Implementation Guide

Hoshin Kanri.

Hoshin Kanri, also called policy deployment, is a management system for turning a handful of breakthrough objectives into aligned targets, KPIs and initiatives at every level of a plant, then reviewing them monthly for a full year. It lives or dies on two numbers and one verb: three breakthrough objectives, not thirty, and catchball, not cascade by email. The X-Matrix is only the record of that conversation. If the conversation did not happen, the matrix is wallpaper.

What Hoshin Kanri is

Hoshin is Japanese for compass needle; kanri means management. The method grew out of the Japanese quality movement of the 1960s and travels in English as policy deployment or strategy deployment. Strip the vocabulary away and it is three linked activities: choose a very small number of breakthrough objectives, deploy them through every level of the organization so each level knows its measurable contribution, and review progress on a fixed monthly rhythm for the whole year.

Direction, deployment, review. Most manufacturers already have the first: a strategy deck exists somewhere. What is usually missing is the other two thirds. The strategy stays in the boardroom, the floor runs on last year's habits and this week's fires, and the two meet once a year at budget time. Hoshin Kanri is the machinery that connects them, and the connecting mechanism has a name: catchball, the structured negotiation in which each level throws its proposed objectives down, and the level below throws back what it can commit to, what it cannot, and what it would need. Targets go down, means and objections come back up, and the plan changes as a result.

Which is why Hoshin Kanri is not an annual planning event. If it happens once in January it is budgeting with extra steps. The plan is the cheapest part; the system is the monthly review that keeps three numbers alive in every leader's calendar for twelve months. And the famous X-Matrix, the one-page grid linking objectives, initiatives, KPIs and owners, is just the record of the catchball conversation. If the conversation did not happen, the matrix is wallpaper.

Where it sits in the transformation roadmap

On the TeamGuru deployment roadmap, Hoshin Kanri is the strategy deployment practice in the Build the Frame stage. It deliberately comes after the diagnosis, not before it: breakthroughs chosen from gut feel are usually wrong, and a plant that has walked its streams with value stream mapping and established trusted baseline KPIs picks its two or three breakthroughs from evidence. Downstream, the deployed targets need somewhere to land, which is why daily management follows directly.

When not to start it

Hoshin Kanri assumes a plant that can measure, choose and hold a rhythm. Three situations say the assumption does not hold yet:

There is no baseline. If nobody can state current OTD, scrap or downtime with a number the floor trusts, any breakthrough target is a guess. Deploy measurement first, then strategy.
Leadership will not say no to existing projects. Hoshin Kanri is a deselection method wearing a planning costume; if every current project survives contact with the plan, you will get the thirty-objective wallpaper.
Firefighting is so acute that today's problems consume every meeting. Strategy deployed into chaos evaporates by Tuesday. Build the daily management system first, so there is a stable layer to deploy into.

In the third case the right sequence is inverted on purpose: stand up lean daily management first, stabilize the day, and deploy strategy into a system that can carry it. Hoshin into firefighting produces a plan nobody has time to execute and a review nobody attends.

How to run the annual cycle

The cycle below is one year long, but only steps 1 to 4 are planning, and they should be timeboxed to 4 to 8 weeks total. Steps 5 to 7 are the part most implementations skip, and the part that makes the difference between a management system and a poster.

Step	Who is involved	Output
1. Establish the vision and breakthroughs	Plant manager and the leadership team, informed by the diagnosis	Two or three breakthrough objectives on a 3 to 5 year horizon, each with a number and a date. Not themes, not values: measurable end states the diagnosis says matter most.
2. Define this year's annual objectives	Leadership team, one working session plus one revision	For each breakthrough, the slice this year must deliver, stated as a measurable annual objective, plus a written list of projects that will be stopped or parked to fund it.
3. Catchball down through the levels	Each leader with their direct reports, level by level, in both directions	Negotiated department objectives, KPIs and initiatives. The people who will deliver the targets have challenged the numbers, proposed the means, and changed at least something.
4. Finalize the X-Matrix	Leadership team plus one deployment owner who curates the document	One X-Matrix per level, linking breakthroughs, annual objectives, initiatives and KPIs, with a single named owner in every row. The matrix records the agreement catchball produced.
5. Execute with monthly reviews	Objective owners and the leadership team, monthly, timeboxed	KPI actuals against target, a countermeasure with an owner and a date for every red, and decisions: initiatives unblocked, resourced, or stopped.
6. Adjust mid-year	Leadership team, one scheduled checkpoint around month six	Revised targets or initiatives where reality diverged from the plan, and a written note of what changed and why, so the annual reflection has facts to work with.
7. Annual reflection (hansei)	Everyone who owned an objective or initiative	An honest assessment of what the process itself got wrong: targets set blind, catchball skipped, reviews that drifted. The findings feed step 1 of the next cycle.

