Clasification Merchandising
Content
CYRUS C. WILSON
CHARLES D. GREENIDGE
Classification Merchandising:
An Overlooked Opportunity
for Increasing
Merchandising Profitability
EPITOME
The development of sophisticated information
technology in retailing has outpaced the development of sophisticated information systems. Up to
this point most installations of EDP equipment
liave been justified on the basis of cost displacement or savings through increased productivity.
Merclwindising applications have empha.sized elaborate unit control systems.
Profitability problems in retailing are most susceptible to improved dollar-based inventory management systems similar to NRMA's classification
merchandising concept. In this article research results of retailers using an advanced dollar system
are analyzed and conclusions are drawn. In addition, the suitability of advanced dollar systems for
meeting a variety of emerging retailing management problems is thoroughly discussed.
P^
THE USE OF ELECTRONIC DATA PROCESSING
in
business has generally been justified on one or both
of two main grounds. For one thing, EDP has been
used to automate certain highly specialized and already well-defined manual processes and, hence,
has served as a means of reducing expenses and improving productivity. Second, EDP has offered
management a broader information base for decision making and has thus furnished a means of improving overall performance and increasing profitability.
FALL / 1969 / VOL. X I I / N O . 1
The first of these two applications—which is often
called the housekeeping function of EDP—has generally been the best understood and most widely accepted. In fact, since the late 195O's, when the giant
retailers first began installing EDP, most applications have been limited to the automation of manual accounting functions where productivity gains
could be quickly exploited.^
The second application—tlie one which rationalizes EDP as a management decision-making toolhas been until recendy a rather vague and poorly
understood promise held out for some unspecified
point in the future. And yet it is this broader application which promises the greatest rewards, especially in retailing.
EDP as a merchandising management tool. In retailing, the logical area for information system development is merchandising management. Since
merchandise is the main source of revenue, a major
creator of expenses, and a key asset item, the overall profitability of a retail operation is largely a
function of the profitability of inventory investments. This is why students of retailing have long
felt that EDP would yield its greatest payout by
generating information for merchandising management.In harnessing EDP to the merchandising function, retailers at first concentrated on automating
unit control systems. This was where the great
speed and mass data-handling capacities of EDP
53
equipment seemed most useful. By 1963, however,
the National Retail Merchants Association
(NRMA) had become interested in the possibilities
of dollar-based information systems, and the publication of NRMA's Standard Classifications in 1967
was a big step toward a more systematic way of organizing dollar information on sales, reductions,
gross margin, and inventory investment.^
The NRMA system, which is now known as classification merchandising, can be used by a broad
range of retailers—from small specialty stores to
giant merchandise organizations. It is useful whenever the retail manager must plan and control a
large number of merchandise items of low- to
moderate-unit value, and it is especially useful in
multidepartment or multistore operations. Most importantly, classification merchandising systems provide profitability information for relatively narrow
segments of merchandise investment; because of
this they are well suited to dealing with today's adverse profitability trends in retailing.
In fact, we believe that retailers-especially large
ones—are overlooking a major opportunity for increasing merchandising profitability. Field research
findings show that a controlled sample of small
men's wear retailers improved their profit performance through a classification merchandising approach to inventory management. At the same time,
readings in the current periodical literature and
personal interviews show tliat the great majority of
retailers, especially department stores and other
large concerns, are not using or do not understand
the classification merchandising concept.
What is the potential of classification merchandising? Our evaluation is organized in five steps:
1 / A brief analysis of merchandising profitability
trends, using the department store industry as an example.
Cyrus C. Wilson is Assistant Professor of
Marketing, Indiana University Graduate
School of Business and a Faculty Associate
of Management Horizons, Inc. He has done extensive research in automated merchandising.
Charles D. Greenidge is Assistant Professor
of Marketing at the University of Colorado
and a Faculty Associate of Management
Horizons, Inc. Retail management information systems are his special interest.
54
2 / Discussion of the limitations of unit systems for
solving profitability problems.
3 / Description of the classification merchandising
concept and evidence indicating the degree of misunderstanding that exists among large retailers.
4 / Profit improvement results from a panel of men's
wear specialty stores which used a classification merchandising approach to merchandising management.
5 / Discussion of the potential applications of classification merchandising for retailers in general, with special emphasis on large retailers.
Department Stores
Over the last ten years, several disturbing trends
have developed in retailing performance—especially in department stores. We have chosen the department store industry as an example because it
covers the broad spectrum of merchandising experience and because it stands to benefit greatly from
classification merchandising.
Merchandising performance data for 1957-1966
are summarized in Table I. In those ten years, a
slight increase in gross margin as a percentage of
net sales was more tlian offset by declining inventory turnover—with the net result of a decline in
overall merchandising profitability.
There are several explanations for this. First, the
past decade has seen continued expansion into suburban branch store locations. By 1965, in fact, over
50 per cent of department store sales were in
branch stores.*
This expansion of branch operations has actually
weakened merchandising performance. For one
thing, downtown department managers have usually bought the merchandise for all locations hecause this seemed the best way of keeping buying
expenses low and of avoiding duplication of effort
in dealing with suppliers. But the weakness of this
approach is that it dilutes the department manager's knowledge of merchandise movement. Traditionally, top management has felt that department
managers should spend a lot of time on the selling
floor so that they could personally observe customer
reaction and sales activity in the department. This
knowledge was considered essential for making
merchandise-buying decisions. But \n\h several
branch locations, such personalized information
feedback was not possible. Some other way was
needed to get merchandise movement information
into department managers' hands.
Califomia Management Review
TABLE I. MERCHANDISING
1957
1958
1959
PERFORJUING D A T A FOR DEPARTMENT
Median Performance
1960
1961
1962 1963
1964
1965
1966
Gross margin to
sales (%)
36.4
36.1
36.4
36.1
36.2
36.2
36.5
36.9
37.2
37.4
Net sales to average
inventory at cost
(Times)
6.3
6.4
6.5
6.2
6.3
6.3
5.6
5.7
5.7
5.7
Gross margin / dollar
inventory at cost ($)
2.30
2.30
2.36
2.24
2.29
2.28
2.03
2.11
2.11
2.12
Net profit (A.T.) / dollar inventory at cost ($)
.18
.20
.18
.15
.16
.17
.13
.15
.18
.17
Inventory turnover
(Times)
3.8
3.8
3.9
3.8
3.9
3.6
3.4
3.5
3.5
3.4
Per Cent
Change
1957-1966
2.2%
(9.5%)
(7.8%)
(5.5%)
(10.5%)
SouncES: Controllers' Congress, National Retail Merchants Association, New York, and Di\ision of Research, Graduate School of
Business, Han-ard University.
* Data for 1957-1962 are for all owned departments, irrespective of the inventory accounting system used. Data for 1963-1966 are for
departments using the retail method of accounting only.
Anotlier factor has been the development of the
"shop" or "boutique" concept of merchandise presentation. With this approach, merchandise from
many separate departments is assembled in a distinct display focused on a particular shopping need.
Again, the individual department manager's ability
to gather information visually is weakened.
Still another factor is the introduction of many
new merchandise items. The investment required to
offer a full assortment in a given line of merchandise has grown persistently large. Staple merchandise in many instances is developing a "fasliion"
component, and assortment requirements, even in
traditional fashion areas, are broadening. The implications for the department manager are obvious.
Even if he did not have to contend with multilocation merchandising, tlie proliferation of items has
greatly increased his information needs.
However compelling the reasons for the declining
profitability in department stores (and in retailing
generally), they are still only explanations for an
unsatisfactory situation. And explanations cannot be
put in the bank. They cannot be distributed to
stockholders as dividends, and they will not finance
expansion and growth. Rather, retailers need to reverse current trends in merchandising profitability.
