Understanding Office Kaizen
- Raphael L. Vitalo, Ph.D. and Joseph P. Vitalo
Some people distinguish between “shop Kaizen” and “office
Kaizen.” They describe shop Kaizen as applying to manufacturing work processes
and office Kaizen as applying to all else. Kaizen has made powerful contributions
to improving the productivity of manufacturing businesses, but now we need to
achieve the same significant improvements in our administrative processes and
the fastest growing sector of our economy—the service sector. How do we
transfer the use of Kaizen from "shop" to "office" or service
settings? We think the distinction of shop versus office misses the mark. The
issue is not location or even industry; it is the nature of the work and the
characteristics of the context in which the work is done. Adjust for these factors
and you can realize the benefits of Kaizen with any work process by using the
in the Kaizen Desk Reference Standard.
Office Work Processes
Types of Tasks and Work Settings
Psychologists distinguish between instrumental and cognitive tasks. Shop or
manufacturing work processes are dominated by instrumental tasks while office
work processes are dominated by cognitive tasks. Instrumental tasks are largely
physical in nature and involve the manipulation of things (e.g., lifting, polishing,
drilling, driving, sculpting). Cognitive tasks are largely mental in nature
and involve the manipulation of data, information, or knowledge (e.g., planning,
evaluating, or designing; counting, calculating, or imagining; deciding1,
choosing, determining, identifying, detecting, judging, or selecting).
The qualifier "largely" is used in both definitions for two reasons. First,
no task is purely one or the other—wholly instrumental or wholly cognitive.
Second, the performer may infuse an instrumental task with considerable cognitive
components (see Exhibit 1 for an example).
Exhibit 1. The Difference a Performer
Hitting a baseball may seem a purely instrumental
task, and, based on interviews, that is how Mickey Mantle approached it.
On the other hand, Ted Williams approached the task very differently.
Both players were exemplary hitters. Mantle was an intuitive performer
who blended talent, practice, and strength to create skill and achievement.
He could not, by his own admission, explain how he hit the ball; he just
did. When he slumped, he would use trial and error to try different ways
to break out of it. Mostly, he just kept trying and let things work themselves
out. Williams was a knowledge-driven performer. He could explain in minute
detail how he executed a particular swing and what swing he would use
in a particular circumstance. If he slipped, he used his knowledge to
detect the source of his batting problem and to devise a fix. Here are
two performers executing the same instrumental task, yet one is infusing
it with a great deal of cognitive components.
Differences Between Instrumental and Cognitive
Notwithstanding the exceptions a performer's approach may create, instrumental
and cognitive tasks have important differences (summarized in Exhibit 2) and
these differences are expressed in the work processes they make up. The most
significant differences are:
- Cognitive tasks are not directly observable,
- Cognitive tasks produce different outputs,
- Work processes dominated by cognitive tasks frequently lack standards, metrics,
and feedback, and
- Performer's implementing work processes dominated by cognitive tasks expect
to be given more discretion in how they accomplish what they do.
Not Directly Observable
Cognitive tasks are not directly observable. By definition, cognitive
work occurs in the mind of the performer. It may be tracked by documents,
whether print or electronic, but mental activity is the substance of the
work and that is not directly observable. This mental activity receives
data or information and transfers or transforms it as a result of a series
of decisions which are followed by some activity that documents or communicates
their results. For example, a clerk receives an order for a quantity of
a specialty gas like neon. The clerk must decide to what plant to route
that order based on who the customer is, where the product is to be shipped,
how large the order is, which plants can produce the gas, and host of other
factors. When done, the clerk "routes" the order by physically
pressing some keys on a keyboard, but the substance of the clerk's work
was not the physical activity of pressing the keys; it was the mental activity
of deciding how to correctly route the order. As another example, a technical
representative for a software development company receives a call from a
frustrated customer who is having problems installing their new software
application. The technical representative takes in both the state of the
customer (e.g., energy, affect, attitude, knowledge of computers) and information
about the problem the customer is experiencing. The technical representative
works out a solution and tells it to the customer. The substance of this
customer service activity is not in the physical speech that shares the
answer (as important as that is), but in the mental activity that concludes
what the correct answer is and how to properly communicate with the customer.
In both examples, the activity and result are not immediately observable.
