cost management - lean frontiers...adjust (pdca) to strategic planning, deploy-ment, and execution...

12

Upload: others

Post on 01-Jun-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31 JANUARY/FEBRUARY 2019 COST MANAGEMENT

In a landmark article in this journal,noted accounting expert H. ThomasJohnson declared that “to becomelean” organizat ions must “shedaccount ing .” Johnson contrasted

two models of management control, man-agement by results (MBR) and manage-ment by means (MBM). 1 MBR is aNewtonian or mechanical model of exter-nal control, focused almost exclusively onfinancial results . 2 MBM is a biological lyinspired, cybernetic model of internal orsel f -control , focused on the means bywhich financial results are achieved. John-son seems hesitant to explain how exactlyMBM might be operationalized. Lean orga-nizations such as Toyota or the Danish firmScania, he says, simply do not permit exter-nal accounting controls inside their facto-r ies ; what i s required i s a new way of

th ink ing . Whi le I genera l ly agree w ithJohnson, MBM requires more than think-ing . As we w i l l see , even the th ink ingrequired is not necessarily free; it is regu-lated — by the scientific method. Like mostadvanced control systems, MBM requiresengineering.In accounting, we automatical ly think

of control in terms of budgets , interna laud i t s , and compl i ance . Accord ing toJohnson, however, the so-cal led internalaudits of management accounting constitutea primary example of MBR’s mechanicalor external control. To understand how toeng ineer a s ys tem of s e l f - cont rol , weobv iously need a new way to look at thecontrol problem that does not take auditsfor granted. In their ar t icle, “Who NeedsBudgets?” for the Har vard Business Review ,Jeremy Hope and Robin Fraser suggest that

DON’T SHEDACCOUNTING,REENGINEER IT —

WITH STRATEGYMANAGEMENT AND

A3 THINKINGTHOMAS L . JACKSON

T H O M A S L . J A C K S O N i s manag ing d i re c tor o f Mos s Adams LLP in S eat t l e , Wash ing ton , and a membe r o f t h eRona Con su l t ing Group, Mos s Adams’s l ean con su l t ing p ra c t i c e . Tom ha s b een t ea ch ing and con su l t ing in l ean s ince1988 and i s t h e author o f s e ve ra l book s on s t rateg y dep loyment and l ean hea l th care . P l ea s e addre s s inqui r i e s toTom. Jack son@MossAdams . com.

Forward-thinking organizations should consider ut i l iz ing a management by

means approach that incorporates strategy management to redef ine

accounting and give themselves a competit ive advantage.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

we think instead of contracts. Managementaccounting is characterized, they say, by ase t of f ixed per formance contrac t s , theresult of the tradit ional annual budgetingprocess. These are contracts offered to man-agers on a take-it-or-leave-it basis that t iepersonal compensation to compliance witha set of f inancial targets. Financial targetsare set with “a l itt le back and forth” at thetop of the organizat ion and remain fixedfor the fiscal year. In contrast, MBM is char-acterized by a set of relat ive improvementcontracts. Relative improvement contractstask managers with process improvementsas wel l as f inancial results; improvementsare related in a cause-and-ef fec t way tof inanc ia l outcomes , and compensat iondepends on progress, as measured by processimprovements as wel l as outcomes. A keyfeature of re l at ive improvement i s thatcontrol is not exercised external ly in per-formance rev iews or audits but internal lyin the course of dai ly work by managersthemselves. Managers are trusted to exerciseself-control (see the sidebar entitled “WhyShed Accounting?”). 3

In this article, I will explain how to oper-ationalize or engineer cybernetic manage-ment control using a relat ive improvementcontract known as the A3, which is the coredocument in a strategy management system.Strategy management is the applicat ion ofthe Deming cycle of plan, do, check, andadjust (PDCA) to strategic planning, deploy-

ment, and execution (see Exhibit 1). Firstdeveloped in Japan in the 1950s, strategymanagement is a system for ar t iculat ing,deploying, and enforcing relative improve-ment contracts companywide. It easily meetsstandard engineering criteria for cyberneticcont rol . S t r ateg y management gathers ,processes, and transmits all the informationof t radit iona l management account ing ,including its annual budgeting process. Itperforms the work of feedback and controlnormally done by performance reviews andinterna l aud i t s . It a l so encourage s thethinking — sometimes referred to as A3thinking — that Johnson demands. Exhibit2 presents the major differences betweenMBR and MBM.4 After a short discussionof the A3 process, this ar t icle explores thenature of the new management contract,communication, feedback, and managementcontrol.

Strategy management and the relativeimprovement contractIn contrast to the fixed performance contractof MBR, the relative improvement contractof MBM is not a set of marching ordersoffered on a take-it-or-leave-it basis; it isnegotiable. During negotiations, many prac-t it ioners of strategy management uti l ize acontrac t template known as an A3. (A3refers to the s ize of paper on which thecontract is printed.) The template explicitly

32 COST MANAGEMENT JANUARY/FEBRUARY 2019 STRATEGY MANAGEMENT

P

D

C

A

DO

CHECKADJUST

PLAN

Local decision makers integrate balanced scorecards of target

conditions of results and means into lean processes and agile

projects; the expectation is not to “do or die” but to design or

improve processes to increase customer satisfaction and

competitiveness.