The key element, of course, is the increasing need
for information. Certainly, the power of EDP to automate information feedback has wide currency as
the answer to today's merchandising dilemma. And
unit-based systems in particular have been tlie cen-
FALL / 1969 / VOL. XII / NO. 1
ter of attention. Yet results have been quite limited.
The many promises made for unit systems have not
been kept and perhaps never wiU be.
Even when unit-control systems are brought to full
maturity, the challenge of better inventory management
will not be entirely met. The reason, of course, is that
while unit-control systems can keep merchandise on the
shelf, they do not provide suflBcient details of the elements determining gross margins. . . .^
Properly conceived, unit systems provide information about the number of units of a merchandise
item sold and on hand, or in sight. This information
is thought to be useful in deciding what action a
buyer ought to take on a given item of merchandise. Should it be reordered, transferred, marked
down, or promoted in some way?
The hoped-for results of having such unit activity
information are (1) better in-stock position and,
hence, fewer lost sales from being out of stock; and
(2) fewer overstocked situations and, hence, lower
inventories in relation to the sales level supported.
Tliese results are expected to enhance profitability
tlirough increased sales and faster inventory turnover, and there is evidence that sometimes this has
been achieved. Thus, the current emphasis on unit
systems is not entirely unfounded. But there are
limitations on unit systems. And they are basically
twofold:
• They are extremely expensive to operate.
• They do not provide information about the profitability of stock investments made.
55
In a highly controlled experiment involving one
department in a department store, Joseph Buchan
and Ernest Koenigsberg demonstrated that a unit
control system can achieve better inventory
balance." Tlie difficulty with this and similar controlled demonstrations is that the full-cost impact of
widespread applications is often not taken into account. Increases in sales as well as turnover should
be weighed against the costs of system operation.
According to the data processing manager of a large
department store which is aggressively expanding
its unit systems using EDP, there are cost factors in
large-scale applications which do not stand out in
small-scale experiments. Receiving and marking
costs are examples. When items were added to the
automated unit control program, the need to record
and mark merchandise items by unit, as opposed to
department or classification, increased the flowthrough time by ten. Data processing costs are also
a good example. According to this same manager,
there are no economies of scale in unit systems. In
fact, his experience shows that computer processing
times for unit systems grow geometrically as the
number of stock-keeping units on the system is
increased.^
Even if accurate and relatively inexpensive unit
data were available, it would still provide only part
of the information needed. Unit information can be
very useful in planning and controlling assortments,
but by itself does not show profitability of merchandise investment because it does not attach a dollar
value to sales, reductions, cost of merchandise, or
average stock investment. This point is well made
in all leading textbooks on merchandising, but it is
worth re-emphasizing in view of current industry
performance trends and emphasis on unit systems.^
A Definition
In a sense, there is nothing startlingly new about
classification merchandising.
The preparation of dollar classification reports reqtiires exactly tlie same kinds of infoimation that are
presently used in the preparation of [traditional] department accounting records. These may be broadly classified as follows:"
1 / Sales and returns.
2 / Inventory: opening inventory; purchases and vendor returns; transfers; and additional markons and cancellations.
3 / Markdowns and cancellations.
56
4 / Open-to-buy: merchandise plans; and orders
placed.
The major difference between traditional dollar
control and the classification merchandising approach is the level of application. In fact, tliis very
point has caused confusion among traditional department and specialty stores.
Retailers themselves appear to be divided by the new
classification system these days. There are those that acclaim Classification Merchandising as today's—and tomorrow's—better way of learning customer demand, of
measuring merchandise performance, of planning sales,
of assigning inventory dollars, of space—and those who
claim we have had the likes of this for years.^°
This same problem was brought out in a study
performed by a leading firm of certified public
accounts.^^ Retailers apparently do not have a clear
idea of what is meant by a "classification" or of
what the objectives of "classification merchandising"
are. In the study quoted above, 50 department
stores were asked if dollar data on sales and stocks
were gathered below the department level. These
were the results:
• 32 stores indicated that they coflected at least sales
below the department level.
• 12 stores indicated that they developed stock plans
and open-to-buy controls below the department level.
• None of tlie stores was developing gross margin data
below the department level.
Typical of tlie stores included in this study is
Rich's, a large, progressive department store located
in Atlanta, Ceorgia. Rich's has about 1,400 merchandise breakdowns below the department level
(about 10 per department), and each breakdown
averages an inventory investment of $5,000.^- While
this may seem to be a fairly detailed breakdown of
inventory investment, it does not really approach
the level of analysis anticipated by proponents of
classification merchandising.
To talk in terms of numbers of inventory breakdowns, however, is to miss the point on classification merchandising. The development of classification merchandising basically reflects the demise of
the department as the primary responsibility and
profit center in department store retailing. As Edward Waterbury of Woodward & Lothrop recently
said:
The rapid expansion of our business in terms of the
number of stores, as well as volume and the remoteness
of the buyer from the ever-growing percentage of the
California Management Review
business, accelerated the demand for accurate and timely
information at the classification
In this sense, the merchandise class is becoming the
basic building block of merchandising.
Anticipating tliat "classes are the departments of
the immediate future . . ."" NRMA has developed
standard definitions for a four-level breakdown
based on increasingly specific merchandise characteristics.^'
First level: Broad groupings, most nearly similar to
historical store divisions; but designed as logical groupings of related merchandise for broad customer end-use.
Merchandise Groups.
Second level: Family groupings most nearly similar
to historical store departments; but designed to break
down merchandise by types, not by size groupings,
price ranges, etc. Merchandise Demand Genters.
Third level: Assortment of items within each secondlevel breakdown, on basis of relationship for customer
end-use. Classes.
Fourth level: Groupings on basis of items that are interchangeable or substitutable for specific customer
end-use. Gategories.
The NRMA classification scheme is basically a
four-digit coding system where the first level corresponds to the first digit. This coding system has the
capacity for 9 Merchandise Croups, 99 Merchandise Demand Centers, 999 Classes, and 9999 Categories. NRMA's immediate purpose in promulgating
standard definitions is to facilitate collection and
exchange of merchandising results at the class or
third level.
Unit VS. Dollar Systems
For the individual firm, whether or not it embraces NRMA's definitions, classification merchandising offers many benefits:
^ One is the abihty to measure profitability of narrower and more homogeneous groupings of merchandise.
This perspective on profltabihty allows management to
make better decisions on the allocation of the scarce resources of space, inventory dollars, and promotional expenditures.
• Management should be able to pinpoint expansion
opportunities and isolate problems or poor performance
areas.
I Merchandise plans and controls will be more meaningful at a class (or even a category) level because they
will be focused on a particular customer end-use. Purchasing for one class will not be closed ofF because another class is over budget, as often happens when plans
FALL / 1969 / V O L . X I I / N O . 1
and conbols are fonnulated for a more conglomerate
merchandise grouping, such as a department or dissection.
I Merchandising by class also facilitates planning and
controlling in the modem context of multiple locations
and boutique operations.
Usually unit systems and dollar systems are
thought to complement one another, although they
serve different purposes. Unit systems provide
movement and stock level information in terms of
physical pieces of merchandise, while dollar systems
provide similar information in financial terms. Traditionally, unit systems have been specifically oriented toward merchandising operations, whereas
dollar systems have been used to regulate total investment by major segments of the business. The
classification merchandising approach, however, extends the level of detail of dollar systems to tlie
point where they can also become useful in a merchandising sense. In some cases, fine-line dollar systems can even be substituted for unit systems. If
this is true, what might be the advantages of reallocating development funds and management emphasis from unit systems to fine-line dollar systems?
First, classification merchandising gives dollar information on sales, markups, reductions, margins,
and investments. This is critically needed on a detailed breakdown basis to deal with the adverse
profitability' trends in department stores and in other
sectors of retailing.
