POINT-OF-PURCHASE DISPLAY AND BRAND SALES

POINT-OF-PURCHASE DISPLAY AND BRAND SALES Michael Williamson Casio Electronics Ltd Robert East Kingston and Wendy Lomax Kingston Business School Abstract We report the results of two field experimental studies on the relationship between store display and sales of a new brand. When the space in a single display was increased, sales tended to increase disproportionately. When the number of displays was increased from one to two, the sales effect of the second display was additive. Introduction Window and special in-store displays are a common feature of retailing. We know that such displays can generate substantial increases in grocery sales (IRI 1989; Inman, McAlister and Hoyer 1990) but we lack a more systematic knowledge. Specifically, it is not clear how much sales are affected by the amount of space allocated to a single display in a store and how one store display interacts with a second display in the same store. Most evidence on display effect is obtained in supermarkets where scanners and store location programs can be used to gather data. Display is important in supermarkets because most grocery purchase decisions are made at the point of purchase (Dagnoli 1987). In addition, the choice of grocery brands is generally low-involvement and this may be conducive to more situational prompting. We lack evidence on other retail categories. In the research reported here, the category chosen for study was watches. The purchase of a watch is likely to be more involving than that of groceries, particularly in our study where the brand used, GShock, was an expensive and fashionable product. Research Evidence The effect of shelf space and position Experimental studies on shelf space variation in grocery categories show that a doubling of space leads to modest increases in sales, mostly of the order of 20 per cent (Curhan 1972, Kotzan and Evanson 1969, Krueckenberg 1969). One study was lower than this with only the best of four investigations giving 21 per cent sales increase (Cox 1970). Space use studies are complicated by the fact that sales gains for categories that are given more space are offset by the sales losses of other categories that lose space. The profit implications of space and location changes may be assessed by using proprietary programs that manage the changes. In this way, Drèze, Hoch and Purk (1994) showed an average profit gain of about 15 per cent when the worst to best configurations (including both location and space) were compared but, in practice, the authors estimated that profit gains would be much smaller. Drèze, Hoch and Purk (1994) found that more effect came from location than from space as long as the brand maintained a shelf presence. The effect of special displays Special display, e.g. end-aisle in supermarkets, gives more effect than facing changes. Bemmaor and Mouchoux (1991) examined the combined effect of discount and end-aisle display and found elasticities of –2 to –11 when 12 brands in 6 categories were studied. Brand elasticities are normally in the region of –2 (Tellis 1988a), so this finding suggests that the end-aisle display amplifies the elasticity substantially. This was supported by work conducted by IRI (1989), which showed that a discount alone produced a price elasticity of – 2.3 but that, with display, this increased in magnitude to over –18. Other studies have shown a smaller but nonetheless substantial effect of special displays (Allenby and Lenk 1995, Tellis 1988b). These studies deal generally with established products; displays of new products might show more effect. Although an increase in sales may be expected with more display space, this relationship could be linear or curvilinear. If sales increase disproportionately with space in a single display, storeowners should have few large displays. By contrast, if sales show diminishing gains with increases in display space, as indicated by facings studies, storeowners may benefit from a larger number of small displays, if extra displays do not show such diminishing returns. Thus, to study this matter, we require evidence on the sales effect of space changes in single displays and also the sales effect of extra displays. Do two displays show a synergistic effect? Store displays are forms of advertising and multiple displays may have the effect of multiple ad exposures. The impact on sales of repeated advertising exposures has been controversial. Most research has indicated that incremental exposures produce diminishing effects (Broadbent 1998, Jones 1995, Roberts 1996, Simon and Arndt 1980). For these researchers, an accelerating sales response to extra exposures is likely only when the product is new or complex. Other researchers expect initially increasing returns for most products (Krugman 1972, Naples 1979, Tellis 1988b). Recent work by Roberts (1999) has indicated that exposures that occur close together in time do produce an initially accelerating impact on sales. This suggests a synergistic process in which the first exposure raises the impact of the second; this process could occur either by priming (reviewed by Eagly and Chaiken 1993) or by overcoming interference effects (Burke and Srull 1988). In our study, the displays were likely to be seen close together in time and featured a new brand (the G-shock watch). Thus, it seemed more likely that a synergistic sales response would be found; i.e. two displays in the same store would give more effect than the sum of two separate displays. Evidence on the interaction between displays and different elements of the promotional mix such as discounts and coupon advertising also relates to this