Decisions that shape the cycle

Five decisions determine whether the seven steps produce alignment or paperwork:

Hold the line at three breakthroughs. If a fourth feels essential, it replaces one; it does not join the list. Every objective added halves the attention each one gets.
Catchball through two levels as a minimum: plant leadership with department or value stream leaders, then departments with their supervisors wherever supervisors will own a KPI. Below that, daily management carries the targets, not catchball.
Timebox the whole planning cycle to 4 to 8 weeks. Catchball expands to fill whatever time you give it, and a plan finished in March is a nine-month plan sold as an annual one.
Write the not-do list and publish it with the plan. Naming the projects you are stopping or parking is what frees the capacity; a Hoshin plan that cancels nothing is an addition, not a deployment.
One named owner per objective, KPI and initiative. A department cannot own a target. A person can.

Worked cascade: one breakthrough, followed down

The numbers below are illustrative but internally consistent, for the same 450-person components manufacturer used across the transformation roadmap: 456 units per day across cutting, machining, welding and assembly, on-time delivery stuck at 79 percent. The diagnosis showed most delivery misses tracing back to machining availability and late purchased material. Here is one breakthrough followed all the way down, with the wording changing at every level:

Breakthrough objective, two years

Owner: plant manager

Raise on-time delivery from 79 to 95 percent within two years.

Outcome language. A customer would recognize this sentence, and no department could deliver it alone.

Annual objective, this year

Owner: operations manager

Reach 88 percent OTD by December by halving the delivery misses caused by machining availability and late purchased material.

Still an outcome, but it now names the two causes this year's work will attack. The causes came from the delivery-miss data, not from opinion.

Department objective, machining

Owner: machining manager

Raise machining schedule adherence from 81 to 93 percent by cutting unplanned downtime on the two bottleneck centers from 26 to 10 hours per week.

Process language. Nobody outside the plant would write this sentence, which is exactly the point: the wording gets more concrete at every level while the arithmetic still adds up to the breakthrough.

What machining actually tracks and runs

KPI 1: machining schedule adherence, weekly, from 81 to 93 percent by Q4
KPI 2: unplanned downtime on centers 4 and 7, weekly, from 26 to 10 hours by Q3
Initiative: planned-maintenance program on centers 4 and 7. Owner: maintenance lead. First 90 days scoped, reviewed monthly.

Notice what catchball did to this thread. The first draft handed machining a 97 percent adherence target. The machining team came back with their downtime and miss data showing that a third of their schedule misses started with late purchased material, which they do not control. The result: machining's target settled at 93 percent, purchasing took its own deployed objective for supplier on-time delivery, and the annual objective's wording was rewritten to name both causes. The plan got more honest because it was negotiated. That revision is not a weakness of the process. It is the process.

The general rule the thread illustrates: each level's objective answers "how" for the level above it and "why" for the level below it. The top of the cascade speaks the customer's language; the bottom speaks the language of machine centers and hours. When every level of a cascade reads the same, no deployment happened, only distribution.

The review rhythm

The monthly Hoshin review is where the system earns its keep, and it is a different animal from a status meeting. Three rules keep it honest. First, the material is the working record: the X-Matrix and the live KPI charts, the same numbers the floor sees, not slides prepared the night before. Second, every KPI that misses its target leaves the room with a countermeasure, an owner and a date; explaining a gap is not the same as closing one. Third, the review decides. Initiatives get unblocked, resourced or stopped in the meeting, by the people in the meeting. A review that ends by scheduling another review has failed at its one job.

Ninety minutes is enough for a plant-level review of three breakthroughs if owners bring their numbers current, and it stays enough only if the objective count stays small. This monthly rhythm sits on top of the plant's existing management review structure rather than beside it: fold the Hoshin objectives into the same monthly meeting instead of creating a second, competing forum, and let the daily and weekly tier meetings handle the short-interval signals so the monthly review can stay at the level of objectives and countermeasures.