Local decision makers self-enforce their A3 contracts by

monitoring lean processes and agile projects continuously and

by meeting frequently to discuss deviations from the

target condition results and means.

Based on evidence, local decision makers make adjustments (−/+) and escalate problems that require intervention; leaders quickly bring resources and expertise to bear on more complex problems with company-wide impact.

START HERE Leaders select and communicate financial expectations and the means of improvement to all managers of lean processes and agile projects, who then confirm by signing relative improvement contracts known as A3s.

relative improvement

contractA3

EXHIBIT 1 Strategy Management: A Framework for MBM

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

incorporates the scientific process that isused to encourage PDCA think ing (seeExhibit 3). 5

The A3 i s a l icense not on ly to th ink but to experiment. Because strategy man-agement is expl icit ly based on PDCA, ittreats results and means not as commands ,as in the case of MBR, but rather as hypothe-ses to be tested empirical ly through exper-imentation. Failed hypotheses are not causefor disincentive, discipl ine, or dismissal ,but are instead treated as oppor tunit iesfor organizat ional learning. In Exhibit 3,we see how the plan phase of PDCA is doc-umented in Sec t ions 1–4. In Sec t ion 1 ,problems are f irst factored in a problemcondition statement, so that managers may

pract ical ly contribute to problem-solv ingwithin their respect ive spans of control.In Section 2, the target condition statementor future s t ate i s descr ibed in terms ofprocess improvements and f inancial out-comes . S ec t ion 3 , the ana lys i s s ec t ion ,invest igates root causes around the gapbetween problem condition and target con-dit ion. The proposed act ions or counter-measures are outl ined in Sect ion 4 at thetop of the next column. In Sect ion 5, thedo phase of PDCA is outlined with a Ganttor milestone chart. The check and adjustphases in Section 6 refer to the continuousmonitor ing of improvement s and sub-sequent adjustments to the ac t ion plan.Inscr ibed convenient ly on one piece of

33STRATEGY MANAGEMENT JANUARY/FEBRUARY 2019 COST MANAGEMENT

LEGEND

CYBERNETIC ENGINEERING

COMPLIANCE WITH CRITERION

strong medium weak

� � rr

TYPE OF SYSTEM MMAMANMANAMANAGMANAGEMANAGEMMANAGEMEMANAGEMENMANAGEMENTMANAGEMENT BMANAGEMENT BYMANAGEMENT BY RMANAGEMENT BY REMANAGEMENT BY RESMANAGEMENT BY RESUMANAGEMENT BY RESULMANAGEMENT BY RESULTMANAGEMENT BY RESULTSS Mechanical / command-and-control

MMAMANMANAMANAGMANAGEMANAGEMMANAGEMEMANAGEMENMANAGEMENTMANAGEMENT BMANAGEMENT BYMANAGEMENT BY MMANAGEMENT BY MEMANAGEMENT BY MEAMANAGEMENT BY MEANMANAGEMENT BY MEANSMANAGEMENT BY MEANS Cyberne!c / biological / self-control

TYPE OF CONTRACT rr Fixed performance contracts � Smart, rela!ve improvement contracts

CCYCY

BCY

BECY

BER

CYBE

RNCY

BERN

ECY

BERN

ETCY

BERN

ETI

CYBE

RNET

ICCY

BERN

ETIC

CY

BERN

ETIC

CY

BERN

ETIC

CY

BERN

ETIC

E

CYBE

RNET

IC

ENCY

BERN

ETIC

EN

GCY

BERN

ETIC

EN

GI

CYBE

RNET

IC

ENG

INCY

BERN

ETIC

EN

GIN

ECY

BERN

ETIC

EN

GIN

EECY

BERN

ETIC

EN

GIN

EER

CYBE

RNET

IC

ENG

INEE

RICY

BERN

ETIC

EN

GIN

EERI

NCY

BERN

ETIC

EN

GIN

EERI

NG

CYBE

RNET

IC

ENG

INEE

RIN

G

CYBE

RNET

IC

ENG

INEE

RIN

G

CYBE

RNET

IC

ENG

INEE

RIN

G

CYBE

RNET

IC

ENG

INEE

RIN

G

FCY

BERN

ETIC

EN

GIN

EERI

NG

FU

CYBE

RNET

IC

ENG

INEE

RIN

G

FUN

CYBE

RNET

IC

ENG

INEE

RIN

G

FUN

CCY

BERN

ETIC

EN

GIN

EERI

NG

FU

NCT

CYBE

RNET

IC

ENG

INEE

RIN

G

FUN

CTI

CYBE

RNET

IC

ENG

INEE

RIN

G

FUN

CTIO

CYBE

RNET

IC

ENG

INEE

RIN

G

FUN

CTIO

NCY

BERN

ETIC

EN

GIN

EERI

NG

FU

NCT

ION

DEGREE OF DECENTRALIZATION rr

Normal decentraliza!on: parallel processing for divisionalized or moderately decentralized organiza!ons