Second, even tliough proponents of unit systems
argue that unit information can be built up into dollar figures, the fact is that executives using unit data
are seldom able to bridge the gap. The problems involve system accuracy—accounting for reductions,
policing tlie entry of unit receipts, and even capturing unit sales data. Buyers also tend to be myopic
about a given merchandise item or stock-keeping
unit, and as a result they are less able to spot trends
across an entire assortment or group of related merchandise items. Furthermore, planning and controlline on the basis of an individual item can be very
hard because of the relatively \vide fluctuations of
item demand around its average value. Such fluctuations tend to be less chaotic for the entire range of
items in a customer end-use category when taken as
a whole.
Third, tlie expenses of establishing and maintainin o- a classification system are generally lower than
those of a unit system. Since most large retailers al-
57
jump in number of classifications was from 16 to
180. The median amount of inventory investment
per classification decreased from $5,000 to about
$600. Information is now being received monthly,
whereas less than half of the retailers enjoyed such
frequency before. More important, the information
is available 3 to 5 days after the end of the month,
as opposed to anywhere from 20 to 90 days using
manual methods. Taken together, such quantitative
and quahtative improvements in infomiation would
logically point to improved performance. But, in
spite of such strong logic, there are several a priori
reasons why improved inventory management perfomiance might not occur, especially during the
first year:
ready have some form of departmental dollar control, all the administrative procedures and information sources for a classification system already exist.
The new expenses of replacing departmental breakdowns with more comprehensive classification
breakdowns are rather marginal. Finally, no matter
how sophisticated the classification system, it is unlikely to exceed 10,000 categories, whereas for a
unit system, 10,000 stockkeeping units doesn't even
constitute a good start.
Certainly, good arguments can be developed for
classification merchandising, but tliey are meaningless unless they can be backed up with evidence of
concrete results. To date, the evidence consists of
isolated success stories, but then no systematic research into the effect of a classification merchandising program on profitability has been previously reported. To help fill this gap, the first-year results of
our continuing study of the impact of classification
merchandising on a sample of thirty smaller men's
wear retailers is presented below.
The inventory management report. An example
of the inventory management report, the heart of
the new information system, is shown in Table II.
This report permits the retailer to divide his inventory investment into as many as 799 classifications.
The report represents a quantum jump for all tlie
retailers studied in terms of sophistication, frequency, and timeliness of information. The median
• Inventories tend to increase initially as a result of
taking corrective action on imbalanced stocks.^°
• There is no history upon which to base planning
for the newly formed classifications.
• The basic tendency of small companies is to shy
away from planning.^^
Results and Evaluation
Results of improved information. Several important findings resulted from the analysis of first-year
data. Dramatic changes in inventory management
performance were observed. Table III summarizes
the first-year results of a comparison of observed in-
TABLE II. AN INVENTORY MANAGEMENT REPORT OF THE TYPE USED BY RETAILERS IN THE PANEL
% Sales
Sales
Total
Gross
Profit
Maint.
ft/1 Q 7.fTI Y\
ijXtXl ^lU
07
/C
Mos.
xnv. on ^
Hand
Additions
Cost
Retail
% Inv.
Mark
Downs
End Inv.
Cost
End Inv.
Retail
—
32.00
—
—
—
8,918.00
3,483.00
2,695.00
289.00
141.00
4,734.00
3,138.00
2,774.00
376.00
634.00
924.00
519.00
200.00
1,267.00
2,792.00
894.00
714.00
25.9
10.1
7.8
34,492.00
100.0
Total
Dept.
2,440.00
400.00
800.00
100.00
50.00
450.00
920.00
960.00
100.00
200.00
300.00
150.00
50.00
750.00
1,850.00
250.00
200.00
24.6
Totals 9,970.00
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
58
2.5
2.0
856.00
103.00
232.00
28.00
16.00
94.00
280.00
384.00
39.00
62.00
99.00
48.00
17.00
285.00
666.00
94.00
60.00
35.0
31.0
29.0
28.0
32.0
32.0
35.0
40.0
39.0
31.0
33.0
32.0
34.0
38.0
36.0
42.0
30.0
100.0
3,363.00
33.7
4.0
8.0
1.0
.5
4.5
9.2
9.6
1.0
2.0
3.0
1.5
.5
7.5
18.6
3.6
8.7
3.4
2.9
2.8
980.00
500.00
960.00
10.7
3.4
2.9
3.8
3.2
3.1
8.5
4.0
1.7
1.5
3.6
3.6
375.00
810.00
520.00
90.00
175.00
130.00
145.00
75.00
350.00
910.00
175.00
160.00
552.00
1,247.00
867.00
148.00
254.00
194.00
213.00
114.00
565.00
1,422.00
302.00
228.00
75.00
67.00
—
5,796.00
2,403.00
1,912.00
208.00
96.00
3,219.00
2,040.00
1,664.00
229.00
437.00
619.00
353.00
132.00
785.00
1,786.00
519.00
500.00
3.5
6,355.00
9,692.00
194.00
22,698.00
—
—
1,508.00
725.00
1,353.00
—
—
—
—
—
—
—
—
—
—
20.00
California Management
.8
.4
13.8
9.1
8.0
1.0
1.8
2.7
1.5
.6
3.7
8.1
2.6
2.1
Review
/entory investment versus predicted inventory investment, and observed gross margin dollars earned
I'ersus predicted gross margin dollars earned. The
predicted values of gross margin dollars earned and
inventory investment are based on a least squares
regression of each value on sales volume for the period 1961-1965. The differences between predicted
and actual value for each variable were converted
to standard units and partitioned to highlight the
degree of change. Changes of greater than one standard deviation in eitlier direction are considered to
be significant. For the panel as a whole, the firstyear results indicate strong improvement in gross
margin doUars earned and an equally strong increase, an unfavorable change, in inventory investment.
TABLE III. COMPARISON OF
FROM PREDICTED VALUES FOR
THE AMOUNT OF CHANGE
GROSS MARGIN DOLLARS
EARNED AND INVENTORY INVESTMENT,
PANEL RETAILERS,
1966
Number of Retailers
Gross Margin
Inventory
Dollars Earned
Investment
Greater than + 1 standard
deviation
Between + 1 and —1
standard deviation
Less than — 1 standard
deviation
5
17
11
11
15
2
SOURCES: Cyrus C. Wilson, "Explaining Management Performance Under Conditions of Improved Information in Smaller
Retail Firms"; and Charles D. Greenidge, "The Effect of Improved Information on Profitability in Smaller Retailing Firms";
both unpublished Ph.D. theses, Ohio State University, 1967.
Interpretation of results. The strong improvement
showTi by panel members on the gross margin-tosales ratio tends to support the idea that classification merchandising is an effective way to improve
profitability. However, panel results showed an unfavorable outcome for the sales-to-inventory ratio.
But there are strong indications that the increases
in this measure are transitory. Improved inventory
investment information on a fine-Hne basis is likely
to result in two kinds of management action:
1 / Increasing inventory on some classes to avoid lost
sales from being out of stock on briskly demanded merchandise.
2/Reducing inventory on some classes which are
presently overstocked in relation to demand.
FALL / 1969 / V O L . X I I / N O . 1
Understocked conditions can be quickly corrected
by placing orders for additional merchandise. Overstocked conditions, however, usually will not be
corrected immediately but will slowly correct themselves as normal sales reduce the excess stocks. The
net result is that inventory will tend to increase
immediately after improved information becomes
available, and better balance will be achieved only
over a period of time, as the excess merchandise is
sold off.
Transitory increases in inventory investment have
been observed in several merchandising information
systems. For example, Buchan and Koenigsberg
noted that inventory investment 'iDulged" immediately after system installation and did not come
back into balance with sales volume until six
months later.^^ Of course, their study involved a department in a large department store with brisk customer traffic, whereas our study covers smaller
stores with correspondingly lower traffic. Therefore,
while the comments of our panel retailers and the
results of other studies show that the unfavorable
change in the sales-to-inventory ratio is a temporary
phenomenon, it is very hard to say how long the
"bulge" will last. We might conclude, however, that
tlie ultimate result of improved information will be
a liigher salcs-to-inventory ratio and, hence, an
overall improvement in profitability.
