1. STUDYING CUSTOMER BEHAVIOR IN
RETAIL STORES
WILLIAM APPLEBAUM
Stop 6? Shop, Inc.
TUDIES of customer behavior in re- is influenced by the needs and prefer-
S tail stores usually deal with (i) ences of the consumers for whom the
identification of customers and (2) their products are purchased. As a dog owner
buying behavior patterns. The aim of 1 can testify that our Cocker influences
such studies is to ascertain who buys her purchasing agents in the choice of
where, what, when and how. In addition, brands of dog food. A devoted house-
such studies endeavor to learn about wife will buy for her husband his favor-
customer response to sales promotion de-ite, extra-strong cheese, even though she
vices. The results of these studies are prefers not to subject her sensitive nos-
useful in the solution of an array of trils to the penetrating, if not over-
marketing problems. powering, bouquet.
These studies are gaining importance Here, however, we are primarily con-
in marketing research. It is safe to pre- cerned with customers' buying behavior
dict that interest in them will increase rather than with consumer preferences.
greatly in the next decade. This paper The strict definition of customer as pur-
summarizes the writer's experience with chaser, valid as it may be, is sometimes
such studies in grocery stores.^ However, impractical because it is not always sim-
the principles and techniques discussed ple to identify a purchaser in a store.
here are also applicable to other types of A husband and wife shopping together
retail stores. may represent two customers, or one
The "why" of customer behavior is a customer and one bundle carrier, and it
separate and very difficult subject; it is is not always possible to tell which is
not treated here. A knowledge of cus- which. A boy accompanying his mother
tomer behavior must precede any con- may influence decisions in selection and
sideration of the reasons for the behavior. purchase of several items, even if the
youngster has no purchasing power of
IDENTIFICATION OF CUSTOMERS his own. Therefore practically every
The terms customer and consumer person who enters a store is a potential
are not synonymous. A customer is a purchaser and represents a unit in the
purchaser of a product or a service; a store's customer traffic.
consumer is a user of a product or a Identification of customers seeks to
service. Bed-ridden invalids in hospitals ascertain who the customers are. It is not
are food consumers but hardly food sufficient to study buying behavior pat-
store customers. Purchasers of rat poison terns without knowing whose buying
are not the consumers of the product. behavior is involved. Hence it is neces-
The buying behavior of the customer sary to identify the competition and
origin of customers.
^ In so far as the author himself has developed or le-
fined the ideas and techniques dealt with in this paper, Composition of Customers
by far the major part of the pioneer research work was
dene between 1932-38 vriiile IK wai in du: employ a(Tht This includes many characteristics,
Kroger Company. 8uch as sex, age group, color, economic
172

2. STUDYING CUSTOMER BEHAVIOR IN RETAIL STORES 173
and educational status, occupation, re- Similarly a distinction should be made
ligion, nationality origin, and so on. Both between buying habits and buying be-
consumption and buying behavior are havior patterns. Habit is a tendency
affected by these characteristics; and toward an action which by repetition
the relative significance of each of these has become spontaneous. A pattern is a
characteristics varies greatly, depending design or type. Each customer has his
on the nature of the problem. The extent or her own buying habits. Buying be-
to which any one or all of these customer havior patterns represent the design of
characteristics should be studied can be behavior of a large number of customers.
determined only on consideration of the A run on stores to buy and hoard sugar,
purpose for which the data are to be nylon stockings or toilet paper—in re-
used. In studying customer behavior in sponse to a shortage scare or to proposed
retail stores it is generally impractical rationing—is not a buying habit; it is a
or unnecessary to ascertain all of the manifestation of the imperfections in our
customer characteristics enumerated education, in our faith, and perhaps even
here. in our frequently exalted way of life.
To the retailer, such panicky customer
Origin of Customers behavior is very annoying.
Customer buying habits or behavior
Where do a store's customers come patterns are not permanently fixed, and
from? What is the geographic distribu- certainly not sacred, even though some
tion of their homes and how far do they habits tenaciously resist change. Many
travel to the store ? The answers to these factors are operating in combination to
questions supply useful data on customer change customer food-buying behavior
origin, which in the sense used here is a patterns. Among these are the automo-
composition characteristic. Origin tells bile, the super market and self-service,
us whether the customer is a large-city the progress in the development and
apartment dweller, a suburbanite, a merchandising of frozen foods, prepared
ruralite, a transient, and so on. The food- flour mixes, brown-and-serve baked
buying behavior pattern of a customer goods and concentrated fresh milk; the
who lives in a congested apartment area increasing availability of suitable facili-
of a large city is markedly different from ties in the customer's home for preserv-
the rural customer who has a large ing these and other highly perishable
vegetable garden and a home freezer. raw and prepared foods; and the public's
receptive disposition to easier and less
CUSTOMER BUYING BEHAVIOR time-consuming ways of living. Similar
PATTERNS and perhaps even more pronounced
To buy is to purchase. To shop is to changes are affecting customer buying
visit business establishments for inspec- behavior patterns of non-food commodi-
tion or purchase of goods. Therefore ties.
