The Use of Advertising During Depression

Roland S. Vaile

A common policy among manufacturers has been to gauge the amount of their magazine advertising by general business conditions; advertising space has been increased during prosperity and decreased during depression. Professor Crum1 found that the major fluctuations in the volume of advertising carried by magazines since the war have taken place slightly in advance of similar movements in general business conditions. The figures which he published also showed that the fluctuations in advertising were somewhat wider than in general sales.

There are numerous departures from this general policy, however, in the case of individual firms. These departures are conspicuous, not alone because of the magnitude of the variations, but equally because of the outstanding character of some of the firms. There are many cases where institutions which appear to be directly competing have evidently used markedly different technique in this part of their sales promotion.

The figures in Table 1 are illustrative of these differences in policy. They are adapted from the monograph of the Crowell Publishing Company, National Markets and National Advertising, 1925.

TABLE 1: CHANGES IN MAGAZINE ADVERTISING, 19201924
(EXPENDITURES FOR 1920 = 100)

Firms Actual
expenditure
1920
Relative Expenditures
1920 1921 1922 1923 1924
Clothing:
Hart, Schaffner & Marx $ 188,500 100 150 155 165 150
Kuppenheimer 240,718 100 50 65 65 45
Foodstuffs:
Postum cereal Company 375,615 100 120 160 135 330
Quaker Oats 1,390,108 100 40 40 45 40
Soap:
Palmolive 377,150 100 125 150 185 190
Resinol 424,197 100 60 55 70 40
Paint:
Valentine & Company 184,270 100 120 150 225 320
SherwinWilliams Company 201,815 100 55 50 60 115

The changes in expenditure for magazine advertising during the fiveyear period, 19201924, are contrasted for individual firms in similar lines of business. The money expended by each firm during 1920 is taken as 100, because Professor Crums study showed  that magazine advertising reached a peak in that year, to be followed by marked slump. The expenditures for subsequent years are reported as percentages of this year. In each case the two firms contrasted are engaged in similar lines of business and they have been, in general, heavy users of magazine advertising.

Many similar differences appear in almost every line of goods. The illustrations given seem typical and could be multiplied many times over.

Some of the differences portrayed are doubtless accounted for by changes in policy between the use of magazine advertising and other forms of sales promotion, although in several cases personal correspondence with the sales manager of the company indicates that the movement of magazine advertising fairly represents the general sales policy at the time of the 19201922 depression. Some of the differences may also be due to a willingness in 1919 and 1920 to spend money extravagantly for advertising rather than pay high corporation taxes. After all allowances are made, however, it seems that a real difference in institutional policy exists in regard to sales promotion during times of depression; some concerns have apparently assumed that advertising is one of the costs which may be reduced during depression, while others have looked upon it as a force for making sales when they are hard to make. Both of these points of view have frequently been defended in articles appearing in Printers Ink and elsewhere, largely on a priori grounds. Little factual evidence has been presented in support of either side.

Two questions are suggested by these differences in policy:

  1. What is the effect of each policy upon sales?

  2. Is the effect the same in the case of all commodities? 

In order to partially test these questions, the writer has compiled sales figures for the years 1920 to 1924, inclusive, for about 250 well known firms. A large proportion of these figures were obtained from Poor's Manual of Industrials 1926. These were supplemented, however, by direct correspondence with the companies in cases where Poor does not record sales. In a number of cases the request was made that the sales of individual firms be held as confidential, so that no report is made in this paper of sales by firms.

In approaching the first question, the sales records of the firms were sorted into three groups, namely:

  1. Firms which did no magazine advertising from 1920 to 1924, inclusive.
  2. Firms which increased their magazine advertising from 1920 to 1924.
  3. Firms which decreased their magazine advertising from 1920 to 1924.

Some firms for which sales records were available, were so inconsistent in their advertising policy during these years that they could not be classed in any of the above groups. Such firms were not included in the study, only those with marked trends being retained.

The average (arithmetic mean) sales for the different groups in 1920 were as follows:

  • Nonadvertising firms        $23,000,000
  • Increased advertising        $25,000,000
  • Decreased advertising      $21,000,000

All sizes were represented in each group, with a range from $1,000,000 to nearly $1,000,000,000 in each case. In the same year, 1920, the average expenditure for magazine advertising was $170,000 per firm, both in the group which subsequently increased and in the one which decreased their expenditures. This amount was 0.68 per cent of sales in the first group and 0.81 per cent in the second.