� Radical decentraliza!on: massively parallel processing for radically decentralized lean/agile organiza!ons

CCOCOM

COM

MCO

MM

UCO

MM

UN

COM

MU

NI

COM

MU

NIC

COM

MU

NIC

ACO

MM

UN

ICAT

COM

MU

NIC

ATI

COM

MU

NIC

ATIO

COM

MU

NIC

ATIO

NCO

MM

UN

ICAT

ION

TERMS OF AGREEMENT rr

Contract to “get things done” (i.e., to achieve financial targets by any means available)

� Contract to think (i.e., to solve problems, using the scien!fic method [PDCA] to test hypotheses about how certain means deliver results)

EXTENT OF DEPLOYMENT �

Tradi!onal budge!ng: top-down communica!on to divisional leaders and some middle managers

� Companywide communica!on of targets for results and means to all managers, including supervisors and team leaders

FFEFEE

FEED

FEED

BFE

EDBA

FEED

BAC

FEED

BACK

FEED

BACK

NEGOTIATION PROCESS rr

Tradi!onal budge!ng: limited give-and-take at the top; take-it-or-leave-it throughout the rest of the organiza!on

Systema!c give-and-take following the improvement kata, with confirma!on memorialized in wri#en agreements (A3s) from top to bo#om of the organiza!on

FOCUS OF NEGOTIATIONS rr Financial results � Means as well as results: the balanced

scorecard

AAUAUT

AUTO

AUTO

MAU

TOM

AAU

TOM

ATAU

TOM

ATI

AUTO

MAT

ICAU

TOM

ATIC

CCOCO

NCO

NT

CON

TRCO

NTR

OCO

NTR

OL

CON

TRO

L

CONTRACT ENFORCEMENT rr

Mechanical control or post hoc compliance via quarterly reports and internal audits

� Automa!c control: daily, monthly, quarterly mee!ngs and visual control, checklists, and fail-safe devices

PERFORMANCE REVIEW rr

Annual performance reviews supported by incen!ves !ed to financial targets but not to means

� Frequent personal coaching supported by balanced paychecks of personal and team incen!ves !ed to means as well as financial results

EXHIBIT 2 An Engineering Perspective on MBR and MBM

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

paper, here is a record of the thinking thatJohnson cal ls for.While MBM and the A3 require thinking,

an MBM control system must be designed,engineered, and instal led to regulate thatthinking. In its essence, cybernetic controls ys tems a re a l l about the immediatedetection and correction — in other words,the automatic regulation — of error. NorbertWeiner describes three criteria for a cyber-netic system: communication, feedback, andautomatic or self-control.6 Applied to man-agement control, these three criteria canbe described as fol lows:• Communication consists of messaging— to al l decision-makers — of a targetcondit ion or strategic posit ion to beachieved (or maintained homeostat i-cal ly).• Feedback consists of information pro-v ided to decision-makers about per-formance dev iat ions from the targetcondit ion or strategic posit ion.• Automatic control consists of adjust-ments made by decision-makers, trig-gered by feedback, to bring themanagement system closer to the tar-get condit ion or strategic posit ion.The PDCA c yc le ( and the s c i ent i f i c

process itsel f ) is an obv ious example ofc yberne t ic regu lat ion . The pr ac t i ce of

strategy management gives us a rich patternfor engineering new, cybernet ic systemsfor management control, systems comprisedof re l at ive improvement cont rac t s thatempower managers to think as wel l as do.Please refer again to Exhibits 1 and 3. Inthe sect ions that fol low, we explore howst rateg y management sat i s f ie s Weiner’scriteria of communication, feedback, andautomatic control.

CommunicationThe first criterion of cybernetic control iscommunication. In normally decentralized(i.e., divisionalized) organizations like GM,it may have been enough to tel l the peoplewhat to do (through the budgeting process)and then to hold them accountable (throughperformance rev iews and internal audits).In the radically decentralized organizationof today, leaders must communicate whatto think about as well — and trust managersto regulate themselves. There are two dis-t inctive features of communication in thesystem of strategy management: systematicor companywide deployment and a balancedscorecard of results and means. Throughcompanywide deployment, a l l decis ion-makers are al igned in solv ing the organi-zat ion’s strategic problems.

34 COST MANAGEMENT JANUARY/FEBRUARY 2019 STRATEGY MANAGEMENT

1) PLAN: PROBLEM CONDITION

2) PLAN: TARGET CONDITION

3) PLAN: ANALYSIS

Experiments designedto reach the target condition

by using data to test the effectiveness of the

reliable means to achieve expected results

Experiments conducted close to the customer under the controlled conditions of

standardized processes, visual work instructions,

and fail-safe devices

4) PLAN: COUNTERMEASURES

5) DO: IMPLEMENTATION PLAN

6) CHECK / ADJUST

Problems factored into discrete, relatable stories and deployable computations

Future state described as a set of results/means conditions—a balanced scorecard of financial and nonfinancial targets

Data gathering and analytics; using data to determine root causes of the problem condition

1. ……….2. ……….3. ……….4. ……….