issue. East (1997) has reviewed IRI (1989) data on this matter. When discount, coupon and display are used together, the sales impact is greater than the sum of the impacts when each component is used separately. However, there is some doubt about the inference of synergy from these studies because they are non-experimental and conditions may differ between the stores offering three promotional features and those offering one (Totten 1986). We conducted two investigations in order to examine the relationship between the sales of the watch brand and display space under actual operating conditions. The watch category is relatively stationary but Casio had recently relaunched the G-shock brand as a stand-alone brand; it used striking designs and was several times more expensive than basic electronic watches. Since it was a new brand, there were no sales without display. Investigation 1 explored the relationship between the size of a single display and sales. Freestanding display cabinets of varied size were used. Investigation 2 was conducted in H. Samuel jewellery and watch stores in the UK. This study explored whether two displays in the same store had a synergistic effect on sales, i.e. whether the second display has a disproportionately large effect on sales. Investigation 1: Display space and sales We expect that sales will be a positive function of display space but we are interested in whether increases in space produce linear, increasing or diminishing returns. To test this we examined if variations in retail space (measured in square inches) produced variations in sales (measured in sales of SKUs). Sections of twenty independent fashion and department stores in the UK were used. These stores did not normally stock watches. This environment ensured that there was no immediate effect from competitor brands. The sections used varied in size from 1,600 to 2,890 square feet of retail space. Display showcases ranging from 144 square inches to 7,012 square inches were allocated to sections. As the size of the showcase increased, there was a corresponding increase in SKUs from 6 to 289. Allocation was random subject to a constraint; not all store sections could accommodate the largest cabinets, so these went to bigger sections. Because of this constraint, it is possible that section size affected the sales figures for large cabinets. Data were collected by sales staff, who conducted stock checks of individual SKU’s over a fourweek period. Results 500 400 300 200 Unit Sales 100 0 -100 0 1000 2000 3000 4000 5000 6000 7000 8000 Space Level in Square Inches Figure 1. Scatter plot of space level and G-Shock unit sales Increase in showcase size generated an increase in product sales. The Pearson correlation coefficient between the two variables was 0.88 (p = 0.00). The scatter plot in Figure 1 shows that the relationship between space and sales appears to be curvilinear with accelerating gains from increased display space, though this assessment rests on relatively few cases. Investigation 2: Effect of one versus two displays In Investigation 2 we examined whether window and in-store displays in the same store produced more sales gain than equivalent window and in-store displays in different stores. Sampling and experimental procedure The sample for this Investigation consisted of 66 stores of the H. Samuel retail chain. The selection of this retailer gave geo-demographic representation across the UK population (H. Samuel have 428 stores with a presence in every major town in the United Kingdom). The chain accounts for 10 per cent of watch sales in the UK. The 66 stores were assigned randomly to one of three treatments groups, as shown in Table 1. Each display contained the same 42 G-shock lines with the brand name clearly displayed on a placard. The way in which products were presented (e.g. solus window) was held constant in each display treatment. Other factors such as pricing, stocking and display of other products remained constant over the period of the experiments. Type of display Table 1 Experimental display treatments Form Window In-store Window and in-store 42 G-Shock lines 42 G-Shock lines 42 G-Shock lines (window) + 42 G-Shock lines (in-store) Number of stores 22 22 22 Data collection procedure H. Samuel central offices collected the data by weekly electronic point-of-sale download. Unit sales of each product line for each store surveyed were recorded over six weeks. The data also revealed the stock level in each store, which was used to prevent out-of-stocks. Results Table 2 shows the aggregate sales from window and in-store displays in different stores and these are compared with the sales in shops that have both types of display. Week 1 2 3 4 5 6 Total Table 2. Weekly sales of G-Shock units by display method Window (1) In-store (2) Aggregate: 1 Window and inand 2 store in same store 96 79 175 146 116 75 191 214 113 75 188 214 100 69 169 190 106 80 186 189 131 110 241 201 662 488 1150 1154 In Table 2, the two displays in one store produced virtually the same sales as the sum of single displays, each in a different store. This result indicates an additive response to a second display. A Kruskal-Wallis one-way analysis of variance showed that the difference in sales between window and in-store displays was statistically significant (p<0.05). In assessing these data, we take account of H. Samuel evidence that the sales of Casio watches appeared to be unaffected by G-Shock sales. We cannot rule out some cannibalisation of Casio sales but since G-Shock watches were several times as expensive, any switching was profitable. Discussion and Implications The results from the two studies