Because of the differences in size of individual firms, the annual dollar sales of each were reduced to relatives. The sales for 1920 were arbitrarily taken as a base in all cases, because it was the first year for which records were available for many of the firms, and because it was generally a year of high sales. Relatives were used rather than dollar sales so that the movement in any one firm might have equal weight with the movement in any other. The writer believes that this gives a more accurate picture of the relation between advertising and sales than would a study of total sales, wherein the movements in the large firms might entirely obscure the movements in the smaller firms. The question may still be raised whether advertising during depression is effective with firms of one size than with those of another; the present paper makes no attempt to answer this question.

Table 2 shows some of the characteristics of each group, together with the relative sales associated with each of the magazineadvertising policies.

TABLE 2: CHARACTERISTICS OF EACH GROUP, WITH RELATIVE SALES ASSOCIATED WITH EACH OF THE MAGAZINEADVERTISING POLICIES

Year Firms not
advertising in
magazines
Firms which increased advertising in magazines Firms which decreased advertising in magazines
1920 1924 1920 1923 1920 1923
Number of firms 105 58 67
Average sales 1920 $23,000,000 $25,000,000 $21,000,000
Relative sales {1920 100 100 100
{1921 80 88 74
{1922 92 107 88
{1923 105 128 103
{1924 108 131 105
Average expenditure for magazine advertising 1920 $170,000 $170,000
Relative magazine advertising {1920 100 100
{1921 138 75
{1922 157 50
{1923 195 54
{1924 242 58
Ratio of magazine advertising to sales {1920 0.68% 0.81%
{1921 1.04 0.81
{1922 1.00 0.46
{1923 1.03 0.43
{1924 1.25 0.45

The sales of the nonadvertising 20 per cent lower in 1921 than in 1920. In contrast, the slump was only 12 per cent for the firms which increased their magazine advertising, while it was 26 per cent for the firms which decreased their advertising expenditures. follows: All three groups showed increasing sales from 1921 to 1924, but the increase was most pronounced in the group which increased their advertising, and least pronounced in the nonadvertising group.

While the firms which increased their advertising also increased their sales, they did not do so proportionally. Advertising became relatively more costly in this group, while it became relatively less costly in the group which decreased advertising. Just what effect these changes may have had on net profits has not been determined; it seems impracticable to study the movements of net profits without a more intimate knowledge of the accounting methods of individual firms than is given in the published reports.

In order to make the contrasts clearer, an attempt was made to eliminate both secular growth and cyclical variation, which might affect all concerns alike regardless of their advertising policies. Because of the short period of time involved, no attempt was made to fit curves which might serve as base lines. Rather, the relative sales of the nonadvertising group were taken as the base and counted as 100 for each year, on the assumption that the sales of these firms would give a fair picture of the combined effect of secular and cyclical variations unaffected by the force of advertising. Table 3 shows the resulting figures. For 1920, all the figures are obviously at 100. The subsequent relatives are measures of the sales associated with changes in advertising policy as contrasted to sales of nonadvertising firms. These figures are derived, of course, directly from the preceding table.

TABLE 3: RELATIVE MOVEMENT OF SALES ASSOCIATED WITH DIFFERENT ADVERTISING POLICIES

1920 1921 1922 1923 1924
Increased advertising 100 110 116 121 121
No advertising 100 100 100 100 100
Decreased advertising 100 95 96 98 97

Two facts seem to stand out from these data with considerable clarity: first, a definite spread occurs between the sales of firms which increased their advertising and those which decreased it. Where intensive advertising during depression was a part of the sales technique, sales were maintained in better volume than where advertising appropriations were cut. Second, the nonadvertising firms fared better, especially in 1921, in volume of sales than did those firms which reduced their advertising. Advertising is a form of sales effort, and its reduction at a time of depression seems to have resulted in a greater loss of sales than was experienced by the firms which constantly depended on other types of distribution. This suggests that personal selling to the trade and dealer goodwill may have more lasting importance at such a time than has consumer recognition.

As a further test of the effect of increasing and decreasing advertising during depression, the relative changes of sales of individual concerns from 1920 to 1921 were correlated with the corresponding changes in advertising. Because of the greater range of change in advertising than in sales, a standard deviation was determined for each series, and each item was divided by the deviation of its series. The Pearsonian coefficient for the resulting figures was .689. From this it appears that sales of individual firms may often be materially affected by advertising, even during depression.