Results self-checked: tracked, validated, and

standardized at the front line; high potential

improvements deployed laterally to other areas of the

organization

EXHIBIT 3 Toyota’s A3: The New Relative Improvement Contract

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The f irst feature of communicat ion isthe systematic or companywide deploymentof results and means to middle managersand superv isors, as wel l as to div isionaland other senior leaders. In Exhibit 4, wesee a ful ly ar t iculated deployment tree fora medium-size energy company in the Euro-zone. Deployment has progressed from Tier1, the organizat ion, through its div isionsat Tier 2, to value streams and major projectsat Tier 3, and finally to processes, operations,and indiv idual tasks at Tiers 4 and 5. At

each t ier of management control, we see anA3 document that represents a formal agree-ment between managers at each node ofthis radical ly decentralized information-processing structure. In this case, roughly50 A3s were drafted, including A3s for sup-por t ing operat ions and specia l projects .A3s were then connected to specific qualityand cost improvements in processes at Tier4 and operations at Tier 5.13

The second feature of communicat ionis the so-called balanced scorecard of results

35STRATEGY MANAGEMENT JANUARY/FEBRUARY 2019 COST MANAGEMENT

WHY SHED ACCOUNTING?

Why are we talking about the need for such drastic changes in management accounting?

To put it succinctly, MBR does not support what Hope and Fraser refer to as radicaldecentralization. In any decentralized organization, “responsibility for strategic thinkingand decis ion making” has been shifted “from the center to people c loser to the

customer.”7 Decentralization is good because it increases the speed of organizational

decision-making — much the same way that parallel processing increases the speed

of modern computers — by 35 percent or more.8 But when does decentral ization

become radical? Radical compared to what? What is normal decentralization? Andwhy is radical decentralization such a problem that we cannot deal with it by resorting

to normal accounting systems?

Modern business was first decentralized in about 1919, when General Motors (GM)

was reorganized into the familiar divisions of Chevrolet, Buick, and Cadillac, among

others. Strategic thinking and decision-making was shifted from the chief executive’s

purview to divisional presidents.9 By the 1970s, GM’s divisionalized structure had

become a global norm.10 Thus, we will use GM — prior to its bankruptcy in 2009 — as

the benchmark for normal decentralization. To build on our computer analogy, in its

degree of decentralization, GM was comparable to an off-the-shelf laptop computer

today, which has roughly four to eight information processing cores. In organizations

such as Toyota, responsibil ity for strategic thinking and decision-making has been

shifted as close as possible to the customer. At Toyota, machine operators on the shop

floor are empowered to pause or even stop production when problems arise. This is

radical decentralization in action. Toyota has hundreds of thousands of cores. It is a

massively parallel supercomputer.

But speed comes at a price, and that price grows with the number of empowered

decision-makers. Just as parallel computer programming becomes progressively more

complicated as we multiply the number of cores, as we multiply the number of deci-

sion-makers, organizational leaders face similar “programming” problems — in other

words, problems of strategic thinking and decision-making. Such problems, known

collectively as the “agency problem,” arise because the natural human tendency is to

mind one’s own business. This leads managers to optimize their compensation rather

than the organization’s profits or share price.11 To deal with the agency problem of

normal decentralization, GM invented what eventually became MBR, including the

budgeting process, internal audits, and the annual performance review.12 For a while,

MBR quite successfully addressed the agency problem, allowing GM and its imitators

to dominate the g lobe. The number of empowered dec is ion-makers in rad ica l ly

decentralized lean and agile organizations has grown exponentially as a function of

the number of management tiers engaged in strategy management and the average

staffing ratio. The problem of agency and the need for new forms of control were

certain to reassert themselves.

. . . . . . . . . . .

36 COST MANAGEMENT JANUARY/FEBRUARY 2019 STRATEGY MANAGEMENT

TIER 1

TIER 2

TIER 3

TIER 5

TIER 4

MEA

NS

com

mun

icat

ed a

nd co

nfirm

ed

During the PLAN phase of strategy management: Each decision maker negotiates a balanced scorecard of financial results and the means of improvement.

Durin

g th

e CH

ECK

phas

e of

stra

tegy

dep

loym

ent:

Prob

lem

s tha

t can

not b

e re

solv

ed lo

cally

are

esc

alat

ed u

p !e

r-by

-!er

un!

l the

y ca

n be

reso

lved

in a

!m

ely

way

.C-suite

Division

Value stream/Department

Process/Project

Produc!on cell/Team member

RESULTS com

municated and confirm

ed

Tiers 1 and 2 represent normal decentralization typically found in large companies like GM and GE.

Decentralization beyond tiers 1

and 2 represents radical

decentralization found in

companies like Toyota, Scania,

and Handelsbanken.