Failure modes

Hoshin Kanri fails in well-documented, repeatable ways, and almost all of them are visible within the first two months if you know what to look for:

What bad looks like

Thirty objectives with sub-objectives, none of which anyone can quote from memory
Targets without owners, or owned by a department instead of a person
Catchball theater: the meetings happened, but every number came back exactly as it went down
The X-Matrix lives in one CI engineer's spreadsheet and nobody else can explain a row of it
Monthly reviews that admire the charts, reschedule the hard items, and decide nothing

What good looks like

Three breakthroughs any supervisor can quote without looking
Every objective, KPI and initiative carries one named owner
Catchball changed at least one target or one initiative at every level
Owners maintain their own rows; the matrix is read in reviews, never rebuilt for them
Reviews end in decisions: countermeasures assigned, initiatives stopped or unblocked

What happens next

Two threads continue from here. The first is the working record: the X-Matrix guide covers how to build and read the one-page record itself, and the catchball guide goes deep on running the negotiation, including what to do when a level pushes back hard. The second thread runs downward: the deployed KPIs have to connect to the boards the floor runs every day, which is the territory of the KPI tree and lean daily management. A breakthrough that never appears on a tier board is a breakthrough the plant is not actually working on.

The practical failure point in both threads is the same: the links between objectives, KPIs, initiatives and owners decay the moment they live in disconnected spreadsheets. This is where TeamGuru's strategy deployment use case does its work, keeping every objective connected to its KPIs, its initiatives and its named owners from the boardroom down to the weekly actions, so the monthly review reads live data instead of a reconstruction. On the roadmap, the next practice is daily management: the system that turns this year's three numbers into every morning's conversation.

Hoshin Kanri implementation diagram (TeamGuru guide) — Take this with you: free to reuse in internal training and workshops. Download PNG

Frequently asked questions

What does Hoshin Kanri mean?

Hoshin is Japanese for compass needle or direction, and kanri means management or control, so the phrase translates roughly as direction management. In English it is most often called policy deployment or strategy deployment. All three names describe the same thing: a system for deploying a few breakthrough objectives through every level of an organization and reviewing them on a fixed rhythm.

What is the difference between Hoshin Kanri and OKRs or MBO?

They share DNA: few objectives, measurable results, regular check-ins. OKRs typically run quarterly and stop at goal setting, leaving the review system to whoever implements them, while classic MBO cascades targets downward without negotiating the means and was historically tied to individual appraisal. Hoshin adds three things the others usually lack: catchball negotiation of both targets and means, an explicit monthly review and mid-year adjustment cycle, and multi-year breakthroughs above the annual goals. A plant already using OKR vocabulary can keep it and add the catchball and review discipline.

How many objectives should a Hoshin plan have?

Three breakthrough objectives, five at the absolute maximum, and a similar number of annual objectives beneath them. The count is the whole method: every objective you add dilutes the attention, capacity and review time available for the rest. If leadership cannot cut the list to three, that disagreement about priorities is the real problem, and Hoshin has just surfaced it early.

How long does Hoshin deployment take?

The planning cycle, from breakthrough selection through catchball to a finalized X-Matrix, should be timeboxed to 4 to 8 weeks. The full cycle is a year: monthly reviews, a mid-year adjustment, and an annual reflection that feeds the next round. Expect the first year to be the roughest, because the baseline data, the review discipline and the habit of saying no are all being built at the same time.

What level should catchball reach?

As deep as KPI ownership goes, and no deeper. In most plants that means two rounds: plant leadership with department or value stream leaders, and departments with supervisors where supervisors own a deployed KPI. Operators meet the strategy through the daily management system, boards and tier meetings, not through catchball sessions.

What happens mid-year when reality changes?

Small deviations are handled by the monthly reviews: a red KPI gets a countermeasure with an owner and a date, not a revised target. Genuine shifts, a lost customer, a demand collapse, a failed initiative, are handled at a scheduled mid-year adjustment where targets and initiatives are formally revised and the change is written down. What kills credibility is the third path: silently abandoning the plan while the matrix stays on the wall.