shopping is an element of customer be- Customer buying behavior patterns
havior in buying. A customer placing an can be grouped in relation to:
order over the telephone is buying, not 1. Place of Purchase
shopping. For this reason it may be 2. Items Purchased
desirable to standardize on the use of the 3. Time and Frequency of Purchase
term buying rather than shopping when 4. Method of Purchase
the totality of customer behavior is un- 5. Response to Sales Promotion De-
der coQsideration. vices

3. 174 THE JOURNAL OF MARKETING
Place of Purchase chased by a relatively small group of
In general, customers divide their pur- customers. The purchase of some items
chases among a number of stores. They has a strong regional demarcation—
shop in more than one department store grits or live lobster, for example.
and in many specialty stores. Even in Rarely does a customer purchase a
buying food there is a division of pur- single potato or a single carrot. On the
chases. Many customers do not buy their other hand, very seldom does a customer
meats or fresh fruits and vegetables purchase more than one watermelon at
where they buy their dry groceries, al- a time. The amount of each item pur-
though all these goods may be available chased depends on many factors, of
in the same store. The specialty food which the following are probably the
store, the milkman, the produce huck- most important: number of consumers
ster, and the "hole-in-the-wall" are all for whom the item is intended; perish-
getting a share of the total food business. ability of the item; storage requirements
Where customers have the choice of and facilities available; purchasing power
purchasing the same goods in a number and ready cash; unit of sale; and price.
of stores, their patronage loyalty to any The introduction of new products and
one store is by no means permanent. changes in dietary habits also affect the
Witness the grand opening of a new su- customer's choice of items and the
per market! Many of the customers who amount purchased.
flock to the opening are abandoning old As previously pointed out, in so far as
patronage loyalties. Hence, length of the customer is the purchasing agent for
patronage also deserves study. a family or a number of consumers, the
Studies of customer buying behavior purchases reflect the characteristics of
patterns with respect to place of pur- all the consumers involved.
chase are useful in selecting store loca- From the distributor's, manufacturer's
tions, in choosing distributors for a prod-and producer's point o? view it is es-
uct, and in merchandising. sential to study what items and how
much of each the customer buys by
Items Purchased brand or quality, by size or weight, by
Every customer purchase and every price, by type of container, and by sea-
store sale consists of a transfer of one son.
or more specific commodities. No one
customer purchases all the different items Time and Frequency of Purchase
for sale in a store. Over a period of time Store operations must be geared to
a customer will purchase a substantial mesh with the customers' time of pur-
selection of the total items available in chase pattern. Store buyers and mer-
the store, but that selection will vary chandisers must keep on schedule with
somewhat with each customer. There- it. Merchandise must be available in the
fore, in studying customer buying be- store in adequate supply if maximum
havior patterns it is necessary to ascer- sales are to be achieved. Woe to him
tain (a) what items and (b) how much whose Christmas trees arrive to market
of each item customers purchase. on December twenty-sixth!
The purchase of many necessity items In studying time of purchase patterns
—soap or bread, for example—is com- it is necessary to relate these to the sea-
mon and ubiquitous. Luxury items, such sons, weather, and regional differences.
as caviar or avocado pears, are pur- Every retailer knows from experience

4. STUDYING CUSTOMER BEHAVIOR IN RETAIL STORES 175
that his volume of sales is not uniform basis, whether a customer shops alone
by days of the week, nor by hours of the or is accompanied by someone else, and
day. The variations are very pronounced, whether a customer walks or rides to
especially in the food business. the store are some of the elements in
Attempts^ by retailers to modify cus- method of purchase.
tomer time of purchase behavior pat- The importance of providing ade-
terns, with the view of improving service quate parking facilities to accommo-
to customers or raising efficiency of date the customer who shops or who
operation, have by no means been en- would like to shop by automobile needs
tirely successful. The long lines of cus- no elaboration. Even department stores
tomers waiting impatiently to be checked are building branch units to meet,
out in super markets during peak periods among other things, the parking prob-
in contrast to the buying inactivity at lem. Size and frequency of purchase in
other times illustrate a continuing grocery stores are definitely affected by
troublesome store operations problem the mode of travel to make the pur-
created by the customers' time of pur- chase.
chase pattern. Couples shopping in super markets
Frequency of purchase depends pri- buy more per transaction than does a
marily upon the type of commodity in- woman or man shopping alone. The same
volved. In the course of a lifetime, a man applies in other type stores. Many a
rarely purchases more than two wedding woman will invite her husband to help
rings—most men try to manage with the her select a purchase; the husband in-
original acquisition, even if it does not fluences the choice and frequently ap-
turn out to be quite the bargain ex- proves the acquisition of a more expen-
pected. An automobile—which in Ameri- sive item.