You must do something additional to make them observable. Usually, you document
the decision or changed information or you survey your service recipient
and document their judgments as to whether they were satisfied with the
service they received.
Produce Different Outputs
Work processes dominated by cognitive tasks always produce information and,
in service contexts, produce the changed state of a person. The service sector
is the fastest growing component of our economy and this is one reason for
the popularity of office Kaizen. Some examples of the changed state of a person
are: a customer satisfied with the response to his or her inquiry, a trainee
who can now perform a task he or she could not previously perform, a patient
comforted by the nursing care he or she received, a customer who can now use
your product to accomplish a new business function. In such service situations,
the initial state of the person is one of the inputs to the work process and
the final state is its key output.
Frequently Lack Standards, Metrics, and Feedback
In our experience, work processes that are predominately cognitive are far
less likely to have precise specifications for inputs or outputs, work standards
to govern their execution, or process and output monitoring that feeds back
information on the quantity and quality of performance.2
Anyone experienced in doing process observations has learned that you always
make sure you are observing the work process as it is supposed to be
implemented. Indeed, when you walk out on a shop floor and ask someone to
show you a work activity, it is not uncommon to have the question come back
to you"Which way, by the book or the way we do it?" Goal-oriented
workers try to find ways around processes that obstruct them. Nevertheless,
in shop settings, workers know that they are doing "a work around"
and that it is not condoned. Not infrequently, the organizations within which
office processes occur permit more performer discretion in how work is accomplished.
They may not precisely define how decisions are to be made—e.g., what
rating to assign a worker in his or her performance appraisal or when do you
decide if there is no solution to the customer's difficulty in installing
your software product. In these settings, workers tend to expect to be given
more discretion in how they accomplish what they do. Sometimes, like in professional
settings, they even experience it as their right or prerogative. Since frequently
in these settings the performer is the judge of his or her own work, you can
encounter a circularity that sends a service further and further awry from
its original intent.
Key Similarities and Differences Between Instrumental and Cognitive Tasks
and the Contexts Within Which Each Is Performed
- Predominantly physical materials
- More likely to be precisely specified as to content and characteristics
- Predominantly information which may be contained in a document or
acquired by conversation
- Includes the initial state of a customer in service processes
- Less likely to be precisely specified as to content and characteristics
- Predominantly physical products but includes some information output
- More likely to be specified precisely with respect to form, fit, finish,
- Problems with outputs more easily observable
- Predominantly information which may be in the form of a decision,
a document, or a report
- Includes the final state of the customer in service processes
- Less likely to be precisely specified
- Problems with outputs less easily detected as they tend to be judgments
not standardized by explicit rules
- Readily observable
- Predominately a sequences of actions
- Not easily observable
- Predominately a sequences of decisions which trigger actions
- Performance tracking and reporting is embedded in the work process
and part of normal management activity
- Output is evaluated against an explicit specification
- Performance feedback is frequently immediate and visible as the component
either passes or fails inspection and the product is produced at its
required volume and speed
- Performance tracking frequently tied to personnel appraisal system
which may not be formalized and, even when formalized, may not be implemented
- Output evaluation is typically a subjective judgment by a supervisor
and may even be the worker's judgment of his or her own work irrespective
of any objective measurements
- Feedback, when it occurs, is usually globalnot instance specific
and occurs at a few set dates during the year
- Detailed work standards more likely to exist
- Production system is more visible as it is marked by work stations
that have different tools or equipment arranged in a sequence that matches
the flow of materials through the work process
- Focus of performers is on output
- Detailed work standards less likely to exist
- Production system is less visible. Frequently a set of similar desks
or work spaces with few cues as to the nature of the work or the flow
- Focus of performers is more likely on process
Implications for Kaizen
The Kaizen process documented in the Kaizen Desk Reference Standard
requires only minor modification to accommodate the differences between cognitive
and instrumental tasks (see Exhibit 3, Guide: Adjusting the Kaizen Process to
Accommodate "Office" Work Processes at the end of this article). By and large,
the steps remain exactly the same.
There are three milestones we complete prior to any event. We document a scope
for the event, analyze whether the event makes sense, and prepare for doing
the event. Little changes in these activities.
We document the scope the same way and use the same form (Milestone A. Document
the Scope of the Kaizen Event in the Kaizen Desk Reference Standard). We prepare
for an office event in the same manner we do for a shop prevent (Milestone C.