NORMAL DECENTRALIZATION

RADICAL DECENTRALIZATION

EXHIBIT 4 Deployment Tree

degr

ee o

f aut

onom

atio

n

TIER LEADERSHIP LEVEL FREQUENCY MECHANISM

100% autonoma!on: human sense and response minimized but not eliminated

100% automa!on: human sense and response eliminated

1 C-suite annually

2 Division monthly

3 Value stream/ dailyDepartment

4 Process/ hour by hourProject

5 Production cell/ cycle by cycleSubteam/ minute by minuteTeam member task by task

real time

SENSE RESPONSE

Periodic reports +

assurance audits

Improvement kata

Daily reports,visual cues,

fail-safe devices

Automation,autonomation

+improvement

kata

EXHIBIT 5 Cycles and Mechanisms of Feedback and Control

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

and means . The Ba lanced Scorecard ofRobert Kaplan and David Norton is in factthe result of an encounter w ith strateg ymanagement, which they observed in thelate 1980s at Analog Dev ices. What theysaw was a method that ver y consciouslycombined the refined financial target ingderived from Peter Drucker’s managementby objectives (MBO) and the equally refinedtargeting of process improvement derivedf rom tota l qua l it y management (TQM).Taken together, MBO and TQM gave riseto the four focal areas of their BalancedScorecard, which include customer, process,growth (of people and technology), and,of course, finance. Each A3 in a deploymentof strategy management contains this infor-mation. In effect, each manager in the systemhas a ba lanced scorecard of resu lt s andmeans. In each A3, a balanced scorecardexpresses the strategic intent of the orga-nizat ion in increasing ly strat i f ied termsthat are under the span of control of theirrespective process owners.14

FeedbackThe second criterion of cybernetic controlis feedback, or information about deviationsfrom the targets for either results or means( s ee the s idebar ent i t l ed “Resu l t s S t i l lMat ter” ) . In a c yberne t ic managementcontrol system, al l decision-makers mustbe informed by feedback. Strategy man-agement prov ides mechanisms to processinternal and external feedback by using thevery same channels of communication estab-lished to deploy targets for both results andmeans (see Exhibit 4). These channels areused as a system for escalat ing dev iat ions,defects, gaps, and other problems that cannotbe solved readily within a manager’s spanof control. Within the system of strategymanagement, feedback has two dist inctivefeatures: the frequency of feedback cyclesand the confirmation of messages througha process known as catchbal l. 15

The first feature of strategy management’sfeedback system is the remarkable frequencyof feedback cycles. In mechanical manage-ment cont rol s ys tems , such a s MBR,managers must often wait for feedback untilthe annual performance review. In contrast,at each t ier in the radical ly decentralizedorganizat ion, in a process known as leader

standard work, managers physical ly v isitt he workpl ace to g ive fe edback on theprocesses under their respect ive spans ofcontrol ( see Exhibit 5) . At Tier 1 , chiefexecut ive s of ver y l a rge organ izat ions ,such a s Toyot a , may v i s i t e ach of the i rfaci l it ies once a year. 17At Tier 2, v ice pres-ident s v i s i t more f requent l y, p e rhapsmonthly. At Tier 3, value stream managersand ag i le team leaders are scheduled tov isit the product ion area week ly or evendai ly. At Tier 4, process owners who havemult iple operat ions under their controlmay revisit the work daily. At Tier 5, super-v isors rev isit the work ever y hour or so.Also relevant at Tier 5, the f inal level ofPDCA is the adherence to and continuousimprovement of standard work by frontlineworkers on the one hand and the adherenceto t he pro j e c t s chedu le by ag i l e t e ammembers on the other hand. Operatorsand team members natural ly rev isit their

37STRATEGY MANAGEMENT JANUARY/FEBRUARY 2019 COST MANAGEMENT

TRUST BUT VERIFY

In the world of public accounting, the

rise of the blockchain has elevated the

possibility that the financial or external

audit may soon be obsolete. In the radically

decentralized organization, the internal

audit may likewise be obsolete, at least

as a form of control. The internal audit

may nevertheless be useful as a form of

assurance.16

RESULTS STILL MATTER

Results stil l matter for Toyota. Not only

are managers required to give account

when they fal l short of their f inancial

targets, they are also required to explain

when they overshoot. Any deviation fromtargets fo r e i ther resu l ts o r means ,

whether positive or negative, is considered

to be a defect. It stands to reason that,

under MBM, managers are twice as likely

to be held accountable as under MBR.

Ironically, this illustrates Johnson’s point

that MBM requires thinking to replace

external, mechanical control. Toyota is

equally interested in the “how,” or causal

mechanism of target achievement, and

the “what,” or the results themselves.

Otherwise, it would not be able to remem-

ber how such results had been achieved.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . work w ith the complet ion of each workcycle. 18

The second feature of feedback is theconfirmation of messages concerning thenegotiation and review of results and means.Under MBR, targets are communicated ina top-down fashion with little or no feedbackfrom the recipients of this information. Incontrast , s t rateg y management requiresthat al l communications be two way. Theprocess of communication is a method ofnegotiation and dialogue commonly knownas catchball. Catchball follows a distinctivepattern that requires s igni f icant ly moregive and take than tradit ional budgeting.The pattern is known as the improvementkata. Kata, a term drawn from Asian martialarts, refers to a defensive or offensive patternof response, of ten drawn from nature. Inthe context of strategy management, thepat tern or k ata ensures messages aboutresults and their means are sent, received,

and confirmed in a way that puts organi-zat iona l l e a rn ing ahead of shor t - termfinancial results. In Exhibit 6, the kata isbroken down into open-ended coachingquestions posed by managers to their directrepor ts at each t ier of decision-making.The kata is used in both the plan phase ofPDCA, to negotiate relat ive improvementcontracts, and in the check phase, to inves-t igate reasons why the means agreed uponin initial negotiations have failed to producethe expected results. 19