seem paradoxical. When single display space is expanded there is a disproportionate increase in sales but when two displays are used, the sales increase is close to the sum of the increases obtained when single displays were used. In the first study, there was some confounding between the size of the store section and the size of the display cabinet. Because of this, it is possible that the disproportionate rise in sales with showcase size was artifactual. However, another explanation for the finding from the first study was that the larger showcases were physically obstructive and unusually large and that this pre-empted attention. If replicated, this finding has implications for display space management and adds to our understanding of point-of-sale effects. Of the two studies, the second is the more fully controlled. The apparently additive effect of displays is of particular interest. The evidence from advertising research suggested that the effect should be synergistic, particularly with a high-involvement new brand and when the two displays were likely to be seen within minutes of each other. If most customers saw only one display, an apparently additive effect would be observed but H. Samuel shops are not large and many customers should see both displays. We conclude that our results run against evidence from advertising research. Results are also at odds with space elasticity findings from facings research. Here elasticities of about 0.2 are common; our data indicate an elasticity about five times greater. The superiority of window displays in the second study is, perhaps, not surprising. Windows are accessible to a larger number of potential customers when compared with in-store displays. IRI findings have suggested that displays produce synergistic effects when combined with discounts and local advertising. However, this work is non-experimental, as noted by Tellis (1986). To our knowledge, our study is the first field-experiment that tests synergy in the mix and, although our work differs from the IRI research, our lack of evidence of synergy in the second study suggests that we should be cautious about making assumptions that it occurs elsewhere. Taken together, these studies make it unlikely that increased special display gives diminishing sales gains in the retail category studied. An extra display seems to give a good return. Presumably, a point occurs when further display produces diminishing sales gains but our studies do not reveal this. References Allenby, Greg M. and Lenk, Peter J.1995. Reassessing Brand Loyalty, Price Sensitivity, and Merchandising Effects on Consumer Brand Choice. Journal of Business and Economic Statistics 13 (3), 281-289. Bemmaor, Albert C. and Mouchoux, Dominique (1991). “Measuring the Short-Term Effect of In-store Promotion and Retail Advertising on Brand Sales: a Factorial Experiment”. Journal of Marketing Research, 28, (May), 202-14. Broadbent, Simon (1998). “Effective Frequency: There and Back”. Admap, May, 34-38. Burke, R.R. and Srull, T.K. (1988) Competitive interference and consumer memory for advertising, Journal of Consumer Research, 15, June, 55-68. Curhan, Ronald C. (1972). “The Relationship Between Shelf Space and Unit Sales in Supermarkets”, 9, (November), 406-412. Cox, Keith K. (1970). “The Effect of Shelf Space Upon Sales of Branded Products”. Journal of Marketing Research. 7, (February), 55-58. Dagnoli, J. (1987). “Impulse Governs Shoppers”. Advertising Age, October 5th , 93. Drèze, Xavier, Hoch, Stephen J. and Purk, Mary E. (1994). “Shelf Management and Space Elasticity”. Journal of Retailing, 70 (4), 301-326. Eagly, A.H. and Chaiken, S. (1993) The Psychology of Attitudes, Orlando, Harcourt, Brace, Jovanovitch. East, Robert (1997). Consumer Behaviour: Advances and Applications in Marketing, Prentice Hall, London. Inman, J.J., McAliser, Leigh and Hoyer, W.D. (1990). “Promotion Signal: Proxy for a Price Cut?”. Journal of Consumer Research, 17, (June), 74-81. IRI (1989). “Larger Sample, Stronger Proof of P-O-P Effectiveness”. Reprint from IRI which enlarges on a report that first appeared in P-O-P Times, March/April, 28-32, 1989. Jones, John Philip (1995). When Ads Work: New Proof That Advertising Triggers Sales, New York, Lexington Books. Kotzan, Jeffrey A. and Evanson, Robert V. (1969). “Responsiveness of Drug Stores Sales to Shelf Space Allocations”. Journal of Marketing Research 6, November, 465-469. Krueckeberg, Harry F. 1969. “The Significance of Consumer Response to Display Space Reallocation”. Proceedings of the American Marketing Association Fall Conference, 30, 336339. Krugman, Herbert E. (1972). “Why Three Exposures May Be Enough”. Journal of Advertising Research, 12, 6, 11-14. Naples, Michael J. (1979). Effective Frequency: The Relationship Between Frequency and Advertising Effectiveness, New York, Association of National Advertisers. Roberts, Andrew (1996). “What Do We Know about Advertising’s Short-Term Effects?”.Admap, (February), 42-45. ---- (1999). “Recency, Frequency and the Sales Effects of TV Advertising”. (February), 40-44. Admap, Simon, J. L. and Arndt, J. 1980. “The Shape of the Advertising Response Function”. Journal of Advertising Research, 20 (4), 11-28. Tellis, Gerard J. (1988a). “The Price Elasticities of Selective Demand: A Meta-Analysis of Econometric Models of Sales”. Journal of Marketing Research, 25, 331-41. ---- (1988b). “Advertising Exposure, Loyalty, and Brand Purchase: A Two-Stage Model of Choice”. Journal of Marketing Research, 25, 134-144. Totten, John C. (1986). “Measuring Retail Sales Response to Retail Sales Promotion”. Paper presented at the ORSA/TIMS Marketing Science Conference, March.