Turning now to the second question suggested above: Is the effect the same in the case of all commodities? The firms for which data are available are divided about equally among seven industries, namely:

  1. personal items, such as tobacco, perfumes, and toilet articles; 

  2. house furnishings, from kitchenware to pianos; 

  3. clothing, including mens and womens suits, hats and coats; 

  4. automobile equipment frequently purchased directly by consumers, such as tyres, shock absorbers, chains; 

  5. automobiles, of which the Ford was the only nonadvertising representative; 

  6. groceries; and 

  7. building materials.

The sales figures used throughout this study are the total annual sales in dollars uncorrected for changes in price level. One of the advantages, therefore, of dividing the data into industrial classes is the fact that within these classes price changes are likely to affect one firm much as they do another.

The same methods of comparison were applied to the data for the several industries as were used in analysing the changes in sales for all the firms together. Tables 4, 5 and 6 present the summarized data by industries giving, first, the relative advertising expenditures both for the group which increased and the one which decreased; second, the growth in sales for each of the advertising groups; and, third, the relative movement of sales of the advertising groups as contrasted to the nonadvertising firms.

TABLE 4: RELATIVE ADVERTISING EXPENDITURES

Firms which increased,
1920 1923
Firms which decreased,
1920 1923
Industry
1920 1921 1922 1923 1924 1920 1921 1922 1923 1924
Personal items 100 115 185 180 220 100 60 50 30 35
House furnishings 100 130 160 250 275 100 70 50 70 80
Clothing 100 180 170 290 360 100 70 35 60 50
Automobile equipment 100 125 150 175 280 100 75 50 60 75
Automobiles 100 145 145 185 250 100 75 25 35 60
Groceries 100 115 130 120 135 100 75 90 75 70
Building materials 100 125 160 165 170 100 95 40 50 35

TABLE 5: GROWTH OF SALES ASSOCIATED WITH DIFFERENT ADVERTISING POLICIES

Industry Number
of firms
Dollar sales
1920
Relative sales
1920 1921 1922 1923 1924
Personal items:
Increased advertising 10 $34,000,000 100 93 107 112 120
No advertising 12 20,500,000 100 80 83 90 95
Decreased advertising 10 22,000,000 100 73 82 90 93
House furnishings:
Increased advertising 7 19,200,000 100 83 105 120 108
No advertising 11 8,540,000 100 70 84 104 95
Decreased advertising 7 13,000,000 100 71 86 101 90
Clothing:
Increased advertising 8 16,500,000 100 102 112 118 106
No advertising 20 18,800,000 100 86 96 100 96
Decreased advertising 9 13,600,000 100 77 74 73 70
Automobile equipment:
Increased advertising 7 56,000,000 100 77 96 117 120
No advertising 11 58,000,000 100 62 86 107 110
Decreased advertising 14 28,000,000 100 72 80 101 100
Automobiles:
Increased advertising 8 100 78 137 184 177
No advertising 1 100 95 126 188 187
Decreased advertising 9 100 73 124 159 165
Groceries:
Increased advertising 10 22,000,000 100 87 96 103 120
No advertising 38 21,000,000 100 85 100 103 115
Decreased advertising 10 24,000,000 100 77 87 94 102
Building materials:
Increased advertising 9 14,000,000 100 81 90 120 125
No advertising 12 13,500,000 100 75 87 123 117
Decreased advertising 8 15,000,000 100 78 92 118 120

TABLE 6: RELATIVE MOVEMENT OF SALES ASSOCIATED WITH DIFFERENT ADVERTISING POLICIES

Points spread between firms which increased or decreased advertising
Industry
1920 1921 1922 1923 1924 1921 1922 1923 1924
Personal items:
Increased advertising 100 116 129 125 127 25 30 25 28
No advertising 100 100 100 100 100
Decreased advertising 100 91 99 100 99
House furnishings:
Increased advertising 100 118 117 118 111 28 30 35 28
No advertising 100 100 100 100 100
Decreased advertising 100 90 77 73 73
Clothing:
Increased advertising 100 118 125 115 115 16 23 18 20
No advertising 100 100 100 100 100
Decreased advertising 100 102 102 97 95
Automobile equipment:
Increased advertising 100 115 112 109 109 8 19 15 18
No advertising 100 100 100 100 100
Decreased advertising 100 107 93 94 91
Automobiles:
Increased advertising 100 80 109 98 95 3 11 13 7
No advertising 100 100 100 100 100
Decreased advertising 100 77 98 85 88
Groceries:
Increased advertising 100 102 96 100 104 12 9 9 15
No advertising 100 100 100 100 100
Decreased advertising 100 90 87 91 89
Building materials:
Increased advertising 100 108 103 98 107 4 2 2 5
No advertising 100 100 100 100 100
Decreased advertising 100 104 105 96 102

It will be noted from the above data that in every industry there is a measurable spread between the relative sales of the firms which increased their advertising and of those which decreased. Moreover, in every case except that of building materials, the firms which decreased their advertising suffered a greater reduction of sales than did the firms which were doing no advertising. Thus the general relationship between advertising and sales seems to be consistent in all industries, but there are differences in degree which are worthy of comment.

