The study, conducted by independent research firm Knowledge Networks-PDI (KN/PDI: www.knowledgenetworks.com/pdi) and commissioned by PromoWorks (www.promoworks.com), is among the first to utilize frequent shopper data to follow consumer behavior long after the day of the sampling event to test the impact on sales. KN/PDI implemented a matched panel test and control research design using frequent shopper-based household panel data from their National Shopper Lab of over 16 million frequent shopper households.
Aptly entitled R.I.S.E. (Report on In-store Sampling Effectiveness), the study showed in-store sampling is a cost effective way to drive trial and sales in ways previously unimagined by marketers. Specifically, the study found:
•In-store sampling works—driving trial and sales. The average cumulative trial for the sampled items was +58% over 20 weeks. Sampled items in multiple categories showed an average +475% cumulative sales lift on the day of event. •In-store sampling impacts sales in ways never thought possible: oIn-store sampling drives additional repeat purchase. The average cumulative first repeat purchase for sampled products was +11% and +6% for the brand franchise over a 20 week period. oIn-store sampling drives sales for existing products and line extensions. The sales lift for the existing product sampled was +177% for day of event and +57% after a 20 week period. The sales lift for the line extension product sampled was +919% for day of event and +107% after a 20 week period. oIn-store sampling drives brand franchise trial and sales. The sampling was found to have a significant impact for the parent brand of the sampled products with +107% average sales lift on the day of event and +21% average sales lift after a 20 week period. The average cumulative trial for the brand franchise was +19% over a 20 week period. oIn-store sampling delivers new buyers - to the sampled items and to the brand franchise. The average cumulative new buyers for sample products was +85% and +23% for the brand franchise over a 20 week period. oIn-store sampling increases the average household shopping basket size. As a result of the sampling event the involved consumers’ overall shopping basket expenditure increased +10%, as compared to the average frequent shopper basket in the participating retailer. This suggests sampling contributes to incremental growth and does not cannibalize other items within the brands’ own franchise. •In-store sampling impacts sales long after the day of event, making it incredibly cost effective. By utilizing frequent shopper data, the actual sales impact of the in-store sampling can now be measured. Over the 20-week period, the sampled items saw an average cumulative sales lift of +74%. In-store sampling’s cost effectiveness when applied to all of sampling’s benefits over time is extraordinary.
“The R.I.S.E. study fundamentally changes the way marketers use and measure in-store sampling conducted by PromoWorks,” says John Stermer, PromoWorks’ EVP-sales & marketing. “Our clients—manufacturers and retailers alike—are all looking for ways to get closer to the consumer and have long sought a better understanding of the impact of in-store sampling. We fully expect to conduct similar studies in various categories in the future.”
Marketers have long recognized the value of in-store sampling for new products to generate trial on the day of the event. However, day of event analysis ignores repeat consumer purchase behavior, and the absence of valid metrics to measure the extended benefit to the brand leaves the perception that sampling events are costly, thus making it an impractical tactic for long-term sales of existing products. The R.I.S.E. study now provides the needed metrics.
“The R.I.S.E. study is innovative given the use of loyalty-card data (rather than conventional POS or panel data) and the study’s experimental design is impressive,” says Bill Schober, managing director-content and editorial for the In-Store Marketing Institute. “Most important are the initial results that sampling is indeed measurable; that the long-term sales impact of a sampling event is far more significant than traditional 'day-of sales-lift' measures; and that sampling's utility extends beyond new product introductions.”
“Frequent shopper data allows us to measure consumer behavior with unprecedented specificity,” says Neal Heffernan, SVP-GM, of Knowledge Networks/PDI. “We were delighted to apply this level of understanding to in-store sampling.” Utilizing its National Shopper Lab, KN/PDI’s systematic approach included the following:
•Identified households who were present in the store during the time of the sampling event. The households had to meet a standard static criteria to ensure they shop regularly in the store as well as a minimum shopping basket size on the day of the event to ensure they were in the store long enough to be exposed to the event. Potential control households were identified from shoppers who were not in the store during the event but who did shop in the same store in the two weeks surrounding the event. The same static and market basket criteria were applied to the control households. The potential test and control households were matched by store shopped based on dollar sales volume for the total category, the brand, and the sampled item(s) during a 52 week base period. The resulting sample size varied by sampling event, ranging from 3,000 households to 30,000 households per cell. •To account for any differences in the test and control household level cells, an Analysis of Covariance (ANCOVA) was used to adjust (extraneous non-sampling variables) sales over a 20 week test period beginning with the day of the first sampling event. Any remaining difference in dollar sales between cells in the test period was then attributed to the PromoWorks in-store sampling event. Because these shoppers were identified at the household level, KN/PDI was able to conduct additional analysis, such as; percent of households trying, percent of households repeating, and previous brands/items purchased to comprehensively understand the overall effectiveness of in-store sampling.