Prepare for the Kaizen Event in the Kaizen Desk Reference Standard). Obviously,
we will not need certain equipment (e.g., hard hats, safety glasses), but our
checklist handles both settings.
While the characteristics of the typical office work setting do underscore
caution and care as you evaluate whether doing a Kaizen event on a work process
makes sense (Milestone B. Analyze Whether to Conduct the Kaizen Event in the
Kaizen Desk Reference Standard), you do not need to adjust the process itself.
As you pass through the decisions you need to make about the appropriateness
of the work process, the alignment of people, and the business case for the
event, you apply the same knowledge in making this evaluation for a "shop"
work process as for “office” work processes. We do find, however,
as we apply our evaluation process that the typical office or service work process
is less likely to be ready for a Kaizen event because its work process may not
be standardized, have no measurements of their performance, and its people may
not be aligned with respect to supporting such an event. When there is no documented
work process, you cannot do the event. You must first build a standard work
process and establish a baseline of experience in using it. Similarly, if people
are unaligned with respect to using a standard, objective feedback on the results
of their work, or changing how they perform their work, you need to address
these “cultural” issues, not do a Kaizen event.
During the Event Activities
Exhibit 4 depicts the process we follow during a Kaizen event.
4. Tasks Completed During the Kaizen Event
We begin each event by introducing the event and doing an ice breaker exercise
to get people familiar with each other. Next, we overview the Kaizen process,
set ground rules for working together, and take care of administrative issues.
Before building a detailed picture of the work process itself, the team completes
a warm-up exercise in which members share with each other what seems to work
well in their work process and what is problematic. This exercise gives us an
opportunity to start thinking about the work process. It uncovers both the team's
concerns and the concerns that team members have heard from other employees.
We build a list of pluses and minuses with respect to how the work process currently
operates, post this list, and used it as a reference during the event. None
of these start-up activities change.
Our first set of tasks focus the Kaizen event based on the facts directly uncovered
by the team. As with all events, we have a strawperson direction (mission, goals,
and a list of “do's and don'ts”), which we build before the event
from a scope document and through preliminary conversations with key stakeholders.
None of this changes. In focusing the event, there are minor adjustments in
three tasks—building a description of the target work process, doing the
walk through, and evaluating waste in the work process.
Building the Description of the Target Work Process
The nature of cognitive work results in some differences in the work process
map you produce (see Step D1-S1. Build a Description of the Target Work Process,
Kaizen Desk Reference Standard), but the process for building this map is
unchanged. The map will almost always be multi-departmental, depict more decisions,
and require additional documentation of the knowledge guiding those decisions3
(see Exhibit 5).
Exhibit 5. Example of a Typical “Shop”
and “Office” Work Process
Decisions in a cognitively dominated work process are the equivalent of the
critical material processing operations in a shop or manufacturing work process.
Decisions constitute the activities of analysis and interpretation and direct
the transfer or transformation of information or the execution of actions.
Decision-making is the equivalent of the skillful actions of machine operators
and craftspeople who shape with tools a raw or semi-finished material into
a final product. The “tool” used by the cognitive task performer
is knowledge, usually in the form of rules. For example, one task a personnel
officer performs is to process retirements. As part of that activity, he or
she needs to determine whether that worker qualifies for a pension. The officer
begins with information about the worker including his or her age and length
of service. The “tool” the officer uses to transform this information
into a pension determination is a rule—let’s say the rule of 80.
This rule specifies:
||the worker is at least 50 years of age
||the worker's length of service equals or is greater than 30 years
||the worker qualifies for a pension.
By applying this rule to the information the officer has, he or she fashions
a judgment which is the output of the officer's task.
Apart from the heavy presence of decisions in the work process map, be prepared
to document more information elements as information constitutes the major
inputs and outputs of tasks in a cognitive work process. Our process already
includes capturing this content as part of the descriptive information about
each task. Just recognize that there will be more to capture.
Doing the Walk Through
The nature of cognitive tasks also requires two adjustments in how you conduct
the walk through in your Kaizen event (Step D1-S2. Walk Through the Target
Work Process, Kaizen Desk Reference Standard). The first adjustment is to
your thinking as the event leader. The second adjustment is to how the walk
In preparing team members for the walk through, we teach them how to detect
waste. We have added to our Kaizen Tool Kit (Version 1.3), a new exercise
that uses only office examples of waste so that the practice team members
do is directly aligned with the process they will walk.
