Team

1 owned

Quality Team

0 owned

Dark objectives

Cut the late-shipment rate from 14% to 6%

Narrative

Two of the plan's three annual objectives have a real, committed owner at the team level: the CNC Shop Floor has taken on the lead-time objective, and Scheduling has taken on the priority-queue launch. The late-shipment objective is still only proposed to Quality — raised at the Q1 cascade review, but nobody has actually agreed to own it yet. Cascade coverage sits at 66.7% on that basis, not 100%, because a proposal isn't the same thing as a commitment.

Next steps

Get a named owner and a committed local objective from Quality for the late-shipment target before the Q2 business review.
Cut the late-shipment rate from 14% to 6% — proposed to Quality, not yet committed · Impact: high · Effort: low · Oksana Petrova · 2027-02-15
Confirm the CNC Shop Floor's second-shift hiring is actually on track to support the 12-day target, not just committed on paper.
CNC Shop Floor — committed to the lead-time objective · Impact: medium · Effort: low · Priya Nair · 2027-02-28

5 · Common mistakes

Common mistakes

Marking a link 'committed' when it's really still just an assignment on paper.
A unit that was told to own an objective, but never actually agreed to it, hasn't really taken it on — the whole point of tracking 'proposed' separately is catching this before the year ends and nobody delivered.
Copying the parent objective's text down as the local objective instead of making it genuinely more specific.
A cascade that doesn't get more specific at each level isn't actually cascading — it's relabeling. The local objective should say exactly what this unit is going to do, not restate the parent's aspiration.
Building the cascade once at the start of the year and never updating it as proposals turn into commitments or quietly fall through.
Same living-document discipline as the X-Matrix itself — a cascade map that was accurate in January and never touched again is stale by the first business review.

6 · What it connects to

What it connects to

upstream

Hoshin Kanri — X-Matrix
The parent plan this cascades from — a cascade map has nothing to track without a real X-Matrix's annual objectives behind it.

downstream

KPI tree
Once a unit actually commits to a local objective here, a KPI tree is often what that unit builds next to track the driver metrics that move it week to week — the cascade map records the hand-off, the KPI tree is what happens after.
A3 problem solving
When a committed local objective isn't moving, an A3 is the right tool to find out why — the cascade map shows who owns it, the A3 closes the specific gap.

7 · Where AI helps

Where AI helps

Judgement — stays yours

Deciding whether a unit has actually committed to an objective or merely been assigned it
Deciding which dark objective or unloaded unit to chase down first

Analysis — AI helps

Drafting a first-pass parent-objective and unit list from a description of the parent plan
Drafting the narrative from the cascade actually built

Drudgery — automated

Computing cascade coverage and per-unit load
Flagging dark objectives — parent objectives with zero committed links
Exporting to xlsx in the house format

9 · Rate this kata