Smart contracts andautomatic/autonomatic controlThe third criterion of cybernetic controlis automatic control, adjustment triggeredinstantly by detect ing dev iat ions from thetarget condition. Under the terms of MBR’sfixed performance contracts, the adjustmentprocess is mechanical and painful ly slow;

38 COST MANAGEMENT JANUARY/FEBRUARY 2019 STRATEGY MANAGEMENT

AA3A3** CCaCatCatcCatchCatchbCatchbaCatchbalCatchballCatchball qCatchball quCatchball queCatchball quesCatchball questCatchball questiCatchball questioCatchball questionCatchball questionsCatchball questions ccoonontontrontraontracontractontract nontract neontract negontract negoontract negotontract negotiontract negotiaontract negotiatontract negotiationtract negotiatioontract negotiationontract negotiation

RReRevReviRevieReviewReview qReview quReview queReview quesReview questReview questiReview questioReview questionReview questionsReview questions ccoconcontcontrcontracontraccontractcontract eenenfenfoenforenforccecemcemecemencementcement

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

2Se

ctio

n2 What is the proposed

target condi!on?

Define/review and analyze expected financial results and

other strategic objec!ves

What are we trying to achieve?

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

11 What are the suppor!ng facts?

Review and analyze historical data/informa!on and

empirical observa!ons of current condi!ons

What is happening right now?

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

33 What is your analysis of cause and effect?

Iden!fy significant challenges, devia!ons, defects, gaps, and

problems

What is ge$ng in your way right now?

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

44

What countermeasures do you propose to ensure good results?

Determine means/methods by which financial and other

performance targets may be achieved

What do you propose to do next?

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

55 What are the target achievement dates?

Establish/confirm the milestone dates by which means and results may be

achieved

When can we expect to see results?

SSeSec

Sect

Sect

iSe

ctio

Sect

ion

Sect

ion

55 What resources do you require?

Ensure sufficient human and material resources to execute

the ac!on plan

Do you require addi!onal resources?

*The catchball and review ques!ons that appear in this exhibit can be correlated to par!cular sec!ons of the A3. Please see Exhibit 3.

EXHIBIT 6 The Improvement Kata of Leader Standard Work

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

depending on the size and complexity ofan organizat ion, some processes may notbe audited more frequently than once everyfive years (see the sidebar entit led “Trustbut Veri f y”). Under the terms of MBM’srelative improvement contracts, each deci-sion-maker has the responsibility to checkfor or sense defects as they occur and theauthority, based upon scientific ev idence,to respond by adjusting the means withoutfurther management intervention. In thissense, control becomes automatic or self-control (more accurately termed “autono-matic” control in this instance — see thesidebar entitled “Automatic vs. AutonomaticControl”).As demonstrated in Exhibit 5, there are

several mechanisms involved in “autono-matic” control. At Tiers 4 and 5, the mostgranular levels of management control, visualand audial cues, checklists, and fail-safe or“mistake-proofing” devices are embeddedin processes and specific tasks and installeddirectly on machines where failures are likelyto occur. Flashing lights and musical tunesa ler t operators to defec t s . Automat icdetect ion can be coupled with automaticresponses to quarantine defects. Automaticdetection and correction of abnormal con-ditions can reduce the probability (even tozero) of defects occurring in the first place.Visual cues and fail-safe routines can alsobe installed in computer programs to helpproject managers, serv ice prov iders, andadministrators prevent and preclude defectsin knowledge work. When problems cannotbe resolved without assistance, productionis paused or stopped, and problems are esca-lated tier-by-tier up the chain of commanduntil the resources and expertise requiredto solve the problem can be organized. AtTiers 1–4 , control sys tems becomeincreasingly less mechanical or electrical— less automated and more automatic —but the improvement kata , which focuseson science and learning and eliminates blamefrom the process of investigation, can alwaysbe repeated.

ConclusionIn a quest for greater competitive advantage,first lean and now agile organizations haveadopted radical ly decentralized decision-making structures. Radical decentralization

promises faster organizational learning byeffect ively putt ing frontline managers andemployees in charge and asking them tothink. But legions of new decision-makersc reate new agenc y problems — that i s ,opportunities for self-dealing by managers,the agents of the organization, and its share-holders. As an effect ive system of manage-ment control, MBR — including the annualbudgeting process, the fixed performancecontract, and even the internal audit — isessential ly obsolete. Not only are the com-mand and the control of MBR slow to respec-tively detect and process feedback, but theyseverely l imit the organizat ion’s adaptiveresponses. In place of MBR, lean and agi leorganizations adopt cybernetic systems ofMBM. Those systems — at least those basedon strategy management — are designedto guide and suppor t as wel l as regulatemanagement self-control. MBM replacesthe machine of MBR’s f ixed performancecontracts with a l iv ing network of relat ive

39STRATEGY MANAGEMENT JANUARY/FEBRUARY 2019 COST MANAGEMENT

AUTOMAT IC VS. AUTONOMAT IC CONTROL

A3s are “smart” contracts, not unl ike

the smart contracts of the blockchain,

which can be programmed to transfer

funds instantaneously when the block-

chain senses that contractual conditions

have been met. Strictly speaking, A3s

are not automated, because they are not

connected mechanically or electronically

to the organizational chain of command.