Adjust Your Thinking
You will need to remind yourself that the forms of waste you and the team
will observe in an office work process will usually involve working with
information and especially the paper forms or software screens that carry
information. Frequently, these forms or screens are poorly designed so that
the sequencing and display of information neither coincides with the flow
of work or enables speedy recognition of their contents, thereby causing
search. Sometimes forms contain contents that are unnecessary and obscure
the needed information, also causing search. Sharpen your understanding
of transport as well to include moving information from one paper form to
another or from a paper form to a computer screen. Consider completing incomplete
information or correcting incorrect information as rework. Recognize that
filling out forms or entering information into a computer screen is setup,
but that entering the same information into multiple software applications
because these applications do not “talk to” each other is unnecessary
processing. Again, fix in your mind that, for office processes, information
is the equivalent of the physical materials used in manufacturing processes.
Relevant here is the question of how you categorize requirements documents
and designs, as are used in engineering products whether physical (a new
cabinet) or intellectual (e.g., a new training course or software application).
Let's assume the case where the customer has not requested these as part
of his or her deliverable (an unwise customer choice, in our experience).
Are these examples of waste? Some people see them as waste because they
are not requested by the customer and are frequently ignored by the craftspeople
who build the product from their "sense" for how the product should
be executed or use an iterative process called prototyping. Our opinion
is different, especially given our experience in managing software development
projects. When a worker ignores the blueprint for a part he or she is machining
and "free-wheels" it, do we consider the blueprint waste or the
performance unacceptable? (Let's assume that the blueprint is properly constructed
and readily available.) If we find that the "blueprint" needs
to be redesigned, we consider that waste (rework) and we fix that problem.
We don't throw the blueprint. We see blueprints as "intellectual jigs"
that enable correct performance the first time. We see prototyping as planned
rework which we do consider waste (unless requested by the customer).
Adjust the Walk Through Method
The second adjustment to the walk through involves how it occurs. With
cognitive work processes, you will have the performer do a “talk through”
of his or her work, probably using the work process map as a guide. This
is necessary since the worker’s operations are largely mental and
therefore not observable.
Evaluating Waste in the Work Process
The mental nature of cognitive work also affects how you measure the waste
in the target work process. Again, since many of the operations the worker
is performing are not directly observable, you need to ask the performer to
do introspective reporting. In this method, the performer not only announces
each task as it is started but says what he or she is thinking and deciding
as the task is performed. Introspective reporting makes explicit to observers
what they cannot see— namely, the mental work of the performer. Otherwise,
we conduct the evaluation task just as described in the Kaizen Desk Reference
Standard (Task D2. Evaluate the Target Work Process) and as we do in shop
settings. We do measurements, observations, and interviews. We pool our information,
analyze the amounts and sources of waste, and summarize our findings. We use
the summary of our analyses the check the mission and goals we defined earlier
in the day in order to see if our new information dictates that we alter either.
If necessary, we modify the event's direction.
Solving the Performance Problems and Acting to Improve the Target
As Exhibit 3 indicates, there are no modifications to how we conduct either
of these tasks. Altogether, our experience is that the adjustments to the
Kaizen process are minor, at least when using the standardized approach specified
in the Kaizen Desk Reference Standard.
Guidance for Adjusting the Kaizen Process to Accommodate "Office"
Document a Scope for the Kaizen Event
- No adjustments needed.
- Pay special attention to Items B14 and B14a of scope document. These
items document the interfaces between the target work process and other
work processes. Item 14a specifically identifies organizations that
must be consulted before the target work process may be modified. Work
processes that generate information or service outputs are frequently
back office or support functions and their content and procedures are
sometimes controlled by the other organizations in whose service they
Analyze Whether to Conduct the Kaizen Event
- No adjustments needed.
- Because work processes that generate information or service outputs
are frequently undocumented and lack measurements, you need to be especially
careful in evaluating the presence of a work standard and the inclination
of management and performers to apply a standard and measure performance.
- You also need to evaluate the presence and impact of software systemsspecifically,
whether these systems so constrain the target work process that meaningful
change is not possible. It is not uncommon that waste within the target
work process is a product of the design of software support systems
yet control of the design of the software resides outside the work process.