Conclusions
The evaluation procedure presented earlier also
reflects the argument for greater emphasis on the
classification merchandising approach.
k The department store industry, specifically, and retailing, generally, are faced with a merchandising profitability problem.
I Present emphasis on unit systems for reasons of cost
and kind of Infonnation provided is not solving the
profitabilit}' problem.
^ Fine-line dollar systems have been proposed but
have not received a great deal of support. The executives
of large department and specialty stores seem to misunderstand the classification merchandising concept.
• Smaller specialty stores using a classification merchandising approach have shown strong improvement for
gross margin performance in the first year and can be
expected to show overall improvement for merchandising profitability in the future.
The experiences of panel retailers with fine-line
dollar systems offer some good insights for large re-
59
tailers. For example, the variety of classification
schemes used effectively by the panel retailers suggests that the NRMA standard classifications, while
necessary for figure exchange, may not be the most
fruitful metliod of classification for purposes of merchandise investment management. Several panel retailers did not conform to the general recommendation of classifying on the basis of end-use or detailed merchandise characteristics and employed
bases of price-line, vendor, season, or even color.
These retailers often had specific examples of how
classifying on a basis other than detailed merchandise characteristics had helped them eliminate a
price-line or achieve better control over a vendor.
From this we conclude that many alternative classification schemes might well be used below the third
level, altliough additional research needs to be
done, of course, to find the relative effectiveness of
alternative methods for various lines of merchandise.
NRMA seems, implicitly, to want to avoid classifying on any dimension other than merchandise
end-use or detailed characteristic. Part of the reason
could be a desire to avoid pushing dollar classification systems too deeply into the traditional province
of unit systems—an artificial limitation. As some of
the panel retailers indicated, too much information
can be a limitation to management. Because of the
large numbers of merchandise items, even in
smaller retail stores, managers cannot effectively
plan and control on an item basis. The strategy of
most retailers in the panel was to use fine-line dollar
information as a trigger mechanism. They reviewed
the dollar results of each class, looking for exceptional deviations from expectations. When they
spotted a class with results that seemed unusual,
they focused their attention on that class for further
study.
The next step might be a complete assortment review of the class tlirough a physical inventory to
determine the exact source of trouble. This, of
course, is nothing more than the management-byexception principle—but it's noteworthy that the
panel retailers felt better able to do this with
broader dollar data than witli more detailed unit
data.
For staple merchandise, fine-line dollar data allowed several panel retailers to retain financial control of a classification while delegating the actual
assortment merchandising to a supplier's representa60
tive. This technique is often used successfully in the
hosiery departments of large retail stores where the
supplier is actually operating the unit system under
the supervision of the department manager. Tliis is
another strong argument for classification merchandising as opposed to unit systems. It may very well
be that unit systems, or preoccupation with items, are
more realistically and economically a vendor function rather than a retailer function. Naturally, this
presupposes more formalized vendor-retailer cooperation in planning model assortments and seasonal
merchandising programs. Actually, there is already
a trend toward more tightly structured vendor-retailer relationships—as witnessed by the growing
popularity of programmed merchandising.
The emergence of the classification or category as
tlie basic level of merchandise planning and control
will also affect management practices. The present
departmental and subdepartmental breakdowns
generally have such a heterogeneous merchandise
composition that dollar measures, such as open-tobuy, are often criticized as misleading. Acceptable
performance for the department as a whole may
simply be an average of very good and very bad for
the classifications or categories which are grouped
together to form the department. There has been a
"credibility gap" about the usefulness of such information. But dollar data on a classification or categor)' basis should be more directional since it is
based on more homogeneous merchandise groupings. This should enhance the value of open-to-buy
controls and encourage the use of several other performance measures which, until now, have played
rather minor roles.
More attention will be given to return-on-investment measures on a classification basis to measure
profitability and productivity and to allocate limited
factors. Cross margin return on average inventory
investment is one example of the type of profitability measure that fine-line dollar information will
make more meaningful. Other measures would include: gross margin return on space utilized, margin after direct expenses return on inventory and
space, and return on promotional outlays. With all
these measures we need to develop standards of
performance for classes, and we need to educate
management to their use.
Class performance measures, new and old, should
ease the dilemma created by the breakdown of the
department as a profit center and the growing sepaCalifornia Management Review
ration of buying and selling into separate job responsibilities." In the future, merchandising will be
more and more a joint responsibility, and procedures should be adjusted to give the store operations executive a greater voice in merchandising
tliose classifications from which he is supposed to
sell. He should be able to work with the buying executive in developing the merchandise program for
his location. Without suggesting an adversary relationship, the store operations executive could treat
the buying executive as a "source of supply," and he
could then be held responsible not only for expenses but also for the profitable use of working
capital and space at his location. The operations
man would also need to be able to expand or contract the part of his budget which is allocated to a
given classification. Tlie buying executive, working
from his own budget, could then be held responsible for total classification performance across all locations.
chandise Gontrol," Chain Store Age, April 1962, pp.
E21-E23.
Tlirough such joint merchandising responsibility,
the store could establish a store operations profit
center for each location, and could simultaneously
create a classffication profit center oriented toward
a specific buying responsibility, regardless of selling
location.
In summary, there appear to be many potential
benefits for large retailers who adopt tlie classification merchandising concept. Although this system is
now widely misunderstood, it seems to be an effective means of reversing adverse profitability trends.
Most importantly, it appears to be consistent with
emerging patterns of change in the retailing industry.
10. George Baylis, "Are the Brakes Too Tight to Let
the Train Roll?" Retail Control, Oct. 1965, p. 18.
REFERENCES
1. R. Stanley Laing, "Tomorrow's Retail Systems,"
address to the Ninth Annual EDP Gonference for Retailers, Retail Research Institute, National Retail Merchants Association (NRMA), Sept. 26, 1967.
2. For example, see "EDP's Major Justification: Mer-
FALL / 1969 / VOL. XII / NO. 1
3. Standard
1967).
Classifications
(New
York:
NRMA,
4. Gontrollers' Gongress, Merchandising and Operating Results of Department and Specialty Stores in 1966
(New York: NRMA, 1967), p. 2.
5. Laing, "Tomorrow's Retail Systems."
6. Joseph Buchan and Ernest Koenigsberg, Scientific
Inventory Management (Englewood Gliffs, N.J.: Prentice-Hall", 1963), p. 75
7. Data in this paragraph are provided by the data
processing manager of a leading department store. The
name of the individual and his finn are withheld by request.
8. For example, see William R. Davidson and Alton
F. Doody, Retailing Management (New York: The
Ronald Press Gompany, 1966), pp. 404-405.
9. Albert 1. Schott as quoted in Putting Classification
Merchandising to Work (New York: NRMA, 1967), p.
20.
11. GarroU E. Ebert, "Merchandise Glassification,"
ibid., Feb. 1966, p. 23.
12. Authors' conversation with John Bradley, Gonti^oller of Rich's.
13. Edward S. Waterbury, "Gathering Glassification
Data by Manual and Mechanical Systems," Retail Control, Feb. 1966, p. 20.
14. William Buiston in Standard Classifications, p.
31.
15. Putting Classification Merchandising to Work, p.
65.
16. Buchan and Koenigsburg, p. 75.
17. See Roger A. Golde, "Practical Planning for
Small Business," Harvard Business Review, XLII:6
(Nov.-Dec. 1964), 148
18. Buchan and Koenigsburg, p. 75.
19. Wheelock H. Bingham and David L. Yunich,
"Retail Reorganization," Harvard Business Review,
XUUA (July-August 1965), 129-146.