can life has become indispensable to the
pursuit of happiness—is turned in for a Response to Sales Promotion Devices
new model every few years. On the other Those who have goods to sell use many
hand, the addicted pleasures of ciga- devices to induce consumption and to
rettes go up in smoke so rapidly that the
promote purchase of these goods. The
frequency of purchase is generally a daily
sales promotional devices used in stores
performance and, allegedly, men will can be grouped under the following
walk a mile for their favorite brand. headings:
Frequency of purchase also varies a. Displays
among customers. Some shop in food b. Pricing
stores daily, others only once a week. c. Demonstrations
The size of the total purchase, the num- d. Sales Talks
ber of items and the quantity of each
item bought all vary with frequency of Displays. In so far as possible or prac-
purchase. The more frequently a cus- tical, consumer goods are packaged to
tomer visits a store, the more is thatcreate eye appeal when put on display.
Thus the shape, size, label and packag-
customer exposed to the impact of sales •
promotional devices used in the store.ing material of the product all play their
part in sales promotion. The manner
Method of Purchase in which an item is arranged on a shelf
Whether a customer buys on a cash or table, in a display case or window,
and carry or on a charge and delivery the space and position given to it, and

5. 176 THE JOURNAL OF MARKETING
the type of point of sale promotional these suffer from too many defects in
material (posters, signs, etc.) used for research methodology) to be more than
reinforcement are also influencing fac- merely suggestive, preliminary intro-
tors. ductions to this subject. This, however,
Pricing. Regular and bargain prices, does not hamper the courage,of the eager
combination deals, coupons, prizes, con- users of these published data on impulse
tests, and unit pricing—where the price buying in their advertising and sales
quoted is for two or more units of an talks. Perhaps it is another manifesta-
item—are all pricing devices to promote tion of "where to know little is to dare
sales. easily.
Demonstrations. This device aims to RESEARCH TECHNIQUES
influence customer purchases by getting
them to sample a product or to learn The marketing research techniques
about the uses or other merits of the used in studying customer behavior in
product. retail stores are:
Sales Talks. Whether expressed orally 1. Analysis of Records
or in writing, in advertisements or by a 2. Observation
sales clerk, the aim of all sales tdks 3. Interviewing
needs no explanation. Self-service mer- 4. Controlled Experimentation
chandising has somewhat shaken the
retailer's faith in the efficacy of the store Analysis of Records
sales clerk. Yet, the magic of a winning In a well-run business records are
personality, a well-turned phrase, and kept on many operating and merchan-
even a friendly smile of an attractive dising results, but few businesses—and
blonde are still potent attributes or de- this applies even to the most progres-
vices for causing cash registers to ring. sive—come anywhere near making maxi-
Study of customer behavior in re- mum use of these records. Many a
sponse to sales promotion devices in sales manager might be inspired with the
retail stores deserves a great deal more by-products which a skilled analyst
attention than has been given to it could distil from these records.
thus far. Much energy and money is Records of what was sold, when it was
spent on sales promotional devices with- sold, and at what price it was sold in
out factual knowledge of what these a store or group of stores are a general-
really accomplish. In this terra incognita ized history of customer buying be-
every merchandiser remains at liberty havior patterns. Stores with credit and
to indulge in his personal pet beliefs, delivery and with itemized individual
prejudices, and hunches. Where there sales transaction records are in posses-
are no faf ts, there is much conjecture. sion of data from which a wide array
In recent years considerable attention of customer behavior patterns can be
has been focused on impulse buying— determined.
buying which presumably was not The more complete and accurate the
planned by the customer before entering records are and the longer the period
a store, but which resulted from a stimu- they cover, the more valuable they can
lus created by a sales promotional de- be. Commodity movement records
vice in the store. The surveys on im- should show units sold by brand, size,
pulse buying in food stores which have style of package, and price. Dollar sales
come to my attention are too few (and data are desirable by departments, by

6. STUDYING CUSTOMER BEHAVIOR IN RETAIL STORES 177
weeks, by days of the week and by peri- All observed data need to be recorded.
ods of the day. Customer sales transac- Such records should be made in code
tion records are desirable by depart- whenever possible. The recording can
ments, amount of purchase, and (if ob- be done in a notebook, on a suitable
tainable) by items purchased. form, or on a specially designed mark-
Most cash registers can be fixed to sense card. The data are processed manu-
retain an exact duplicate copy of every ally or mechanically.