The spread between the sales of firms which increased their advertising and the sales of those which decreased is much more pronounced in some industries than in others, as shown in Figure 1. It is particularly large in the cases of personal items and clothing, and fairly large with automobile equipment and house furnishings, while it is small in the cases of automobiles and groceries, and almost negligible with building materials. This would indicate that magazine advertising is a more potent force in the sale of personal items and clothing than it is in the sale of the other lines. This is further borne out by the fact that the sales of groceries and building materials were only slightly increased by increased advertising, when compared with the sales of the nonadvertising firms in these two fields.

The data of this study of advertising force in various industries seem in harmony with other knowledge of consumers buying habits. Personal items and clothing, as included in this study, are purchased largely by and for men, as are also automobile equipment items. Men are said to be especially susceptible to brand advertising because they do not care to shop and because they are willing to accept the same article time after time. Moreover, there are several instances within these groups where the use is a habitual one such as smoking and in these cases the brand name may become associated with the habit.

Groceries, on the other hand, are often bought on the recommendation of the retailer. George Louis, writing in System, the Magazine of Business, for June, 1912, reported a study in which it was found that unknown brands of foodstuffs could be substituted for advertised brands with little opposition on the part of the consumer. Groceries are convenience goods; there is little defined character to individual brands, so that the suggestion of the retailer has much to do with the consumer's choice. 

Building materials, even though adver tised to consumers, are generally sold to contractors and others who regard them as industrial goods; profit on the trans action rather than immediate enjoyment of the article is a prime consideration. These buyers are likely to be in close touch with the trade so that they do not need the manufacturers advertising as an aid in purchasing. For these reasons it was to be expected that consumer advertising would not be as important a factor with such commodities as with strictly consumer goods. It is of interest that there is any measurable effect at all from such advertising at a time of
depression. 

The fact that automobile sales of the companies which increased their magazine advertising fell off in 1921 more than did the sales of the non advertising firms deserves special comment. The comparison here made is with the Ford car only, and it is well known that the Ford Motor Company has long received large amounts of free publicity; it is hardly fair, therefore, to consider it as a nonadvertising concern. Moreover, it was to be expected that there would be a movement towards
lowpriced cars during the depression, which doubtless helped the Ford sales in comparison with the sales of other groups. The Ford price in 1921 was just about onehalf of the average price of the cars included in the other groups. The relatively small spread between the sale s of the firms which increased their advertising and those which decreased suggests that automobiles are sold to a large extent by personal solicitation, and that advertising is of less importance with them than with other commodities.

To summarise the suggestions that grow out of this study, it would seem:

  • That increased magazine advertising during depression generally resulted in an increase in sales relative to the sales of competitors who did not make such increases.
  • That reduction of advertising during depression generally resulted in a greater falling off of sales than occurred with firms which did no consumer advertising.
  • That different classes of commodities respond in different degrees to changes in advertising. In general, it may be said that these differences are in keeping with accepted statements concerning consumers' buying habits.
  • If, in some cases, the decrease in magazine advertising has been offset by an increase in the use of newspapers or other media, the comparisons indicate that such change of medium has not resulted in larger sales.
  • The increase in sales which apparently accompanies increased advertising during depression does not necessarily indicate an increase in profits; it has been impossible to consider the profit aspect of the problem in this study.

NOTES

The above article was drawn to the attention of the Journal by Professor Harry Henry. It first appeared in Harvard Business Review, April 1927. We are grateful to the Harvard Business Review for their permission to reprint it here. Copyright 1927 by the President and Fellows of Harvard College; all rights reserved.

1. Crum, W. L., Harvard Economic Service, Weekly Letters, Vol. 3, p. 47.

NOTES & EXHIBITS


Some Observations on the Effects of Newsprint Rationing (1939�1959) on the Advertising Media

FIGURE 1: relative movement of sales � associated with different advertising polices