61
CHARLES D. GREENIDGE
Classification Merchandising:
An Overlooked Opportunity
for Increasing
Merchandising Profitability
EPITOME
The development of sophisticated information
technology in retailing has outpaced the development of sophisticated information systems. Up to
this point most installations of EDP equipment
liave been justified on the basis of cost displacement or savings through increased productivity.
Merclwindising applications have empha.sized elaborate unit control systems.
Profitability problems in retailing are most susceptible to improved dollar-based inventory management systems similar to NRMA's classification
merchandising concept. In this article research results of retailers using an advanced dollar system
are analyzed and conclusions are drawn. In addition, the suitability of advanced dollar systems for
meeting a variety of emerging retailing management problems is thoroughly discussed.
P^
THE USE OF ELECTRONIC DATA PROCESSING
in
business has generally been justified on one or both
of two main grounds. For one thing, EDP has been
used to automate certain highly specialized and already well-defined manual processes and, hence,
has served as a means of reducing expenses and improving productivity. Second, EDP has offered
management a broader information base for decision making and has thus furnished a means of improving overall performance and increasing profitability.
FALL / 1969 / VOL. X I I / N O . 1
The first of these two applications—which is often
called the housekeeping function of EDP—has generally been the best understood and most widely accepted. In fact, since the late 195O's, when the giant
retailers first began installing EDP, most applications have been limited to the automation of manual accounting functions where productivity gains
could be quickly exploited.^
The second application—tlie one which rationalizes EDP as a management decision-making toolhas been until recendy a rather vague and poorly
understood promise held out for some unspecified
point in the future. And yet it is this broader application which promises the greatest rewards, especially in retailing.
EDP as a merchandising management tool. In retailing, the logical area for information system development is merchandising management. Since
merchandise is the main source of revenue, a major
creator of expenses, and a key asset item, the overall profitability of a retail operation is largely a
function of the profitability of inventory investments. This is why students of retailing have long
felt that EDP would yield its greatest payout by
generating information for merchandising management.In harnessing EDP to the merchandising function, retailers at first concentrated on automating
unit control systems. This was where the great
speed and mass data-handling capacities of EDP
53
equipment seemed most useful. By 1963, however,
the National Retail Merchants Association
(NRMA) had become interested in the possibilities
of dollar-based information systems, and the publication of NRMA's Standard Classifications in 1967
was a big step toward a more systematic way of organizing dollar information on sales, reductions,
gross margin, and inventory investment.^
The NRMA system, which is now known as classification merchandising, can be used by a broad
range of retailers—from small specialty stores to
giant merchandise organizations. It is useful whenever the retail manager must plan and control a
large number of merchandise items of low- to
moderate-unit value, and it is especially useful in
multidepartment or multistore operations. Most importantly, classification merchandising systems provide profitability information for relatively narrow
segments of merchandise investment; because of
this they are well suited to dealing with today's adverse profitability trends in retailing.
In fact, we believe that retailers-especially large
ones—are overlooking a major opportunity for increasing merchandising profitability. Field research
findings show that a controlled sample of small
men's wear retailers improved their profit performance through a classification merchandising approach to inventory management. At the same time,
readings in the current periodical literature and
personal interviews show tliat the great majority of
retailers, especially department stores and other
large concerns, are not using or do not understand
the classification merchandising concept.
What is the potential of classification merchandising? Our evaluation is organized in five steps:
1 / A brief analysis of merchandising profitability
trends, using the department store industry as an example.
Cyrus C. Wilson is Assistant Professor of
Marketing, Indiana University Graduate
School of Business and a Faculty Associate
of Management Horizons, Inc. He has done extensive research in automated merchandising.
Charles D. Greenidge is Assistant Professor
of Marketing at the University of Colorado
and a Faculty Associate of Management
Horizons, Inc. Retail management information systems are his special interest.
54
2 / Discussion of the limitations of unit systems for
solving profitability problems.
3 / Description of the classification merchandising
concept and evidence indicating the degree of misunderstanding that exists among large retailers.
4 / Profit improvement results from a panel of men's
wear specialty stores which used a classification merchandising approach to merchandising management.
5 / Discussion of the potential applications of classification merchandising for retailers in general, with special emphasis on large retailers.
Department Stores
Over the last ten years, several disturbing trends
have developed in retailing performance—especially in department stores. We have chosen the department store industry as an example because it
covers the broad spectrum of merchandising experience and because it stands to benefit greatly from
classification merchandising.
Merchandising performance data for 1957-1966
are summarized in Table I. In those ten years, a
slight increase in gross margin as a percentage of
net sales was more tlian offset by declining inventory turnover—with the net result of a decline in
overall merchandising profitability.
There are several explanations for this. First, the
past decade has seen continued expansion into suburban branch store locations. By 1965, in fact, over
50 per cent of department store sales were in
branch stores.*
This expansion of branch operations has actually
weakened merchandising performance. For one
thing, downtown department managers have usually bought the merchandise for all locations hecause this seemed the best way of keeping buying
expenses low and of avoiding duplication of effort
in dealing with suppliers. But the weakness of this
approach is that it dilutes the department manager's knowledge of merchandise movement. Traditionally, top management has felt that department
managers should spend a lot of time on the selling
floor so that they could personally observe customer
reaction and sales activity in the department. This
knowledge was considered essential for making
merchandise-buying decisions. But \n\h several
branch locations, such personalized information
feedback was not possible. Some other way was
needed to get merchandise movement information
into department managers' hands.
Califomia Management Review
TABLE I. MERCHANDISING
1957
1958
1959
PERFORJUING D A T A FOR DEPARTMENT
Median Performance
1960
1961
1962 1963
1964
1965
1966
Gross margin to
sales (%)
36.4
36.1
36.4
36.1
36.2
36.2
36.5
36.9
37.2
37.4
Net sales to average
inventory at cost
(Times)
6.3
6.4
6.5
6.2
6.3
6.3
5.6
5.7
5.7
5.7
Gross margin / dollar
inventory at cost ($)
2.30
2.30
2.36
2.24
2.29
2.28
2.03
2.11
2.11
2.12
Net profit (A.T.) / dollar inventory at cost ($)
.18
.20
.18
.15
.16
.17
.13
.15
.18
.17
Inventory turnover
(Times)
3.8
3.8
3.9
3.8
3.9
3.6
3.4
3.5
3.5
3.4
Per Cent
Change
1957-1966
2.2%
(9.5%)
(7.8%)
(5.5%)
(10.5%)
SouncES: Controllers' Congress, National Retail Merchants Association, New York, and Di\ision of Research, Graduate School of
Business, Han-ard University.
* Data for 1957-1962 are for all owned departments, irrespective of the inventory accounting system used. Data for 1963-1966 are for
departments using the retail method of accounting only.
Anotlier factor has been the development of the
"shop" or "boutique" concept of merchandise presentation. With this approach, merchandise from
many separate departments is assembled in a distinct display focused on a particular shopping need.
Again, the individual department manager's ability
to gather information visually is weakened.
Still another factor is the introduction of many
new merchandise items. The investment required to
offer a full assortment in a given line of merchandise has grown persistently large. Staple merchandise in many instances is developing a "fasliion"
component, and assortment requirements, even in
traditional fashion areas, are broadening. The implications for the department manager are obvious.
Even if he did not have to contend with multilocation merchandising, tlie proliferation of items has
greatly increased his information needs.
However compelling the reasons for the declining
profitability in department stores (and in retailing
generally), they are still only explanations for an
unsatisfactory situation. And explanations cannot be
put in the bank. They cannot be distributed to
stockholders as dividends, and they will not finance
expansion and growth. Rather, retailers need to reverse current trends in merchandising profitability.
