Reaching Some Conclusions - Promotion Marketing Association researches sampling - Brief Article

Brandweek, May 7, 2001

An analysis of new data is the first step in quantifying the effect of sampling on sales

The Promotion Marketing Association's Product Sampling & Demonstration Council provides education and information on issues affecting all aspects of sampling. The Council enables companies with an interest in sampling to share ideas, discuss common concerns and represent their views within the general PMA membership (the views of the Council do not necessarily represent the views of other PMA members).

To demonstrate that sampling is an effective promotion tool, the PS&D Council has sponsored educational seminars, qualitative research and publications on trial and conversion. It has been interested in conducting quantitative research as well, but had not found quite the "right way" to do it. That is, until now.

The Council teamed with Brand Marketing and ACNielsen for groundbreaking quantitative research on the effectiveness of sampling. Over the next few pages are results of sampling events in four different product areas: health and beauty care, salty snacks, breakfast foods and meal solutions.

Four different distribution methods -- newspaper, direct mail, in-store and special events -- were analyzed.

Newspaper sampling is samples distributed to consumers with a newspaper.

With direct mail, samples are delivered via the U.S. Postal Service.

For in-store sampling, samples are distributed in a retail outlet where the product is sold.

With special events, samples are given out during an activity such as a school event.

The Council was interested in quantifying the ability of sampling programs to drive incremental sales. The report is presented as four case studies. Each includes generalized results of the research along with category trends.

To interpret the results, it is important to understand how the study was organized. Council members were asked to provide the ACNielsen KnowledgeWorks group with details of client sampling events encompassing various distribution methods. They provided a description of the sample, the distribution method, time frame, quantity, geography, and whether or not a coupon was included.

For direct mail events, the KnowledgeWorks team utilized ACNielsen Homescan consumer panel data to assess the effectiveness of the programs. For all other events, the data source was ACNielsen SCANTRACK store data.

The impact of sampling was established through a statistical technique called Analysis of Covariance. Essentially, for each sampling event a test and control group were established. The control group acted as a "baseline" to determine what would have happened to the test group if they had not been exposed to the sampling event. For example:

* For direct mail events, the test group consisted of Homescan panelists who received the samples. The control group was comprised of Homescan panelists who did not receive the samples but whose purchase patterns closely mirrored those of the test group during the time period prior to the sampling event.

* For in-store events, the test group included those stores where samples were distributed. The control group consisted of stores where pre-sampling event sales of the product in question were similar to pre-event sales in stores selected for sampling.

In order to conduct this study the Council adopted a strict confidentiality policy. ACHielsen conducted detailed analyses of each event for which adequate data were available. Brand Marketing is presenting the results generically without revealing the marketer/brand or which sample distribution method was used in order to honor the confidentiality required to present this report.