Instead, we say they are autonomatedor “automated with a human touch.”

Autonomation is a work design philoso-

phy that embeds PDCA directly into man-

ufacturing and project management work-

flows by means of visual cues, checklists,

and electronic sensing devices and ser-

vomotors. When there is actionable feed-

back about a target condition, autonoma-

tion requires decision-makers to stop

and think. Technically speaking, automatic

control differs from autonomatic control

in that the former does not involve human

beings in the adjust cycle of PDCA. While

fu l l automat ion of check and act are

always an option, most lean companies

prefer to leave human beings in the PDCA

loop, unless the work is particularly dirty,

dangerous, or diff icult.20

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

improvement contracts. In the A3, we finda perfect template for those contracts. Thecompanyw ide deployment proce s s ofcatchbal l ensures that MBM’s communi-cation — the negotiation and confirmation— of balanced scorecards of financial targetsand process improvements to dec is ion-makers is clear. In improvement kata , wefind the means of feedback and autonomaticcontrol — that is, automation with a humantouch. That touch is the scientific method,PDCA, and a l icense to think inscribed onevery A3. So, while organizations may “shedaccounting” to become lean — or agi le —they should ser iously consider s t rateg ymanagement as an alternative. It prov idesa blueprint for the control system of thefuture, a system of cybernetic self-controland careful ly regulated, scientif ic think-ing. n

NOTES1 Johnson, H.T., Lean accounting: To become lean,

shed accounting, Cost Management 20, no. 1 (2006):6–17; Johnson, H.T. and Bröms, A., Profit BeyondMeasure: Extraordinary Results Through Attentionto Work and People. (New York: The Free Press,2000): 2.

2 Under the head ing “Externa l Cont ro l ,” Johnsonincludes the internal audit of modern managementaccounting , which is conducted by managementaccountants as external agents, in the sense thatthey are not part icipants in the processes that theyassess and control. Of course, the external, financialaudit of publ ic accounting is external in the samesense. In the shift toward lean and agile processes,however, the role of management accountants ischanging. See, for example, Cunningham, J., TheValue Add Accountant: An Indispensable PartnerSupporting Strategic Improvement Efforts. (Evanston,IL: JCC Press, 2018).

3 Hope, J. and Fraser, R., Who needs budgets? HarvardBusiness Review (Feb 2003): 2–8; Hope, J. andFraser, R., Beyond Budgeting: How Managers CanBreak F ree f rom the Annua l Per fo rmance T rap(Boston: Har vard Business School Press, 2003):26–29. Although Hope and Fraser’s research focusedon Europe and the United Kingdom, by my reckoning,the relat ive improvement contract f i rst emergedas a formal document in Japan in the late 1970s.See Shook, J. and Yoshino, I . , How the A3 came tobe Toyota’s go-to management process for knowledgework, The Lean Post (Aug 2, 2016). Avai lable at:https://lean.org/leanpost/Posting.cfm?LeanPostId=615.Japanese adopters of TQM and strategy managementmay have begun to go beyond budget ing in the1950s and 1960s, long before 1970, when Hopeand Fraser ’s paragon, Handelsbanken, abandonedthe budgeting process. See Akao, Y. (Ed.) , HoshinKanr i : Po l icy Deployment and Successfu l TQM.(New York: Productivity Press, 1991): 192. Economistsmaintain that the fundamental unit of analysis ofo rgan i z a t i ona l s t ruc tu re i s t he con t r ac t . SeeWill iamson, O.E., The Economic Institutions of Cap-i ta l ism: F i rms, Markets, Relat iona l Contract ing.(New York: The Free Press, 1985): 43–84. Natural lyenough, economists see organizat ions as a nexusor bund le o f con t rac ts . See Jensen , M.C . andMeckl ing , W.H. , Theor y of the f i rm: Manager ia lbehavior, agency costs and ownership structure,Journa l of F inancia l Economics 3, no. 4 (1976) :

305–360. These are contracts, essential ly, to solveproblems and achieve targets.

4 Shook, J., Toyota’s secret report, Sloan ManagementReview 50, no. 4 (2009): 30–33; Shook, J., Managingto Learn: Using the A3 Management Process toSolve Problems, Gain Agreement, Mentor and Lead.(Cambridge, MA: Lean Enterprise Institute, 2008).Regarding the system of strategy management, referto Akao (op. cit. note 3 Akao). The Japanese characterfor hosh in represents a compass and refers todirection or strategy. The character for kanri standsfor management or control, in the sense of qualitycontrol. Hoshin kanri, or strategy management, wasinvented in Japan in the 1950s in response to theUnion of Japanese Scientists and Engineers’ (JUSE)inclusion of a “quality policy” criterion in the DemingPrize. The first published book that explained howthe A3 is used to deploy and execute strategy wasJackson, T.L., Hoshin Kanri for the Lean Enterprise:Developing Competitive Capabil it ies and ManagingProfit. (New York: Productivity Press, 2006). Seealso Jackson, T.L., Implementing a Lean ManagementSystem. (Portland, OR: Productivity Press, 1996).