The key element, of course, is the increasing need
for information. Certainly, the power of EDP to automate information feedback has wide currency as
the answer to today's merchandising dilemma. And
unit-based systems in particular have been tlie cen-
FALL / 1969 / VOL. XII / NO. 1
ter of attention. Yet results have been quite limited.
The many promises made for unit systems have not
been kept and perhaps never wiU be.
Even when unit-control systems are brought to full
maturity, the challenge of better inventory management
will not be entirely met. The reason, of course, is that
while unit-control systems can keep merchandise on the
shelf, they do not provide suflBcient details of the elements determining gross margins. . . .^
Properly conceived, unit systems provide information about the number of units of a merchandise
item sold and on hand, or in sight. This information
is thought to be useful in deciding what action a
buyer ought to take on a given item of merchandise. Should it be reordered, transferred, marked
down, or promoted in some way?
The hoped-for results of having such unit activity
information are (1) better in-stock position and,
hence, fewer lost sales from being out of stock; and
(2) fewer overstocked situations and, hence, lower
inventories in relation to the sales level supported.
Tliese results are expected to enhance profitability
tlirough increased sales and faster inventory turnover, and there is evidence that sometimes this has
been achieved. Thus, the current emphasis on unit
systems is not entirely unfounded. But there are
limitations on unit systems. And they are basically
twofold:
• They are extremely expensive to operate.
• They do not provide information about the profitability of stock investments made.
55
In a highly controlled experiment involving one
department in a department store, Joseph Buchan
and Ernest Koenigsberg demonstrated that a unit
control system can achieve better inventory
balance." Tlie difficulty with this and similar controlled demonstrations is that the full-cost impact of
widespread applications is often not taken into account. Increases in sales as well as turnover should
be weighed against the costs of system operation.
According to the data processing manager of a large
department store which is aggressively expanding
its unit systems using EDP, there are cost factors in
large-scale applications which do not stand out in
small-scale experiments. Receiving and marking
costs are examples. When items were added to the
automated unit control program, the need to record
and mark merchandise items by unit, as opposed to
department or classification, increased the flowthrough time by ten. Data processing costs are also
a good example. According to this same manager,
there are no economies of scale in unit systems. In
fact, his experience shows that computer processing
times for unit systems grow geometrically as the
number of stock-keeping units on the system is
increased.^
Even if accurate and relatively inexpensive unit
data were available, it would still provide only part
of the information needed. Unit information can be
very useful in planning and controlling assortments,
but by itself does not show profitability of merchandise investment because it does not attach a dollar
value to sales, reductions, cost of merchandise, or
average stock investment. This point is well made
in all leading textbooks on merchandising, but it is
worth re-emphasizing in view of current industry
performance trends and emphasis on unit systems.^
A Definition
In a sense, there is nothing startlingly new about
classification merchandising.
The preparation of dollar classification reports reqtiires exactly tlie same kinds of infoimation that are
presently used in the preparation of [traditional] department accounting records. These may be broadly classified as follows:"
1 / Sales and returns.
2 / Inventory: opening inventory; purchases and vendor returns; transfers; and additional markons and cancellations.
3 / Markdowns and cancellations.
56
4 / Open-to-buy: merchandise plans; and orders
placed.
The major difference between traditional dollar
control and the classification merchandising approach is the level of application. In fact, tliis very
point has caused confusion among traditional department and specialty stores.
Retailers themselves appear to be divided by the new
classification system these days. There are those that acclaim Classification Merchandising as today's—and tomorrow's—better way of learning customer demand, of
measuring merchandise performance, of planning sales,
of assigning inventory dollars, of space—and those who
claim we have had the likes of this for years.^°
This same problem was brought out in a study
performed by a leading firm of certified public
accounts.^^ Retailers apparently do not have a clear
idea of what is meant by a "classification" or of
what the objectives of "classification merchandising"
are. In the study quoted above, 50 department
stores were asked if dollar data on sales and stocks
were gathered below the department level. These
were the results:
• 32 stores indicated that they coflected at least sales
below the department level.
• 12 stores indicated that they developed stock plans
and open-to-buy controls below the department level.
• None of tlie stores was developing gross margin data
below the department level.
Typical of tlie stores included in this study is
Rich's, a large, progressive department store located
in Atlanta, Ceorgia. Rich's has about 1,400 merchandise breakdowns below the department level
(about 10 per department), and each breakdown
averages an inventory investment of $5,000.^- While
this may seem to be a fairly detailed breakdown of
inventory investment, it does not really approach
the level of analysis anticipated by proponents of
classification merchandising.
To talk in terms of numbers of inventory breakdowns, however, is to miss the point on classification merchandising. The development of classification merchandising basically reflects the demise of
the department as the primary responsibility and
profit center in department store retailing. As Edward Waterbury of Woodward & Lothrop recently
said:
The rapid expansion of our business in terms of the
number of stores, as well as volume and the remoteness
of the buyer from the ever-growing percentage of the
California Management Review
business, accelerated the demand for accurate and timely
information at the classification
In this sense, the merchandise class is becoming the
basic building block of merchandising.
Anticipating tliat "classes are the departments of
the immediate future . . ."" NRMA has developed
standard definitions for a four-level breakdown
based on increasingly specific merchandise characteristics.^'
First level: Broad groupings, most nearly similar to
historical store divisions; but designed as logical groupings of related merchandise for broad customer end-use.
Merchandise Groups.
Second level: Family groupings most nearly similar
to historical store departments; but designed to break
down merchandise by types, not by size groupings,
price ranges, etc. Merchandise Demand Genters.
Third level: Assortment of items within each secondlevel breakdown, on basis of relationship for customer
end-use. Classes.
Fourth level: Groupings on basis of items that are interchangeable or substitutable for specific customer
end-use. Gategories.
The NRMA classification scheme is basically a
four-digit coding system where the first level corresponds to the first digit. This coding system has the
capacity for 9 Merchandise Croups, 99 Merchandise Demand Centers, 999 Classes, and 9999 Categories. NRMA's immediate purpose in promulgating
standard definitions is to facilitate collection and
exchange of merchandising results at the class or
third level.
Unit VS. Dollar Systems
For the individual firm, whether or not it embraces NRMA's definitions, classification merchandising offers many benefits:
^ One is the abihty to measure profitability of narrower and more homogeneous groupings of merchandise.
This perspective on profltabihty allows management to
make better decisions on the allocation of the scarce resources of space, inventory dollars, and promotional expenditures.
• Management should be able to pinpoint expansion
opportunities and isolate problems or poor performance
areas.
I Merchandise plans and controls will be more meaningful at a class (or even a category) level because they
will be focused on a particular customer end-use. Purchasing for one class will not be closed ofF because another class is over budget, as often happens when plans
FALL / 1969 / V O L . X I I / N O . 1
and conbols are fonnulated for a more conglomerate
merchandise grouping, such as a department or dissection.
I Merchandising by class also facilitates planning and
controlling in the modem context of multiple locations
and boutique operations.
Usually unit systems and dollar systems are
thought to complement one another, although they
serve different purposes. Unit systems provide
movement and stock level information in terms of
physical pieces of merchandise, while dollar systems
provide similar information in financial terms. Traditionally, unit systems have been specifically oriented toward merchandising operations, whereas
dollar systems have been used to regulate total investment by major segments of the business. The
classification merchandising approach, however, extends the level of detail of dollar systems to tlie
point where they can also become useful in a merchandising sense. In some cases, fine-line dollar systems can even be substituted for unit systems. If
this is true, what might be the advantages of reallocating development funds and management emphasis from unit systems to fine-line dollar systems?
First, classification merchandising gives dollar information on sales, markups, reductions, margins,
and investments. This is critically needed on a detailed breakdown basis to deal with the adverse
profitability' trends in department stores and in other
sectors of retailing.