sales receipt. Such "register tapes"—
showing the total amount of the pur- Interviewing
chase and the price of each item by de- Some information on customer char-
partments—are useful in analyzing cus- acteristics and buying behavior is most
tomer purchases in super markets, even readily obtainable by interviewing cus-
though only the price but not the name tomers in the store. Questions on mode
of an item is listed on the sales receipt. of travel to the store, frequency of visits
Where the desired store sales records and length of patronage are easily asked
do not exist, it is necessary to set up and readily answered. Many other simi-
procedures to obtain them. The research- lar direct and simple questions can be
er competent to analyze such data will handled successfully in store interviews.
know how to procure the records. The best time to interrogate store cus-
tomers is at the conclusion of the pur-
Observation chase. Such interviews consume very
The sex, color and age grouping of a little time and are inexpensive. Not to
customer can be easily observed either spoil a good thing, however, one should
as the customer enters, leaves, or shops avoid asking too many questions at one
in the store. By observation it is also time.
possible to learn what a specific customer When asked tactfully, a customer will
does in the store. A very wide range of rarely refuse to give his or her home ad-
customer activities are observable with- dress. This important piece of informa-
out the customer knowing that such tion is essential to determine the cus-
observations are being made. This is all tomer's geographic origin and the trad-
to the good, as nothing should be done ing area of a store.* From this, and with
to interfere with "normal" customer be- the aid of other available data (census,
havior. In some studies it is necessary town planning, etc.), it is frequently
to follow the customer in the store in possible to ascertain, approximate, or
order to observe various activities, but surmise various other characteristics,
this is generally a troublesome and costly such as income grouping, cultural levels,
procedure. The checkout stand is the and nationality origins.
meet s t r a t ^ c spot in a super market On occasion it is desirable to follow
for observing the quantity of specific up a store interview with another inter-
items purchased. view in the customer's home, with or
Unless the inv^tigator has had ex- without disclosing for whom the infor-
tensive experience with such customer mation is being obtained. The cus-
observation studies in stores, it is neces- tomer's address is indispensable for such
sary to formulate procedures carefully, a follow-up interview; nam^ are neces-
to pre-test them, and to make sure that ' For further details set William Applebaum and
the observers are thoroughly trained for Richard F. Spears, "How to Measme a Trading Area,"
the work. Chain Store Age, January, 1951.

7. 178 THE JOURNAL OF MARKETING
sary where customers live in multi- group one condition is changed. The
family dwelling units. changed condition is the element or
• Studies of impulse buying depend on variable to be tested. This element can
customer interviewing in the store, sup- be a product, a method, or a sales promo-
plemented by observation. The customer tion device. The element is exposed to
is interviewed on entering the store and the store customers and it is anticipated
again after the completion of the pur- that their buying behavior will be af-
chase. The investigator ascertains fected by it. Results are measured in
whether the customer has a written terms o^ sales produced—not on the
f
shopping list or a mental list. The items basis of opinions. Differences in results
enumerated on such lists are recorded. between the test and control store
On completion of the purchase, the items groups are interpreted to be due to the
purchased are checked against the listed one variable in the experiment.
items. Omissions, substitutions and ad- Where a problem involves more than
ditions are noted. one element, it is necessary to select
The trouble with this technique is several comparable store groups and
that (i) it depends on spontaneous recall conduct a number of tests simultaneous-
which at best is incomplete and (2) it is ly, or to run, in succession, a series of
unable to differentiate between (a) im- experiments in two groups of stores,
pulse, (b) postponement of decision, and each experiment testing only one ele-
(c) studied decision at the point of pur- ment.
chase. The housewife who has no shop- The technique of controlled experi-
ping list (written or mental) may still mentation is very exacting. There can
be deliberate in the choice of the items be no controlled experimentation with-
which she does purchase A more satis- out stringent controls. It is a compli-
factory technique remains to be de- cated, slow and expensive technique,
vised for the study of impulse buying. yet one which merits a prominent place
in the marketing researcher's tool kit.
Controlled Experimentation^
CONCLUSION
Stores offer a unique laboratory for
conducting controlled experiments to The endeavor has been made here to
ascertain the behavior and response of show along what lines and with what
customers to products, methods and techniques studies of customer behavior
devices. From the facts discovered by can be made in retail stores. Possible
such small scale experiments it is possible uses of the results for the solution of
to reach broad conclusions. marketing problems have been suggested
In controlled experimentation two (or or indicated. To spell out fully and sys-
more) groups of comparable stores are tematically the many practical uses to
selected. One of these groups is desig- which such studies can be put by pro-
nated as the control group and the other ducers, manufacturers and distributors
as the test group. In the control group would require a series of articles—per-
"business as usual" goes on. In the test haps a book. This article is intended to
stimulate those who can gain the most
' For a fiill presentation of this technique see William from such studies to take advantage of
Applebaum and Richard F. Spears, "Controlled Ex- the opportunities for learning more about
perimentation in Marketing Research," THE JOURNAL
a, January, 1950. customer behavior in the market-place.