5 Op. cit. note 4 Hoshin Kanri, Introduction and Chapter4 (“Design the Annual Hoshin”).

6 Weine r, N . , Cybe rne t i c s : O r Con t ro l andCommunication in the Animal and the Machine. 2nd

ed. (Cambridge, MA: The MIT Press, 1961): 6–12.7 Op. cit . note 3 Beyond Budgeting, p. 119.8 Wheatley, M., Goodbye, command and control, Leader

to Leader 5 (Summer 1997): 21–28. Toyota is famousfor collapsing production lead times by more than90 percent and development and administrative leadt imes by more than 50 percent , resu l ts I havereplicated myself many times as a lean consultantfor manufacturing, health care, and other industries.

9 Chandler Jr., A.D., Strategy and Structure: Chaptersin the History of American Industr ia l Enterprise.(Cambridge, MA: The MIT Press, 1962).

10Servan-Schreiber, J. -J. , The American Chal lenge.(New York: Atheneum, 1968).

11 The agency problem is compounded by inherentbiases and a lamentable tendency to hide informationfrom others. Op. cit . note 3 Wil l iamson, pp. 279–283.

12Johnson, H.T. and Kaplan, R.S., Relevance Lost:The R ise and Fa l l o f Management Account ing .(Boston, MA: Harvard Business School Press, 1987):93–123; cf. note 3 Beyond Budgeting , pp. 9–15.

13Techn ica l l y speak ing , on ly managers ( inc lud ingsupervisors and team leaders) write A3s. At Tier 5,strategy management rel ies on a different type ofdocumentat ion known as visual management, inwhich the same balanced scorecard of targets forimprovements as wel l as results are recorded andtracked. Op. cit . note 3 Hoshin Kanri , Chapters 5(“Al ign the Organ izat ion Through Catchba l l” ) , 6(“Create a Lean-Th ink ing Env i ronment”) , and 9(“Institutionalizing Hoshin Kanri Through StandardizedWork, Kaizen and Leadership Development”).

14Kap l an , R .S. and No r ton , D.P. , The Ba l ancedSco reca rd : T r ans l a t i ng S t r a tegy i n to Ac t i on .(Cambridge, MA: Harvard, 1996): vi i ; Kaplan, R.S.and Norton, D.P., The Balanced Scorecard—Measuresthat dr ive performance, Harvard Business Review(Jan–Feb 1992): 71–79; Kaplan, R.S., “Analog devices:The half- l ife metric,” Harvard Business School Case190-061, March 1990 (revised June 1993).

15Op. cit . note 3 Hoshin Kanri , Chapter 5.16Conce rn ing the fu tu re o f f i nanc i a l aud i t s , see

“Blockchain technology: A game-changer in account-i ng?” De lo i tt e (Ma rch 2016 ) . Ava i l ab l e a t :https://www2.deloitte.com/de/de/pages/innovation/topics/blockchain.html. For an i l lustrat ion of howToyota’s production system performs the functionof an internal audit more or less in real t ime, seeJackson, T.L., Decoding value at the source: TheABCs of lean production metrics, Cost Management32, no. 4 (2018): 31–42.

40 COST MANAGEMENT JANUARY/FEBRUARY 2019 STRATEGY MANAGEMENT

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17The annual visit of the top executive is known asthe pres ident ’s d iagnosis , a method that is thefoundation of most qual i ty and productivity pr izes.See Jackson, T.L., Corporate Diagnosis: Setting theG loba l S t anda rd f o r Exce l l ence . (New Yo rk :Productivity Press, 1996).

18The seminal contr ibution on leader standard workis Mann, D. , Creat ing a Lean Cu l ture : Too ls toSustain Lean Conversions . 3rd ed. (Boca Raton, FL:CRC Press, 2015). For a detailed exploration of dailymanagement at Tiers 4 and 5 and its relat ionshipto strategy management, see Galgano, A., Compa-nywide Quality Management. (New York: ProductivityPress, 1994): 271–320.

19Rother, M., Toyota Kata: Managing People for Improve-ment, Adaptiveness, and Superior Results. (NewYork: McGraw Hil l , 2009). The coaching questionsof the improvement kata are deeply rooted in strategymanagement. Ibid. Galgano, pp. 353–360.

20Shingo, S. (translated by Di l lon, A.P.) , A Study ofthe Toyota Production System from an Industr ia lEng ineer ing V iewpoint . (New York : Product iv i t yPress, 1989): 57–60. The principle method of imple-ment ing autonomat ion is poka-yoke or mistake-proofing. See Shingo, S. (translated by Dil lon, A.P.),Zero Qual i ty Control : Source Inspection and thePoka -Yoke System. (Por t l and , OR: Product iv i t yPress, 1986).

41STRATEGY MANAGEMENT JANUARY/FEBRUARY 2019 COST MANAGEMENT