Second, even tliough proponents of unit systems
argue that unit information can be built up into dollar figures, the fact is that executives using unit data
are seldom able to bridge the gap. The problems involve system accuracy—accounting for reductions,
policing tlie entry of unit receipts, and even capturing unit sales data. Buyers also tend to be myopic
about a given merchandise item or stock-keeping
unit, and as a result they are less able to spot trends
across an entire assortment or group of related merchandise items. Furthermore, planning and controlline on the basis of an individual item can be very
hard because of the relatively \vide fluctuations of
item demand around its average value. Such fluctuations tend to be less chaotic for the entire range of
items in a customer end-use category when taken as
a whole.
Third, tlie expenses of establishing and maintainin o- a classification system are generally lower than
those of a unit system. Since most large retailers al-
57
jump in number of classifications was from 16 to
180. The median amount of inventory investment
per classification decreased from $5,000 to about
$600. Information is now being received monthly,
whereas less than half of the retailers enjoyed such
frequency before. More important, the information
is available 3 to 5 days after the end of the month,
as opposed to anywhere from 20 to 90 days using
manual methods. Taken together, such quantitative
and quahtative improvements in infomiation would
logically point to improved performance. But, in
spite of such strong logic, there are several a priori
reasons why improved inventory management perfomiance might not occur, especially during the
first year:
ready have some form of departmental dollar control, all the administrative procedures and information sources for a classification system already exist.
The new expenses of replacing departmental breakdowns with more comprehensive classification
breakdowns are rather marginal. Finally, no matter
how sophisticated the classification system, it is unlikely to exceed 10,000 categories, whereas for a
unit system, 10,000 stockkeeping units doesn't even
constitute a good start.
Certainly, good arguments can be developed for
classification merchandising, but tliey are meaningless unless they can be backed up with evidence of
concrete results. To date, the evidence consists of
isolated success stories, but then no systematic research into the effect of a classification merchandising program on profitability has been previously reported. To help fill this gap, the first-year results of
our continuing study of the impact of classification
merchandising on a sample of thirty smaller men's
wear retailers is presented below.
The inventory management report. An example
of the inventory management report, the heart of
the new information system, is shown in Table II.
This report permits the retailer to divide his inventory investment into as many as 799 classifications.
The report represents a quantum jump for all tlie
retailers studied in terms of sophistication, frequency, and timeliness of information. The median
• Inventories tend to increase initially as a result of
taking corrective action on imbalanced stocks.^°
• There is no history upon which to base planning
for the newly formed classifications.
• The basic tendency of small companies is to shy
away from planning.^^
Results and Evaluation
Results of improved information. Several important findings resulted from the analysis of first-year
data. Dramatic changes in inventory management
performance were observed. Table III summarizes
the first-year results of a comparison of observed in-
TABLE II. AN INVENTORY MANAGEMENT REPORT OF THE TYPE USED BY RETAILERS IN THE PANEL
% Sales
Sales
Total
Gross
Profit
Maint.
ft/1 Q 7.fTI Y\
ijXtXl ^lU
07
/C
Mos.
xnv. on ^
Hand
Additions
Cost
Retail
% Inv.
Mark
Downs
End Inv.
Cost
End Inv.
Retail
—
32.00
—
—
—
8,918.00
3,483.00
2,695.00
289.00
141.00
4,734.00
3,138.00
2,774.00
376.00
634.00
924.00
519.00
200.00
1,267.00
2,792.00
894.00
714.00
25.9
10.1
7.8
34,492.00
100.0
Total
Dept.
2,440.00
400.00
800.00
100.00
50.00
450.00
920.00
960.00
100.00
200.00
300.00
150.00
50.00
750.00
1,850.00
250.00
200.00
24.6
Totals 9,970.00
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
58
2.5
2.0
856.00
103.00
232.00
28.00
16.00
94.00
280.00
384.00
39.00
62.00
99.00
48.00
17.00
285.00
666.00
94.00
60.00
35.0
31.0
29.0
28.0
32.0
32.0
35.0
40.0
39.0
31.0
33.0
32.0
34.0
38.0
36.0
42.0
30.0
100.0
3,363.00
33.7
4.0
8.0
1.0
.5
4.5
9.2
9.6
1.0
2.0
3.0
1.5
.5
7.5
18.6
3.6
8.7
3.4
2.9
2.8
980.00
500.00
960.00
10.7
3.4
2.9
3.8
3.2
3.1
8.5
4.0
1.7
1.5
3.6
3.6
375.00
810.00
520.00
90.00
175.00
130.00
145.00
75.00
350.00
910.00
175.00
160.00
552.00
1,247.00
867.00
148.00
254.00
194.00
213.00
114.00
565.00
1,422.00
302.00
228.00
75.00
67.00
—
5,796.00
2,403.00
1,912.00
208.00
96.00
3,219.00
2,040.00
1,664.00
229.00
437.00
619.00
353.00
132.00
785.00
1,786.00
519.00
500.00
3.5
6,355.00
9,692.00
194.00
22,698.00
—
—
1,508.00
725.00
1,353.00
—
—
—
—
—
—
—
—
—
—
20.00
California Management
.8
.4
13.8
9.1
8.0
1.0
1.8
2.7
1.5
.6
3.7
8.1
2.6
2.1
Review
/entory investment versus predicted inventory investment, and observed gross margin dollars earned
I'ersus predicted gross margin dollars earned. The
predicted values of gross margin dollars earned and
inventory investment are based on a least squares
regression of each value on sales volume for the period 1961-1965. The differences between predicted
and actual value for each variable were converted
to standard units and partitioned to highlight the
degree of change. Changes of greater than one standard deviation in eitlier direction are considered to
be significant. For the panel as a whole, the firstyear results indicate strong improvement in gross
margin doUars earned and an equally strong increase, an unfavorable change, in inventory investment.
TABLE III. COMPARISON OF
FROM PREDICTED VALUES FOR
THE AMOUNT OF CHANGE
GROSS MARGIN DOLLARS
EARNED AND INVENTORY INVESTMENT,
PANEL RETAILERS,
1966
Number of Retailers
Gross Margin
Inventory
Dollars Earned
Investment
Greater than + 1 standard
deviation
Between + 1 and —1
standard deviation
Less than — 1 standard
deviation
5
17
11
11
15
2
SOURCES: Cyrus C. Wilson, "Explaining Management Performance Under Conditions of Improved Information in Smaller
Retail Firms"; and Charles D. Greenidge, "The Effect of Improved Information on Profitability in Smaller Retailing Firms";
both unpublished Ph.D. theses, Ohio State University, 1967.
Interpretation of results. The strong improvement
showTi by panel members on the gross margin-tosales ratio tends to support the idea that classification merchandising is an effective way to improve
profitability. However, panel results showed an unfavorable outcome for the sales-to-inventory ratio.
But there are strong indications that the increases
in this measure are transitory. Improved inventory
investment information on a fine-Hne basis is likely
to result in two kinds of management action:
1 / Increasing inventory on some classes to avoid lost
sales from being out of stock on briskly demanded merchandise.
2/Reducing inventory on some classes which are
presently overstocked in relation to demand.
FALL / 1969 / V O L . X I I / N O . 1
Understocked conditions can be quickly corrected
by placing orders for additional merchandise. Overstocked conditions, however, usually will not be
corrected immediately but will slowly correct themselves as normal sales reduce the excess stocks. The
net result is that inventory will tend to increase
immediately after improved information becomes
available, and better balance will be achieved only
over a period of time, as the excess merchandise is
sold off.
Transitory increases in inventory investment have
been observed in several merchandising information
systems. For example, Buchan and Koenigsberg
noted that inventory investment 'iDulged" immediately after system installation and did not come
back into balance with sales volume until six
months later.^^ Of course, their study involved a department in a large department store with brisk customer traffic, whereas our study covers smaller
stores with correspondingly lower traffic. Therefore,
while the comments of our panel retailers and the
results of other studies show that the unfavorable
change in the sales-to-inventory ratio is a temporary
phenomenon, it is very hard to say how long the
"bulge" will last. We might conclude, however, that
tlie ultimate result of improved information will be
a liigher salcs-to-inventory ratio and, hence, an
overall improvement in profitability.