- Some work processes that generate information outputs or service outcomes
occur over long cycle times because they are intermittent. An intermittent
work process starts then stops between operations. The actual time spent
in executing the process is much less than its cycle time. For example,
a personnel appraisal work process typically has a cycle time of one
year but the actual time spent in executing the process maybe no more
than three to four hours total. There maybe a meeting to establish appraisal
goals, a second meeting to provide feedback at the half-year mark, and
a third meeting to provide final feedback and the appraisal rating.
Pay attention to intermittence when evaluating whether the team can
observe the operation of the target work process within the time available
during a Kaizen event.
- In evaluating the business case, consider not only the direct cost
of operating the work process but also the cost of poor performance.
We had an office work process that was small in total operating cost
but because of its tardiness, it required the company to use expedited
delivery procedures to meet strict timeliness requirements. Its products
were being shipped internationally and expedited delivery resulted in
a highly significant cost which the Kaizen event could eliminate.
Prepare for the Kaizen Event
Perform the Kaizen Event
||Task D1. Focus
the Kaizen Event
Build a Description of the Target Work Process
- No adjustments are needed to building the overview of the target work
- Be prepared to document more decisions in your work process map. Work
processes that emphasize cognitive tasks are predominately a sequence
of decisions which culminate in some action. These decisions are more
complex and have a greater affect on results than the actions the decision's
trigger. Document both what the decision is and the knowledge that should
guide decision making. Be sure to note when there is no formal knowledge
defined to guide decision making. The absence of such knowledge can
have a significant impact on waste, usually in the form of rework.
- The flow of processes that generate an information or service output
frequently mirrors the flow of documents from input to output. For example,
an employee might submit purchase request (an document) to his or her
manager. That manager evaluates the request for relevance, appropriateness,
and against budgetary priorities. He or she decides whether to approve
the request or deny it. The document is updated with the decision and
possible the reason and sent on to purchasing or back to the employee.
You may find that tracing the path of input documents and the transformation
of the information they contain into outputs is an aid in understanding
the flow of work.
||Step D1-S2. Walk Through
the Target Work Process
- Your walk through plan will likely use "talk through" as
a method for cognitive tasks in that you cannot observe these activities
being performed. Much of the talk through may occur as you map the work
process especially if you do a thorough job on documenting the knowledge
that guides decision making.
- You will still want to observe the work area in which the process
is performed with a performer leading you through where each task or
decision is executed and how he or she moves between tasks. You will
want to observe any storage areas that are accessed, software screens
that the performer must work with in doing his or her job, and forms
or other media which convey information inputs or information outputs.
- Apart from the adjustments just mentioned, the walk through process
remains the same.
||Step D1-S3. Build the Mission
||Step D1-S4. Set Goals for
the Kaizen Event
||Step D1-S5. Define the
Do's and Don'ts
||Task D2. Evaluate
the Target Work Process
- Since the most significant work is mental, observing the performance
of a task has limited utility. Modify the observation part of the evaluation
accordingly. Consider using introspective reporting. In this method,
the performer not only announces each task at it is started but says
what he or she is thinking and deciding as the task is completed. It
makes explicit to the observers what they cannot see directlynamely,
the mental work of the performer.
- No other adjustments are necessary.
||Task D3. Solve
the Performance Issue
||Step D3-S3. Conduct an
||Task D4. Act
to Improve the Target Work Process
||Step D4-S1. Measure Results
||Step D4-S3. Conduct a Pilot
Institutionalize the Process Improvements
Published January 2006
Help Us Provide
You Better Content.
Tell us your thoughts about this
Be sure to name the article in your
1 "Deciding" is a very
large category and includes actions like completing forms, interpreting codes
or symbols, translating from one language to another, and many more. In fact,
it is an essential element of every complex cognitive task including, for example,
the ones we list above (planning, evaluating, etc.).
2 These observations are less
applicable the more the work process is related to matters of legal or fiduciary
responsibilitye.g., accounting, purchasing, medical testing, or charting
of the nation's waterways.
3We have added a job aid to guide
documenting decision logic to the Kaizen Tool Kit (Version 1.4 and later).
4The titles in this column refer
to sections in the Kaizen Desk Reference Standard.