Conclusions
The evaluation procedure presented earlier also
reflects the argument for greater emphasis on the
classification merchandising approach.
k The department store industry, specifically, and retailing, generally, are faced with a merchandising profitability problem.
I Present emphasis on unit systems for reasons of cost
and kind of Infonnation provided is not solving the
profitabilit}' problem.
^ Fine-line dollar systems have been proposed but
have not received a great deal of support. The executives
of large department and specialty stores seem to misunderstand the classification merchandising concept.
• Smaller specialty stores using a classification merchandising approach have shown strong improvement for
gross margin performance in the first year and can be
expected to show overall improvement for merchandising profitability in the future.
The experiences of panel retailers with fine-line
dollar systems offer some good insights for large re-
59
tailers. For example, the variety of classification
schemes used effectively by the panel retailers suggests that the NRMA standard classifications, while
necessary for figure exchange, may not be the most
fruitful metliod of classification for purposes of merchandise investment management. Several panel retailers did not conform to the general recommendation of classifying on the basis of end-use or detailed merchandise characteristics and employed
bases of price-line, vendor, season, or even color.
These retailers often had specific examples of how
classifying on a basis other than detailed merchandise characteristics had helped them eliminate a
price-line or achieve better control over a vendor.
From this we conclude that many alternative classification schemes might well be used below the third
level, altliough additional research needs to be
done, of course, to find the relative effectiveness of
alternative methods for various lines of merchandise.
NRMA seems, implicitly, to want to avoid classifying on any dimension other than merchandise
end-use or detailed characteristic. Part of the reason
could be a desire to avoid pushing dollar classification systems too deeply into the traditional province
of unit systems—an artificial limitation. As some of
the panel retailers indicated, too much information
can be a limitation to management. Because of the
large numbers of merchandise items, even in
smaller retail stores, managers cannot effectively
plan and control on an item basis. The strategy of
most retailers in the panel was to use fine-line dollar
information as a trigger mechanism. They reviewed
the dollar results of each class, looking for exceptional deviations from expectations. When they
spotted a class with results that seemed unusual,
they focused their attention on that class for further
study.
The next step might be a complete assortment review of the class tlirough a physical inventory to
determine the exact source of trouble. This, of
course, is nothing more than the management-byexception principle—but it's noteworthy that the
panel retailers felt better able to do this with
broader dollar data than witli more detailed unit
data.
For staple merchandise, fine-line dollar data allowed several panel retailers to retain financial control of a classification while delegating the actual
assortment merchandising to a supplier's representa60
tive. This technique is often used successfully in the
hosiery departments of large retail stores where the
supplier is actually operating the unit system under
the supervision of the department manager. Tliis is
another strong argument for classification merchandising as opposed to unit systems. It may very well
be that unit systems, or preoccupation with items, are
more realistically and economically a vendor function rather than a retailer function. Naturally, this
presupposes more formalized vendor-retailer cooperation in planning model assortments and seasonal
merchandising programs. Actually, there is already
a trend toward more tightly structured vendor-retailer relationships—as witnessed by the growing
popularity of programmed merchandising.
The emergence of the classification or category as
tlie basic level of merchandise planning and control
will also affect management practices. The present
departmental and subdepartmental breakdowns
generally have such a heterogeneous merchandise
composition that dollar measures, such as open-tobuy, are often criticized as misleading. Acceptable
performance for the department as a whole may
simply be an average of very good and very bad for
the classifications or categories which are grouped
together to form the department. There has been a
"credibility gap" about the usefulness of such information. But dollar data on a classification or categor)' basis should be more directional since it is
based on more homogeneous merchandise groupings. This should enhance the value of open-to-buy
controls and encourage the use of several other performance measures which, until now, have played
rather minor roles.
More attention will be given to return-on-investment measures on a classification basis to measure
profitability and productivity and to allocate limited
factors. Cross margin return on average inventory
investment is one example of the type of profitability measure that fine-line dollar information will
make more meaningful. Other measures would include: gross margin return on space utilized, margin after direct expenses return on inventory and
space, and return on promotional outlays. With all
these measures we need to develop standards of
performance for classes, and we need to educate
management to their use.
Class performance measures, new and old, should
ease the dilemma created by the breakdown of the
department as a profit center and the growing sepaCalifornia Management Review
ration of buying and selling into separate job responsibilities." In the future, merchandising will be
more and more a joint responsibility, and procedures should be adjusted to give the store operations executive a greater voice in merchandising
tliose classifications from which he is supposed to
sell. He should be able to work with the buying executive in developing the merchandise program for
his location. Without suggesting an adversary relationship, the store operations executive could treat
the buying executive as a "source of supply," and he
could then be held responsible not only for expenses but also for the profitable use of working
capital and space at his location. The operations
man would also need to be able to expand or contract the part of his budget which is allocated to a
given classification. Tlie buying executive, working
from his own budget, could then be held responsible for total classification performance across all locations.
chandise Gontrol," Chain Store Age, April 1962, pp.
E21-E23.
Tlirough such joint merchandising responsibility,
the store could establish a store operations profit
center for each location, and could simultaneously
create a classffication profit center oriented toward
a specific buying responsibility, regardless of selling
location.
In summary, there appear to be many potential
benefits for large retailers who adopt tlie classification merchandising concept. Although this system is
now widely misunderstood, it seems to be an effective means of reversing adverse profitability trends.
Most importantly, it appears to be consistent with
emerging patterns of change in the retailing industry.
10. George Baylis, "Are the Brakes Too Tight to Let
the Train Roll?" Retail Control, Oct. 1965, p. 18.
REFERENCES
1. R. Stanley Laing, "Tomorrow's Retail Systems,"
address to the Ninth Annual EDP Gonference for Retailers, Retail Research Institute, National Retail Merchants Association (NRMA), Sept. 26, 1967.
2. For example, see "EDP's Major Justification: Mer-
FALL / 1969 / VOL. XII / NO. 1
3. Standard
1967).
Classifications
(New
York:
NRMA,
4. Gontrollers' Gongress, Merchandising and Operating Results of Department and Specialty Stores in 1966
(New York: NRMA, 1967), p. 2.
5. Laing, "Tomorrow's Retail Systems."
6. Joseph Buchan and Ernest Koenigsberg, Scientific
Inventory Management (Englewood Gliffs, N.J.: Prentice-Hall", 1963), p. 75
7. Data in this paragraph are provided by the data
processing manager of a leading department store. The
name of the individual and his finn are withheld by request.
8. For example, see William R. Davidson and Alton
F. Doody, Retailing Management (New York: The
Ronald Press Gompany, 1966), pp. 404-405.
9. Albert 1. Schott as quoted in Putting Classification
Merchandising to Work (New York: NRMA, 1967), p.
20.
11. GarroU E. Ebert, "Merchandise Glassification,"
ibid., Feb. 1966, p. 23.
12. Authors' conversation with John Bradley, Gonti^oller of Rich's.
13. Edward S. Waterbury, "Gathering Glassification
Data by Manual and Mechanical Systems," Retail Control, Feb. 1966, p. 20.
14. William Buiston in Standard Classifications, p.
31.
15. Putting Classification Merchandising to Work, p.
65.
16. Buchan and Koenigsburg, p. 75.
17. See Roger A. Golde, "Practical Planning for
Small Business," Harvard Business Review, XLII:6
(Nov.-Dec. 1964), 148
18. Buchan and Koenigsburg, p. 75.
19. Wheelock H. Bingham and David L. Yunich,
"Retail Reorganization," Harvard Business Review,
XUUA (July-August 1965), 129-146.
61
