We're seeing a lot of published information on various types of data lately. So let's take a brief look market research data used in the food industry.
There are two distinct types of data: retail sales data and consumer-specific data. Retail sales data tells how much of a specific product was sold, and what factors existed at the time that might have caused those sales (ads, displays, price discounts, etc.). Consumer-specific data tells who bought the product, and sometimes describes their characteristics, what else they bought, and what factors might have caused them to buy it (same as above -- ads, displays, price discounts, etc.). I'll review retail sales data this issue, and consumer-specific in a future column.
Since the 1930s, retail sales data have been audited manually by A.C. Nielsen Co. Auditors counted inventory at the shelf and in the back room, factored in product deliveries and returned product, and calculated retail sales. These retail audits were conducted on a rolling bi-monthly schedule and results were reported to customers during the following period.
In the 1970s, warehouse withdrawal data developed as a popular method of measuring national, regional and local sales. It was collected and provided by SAMI (Selling Areas Marketing, Inc.). Warehouse withdrawal represented product moved from the warehouse to the stores, not what went out the retail doors. SAMI dropped out of the business when scanner data supplanted warehouse withdrawal information.
In 1979, Nielsen introduced ScanTrack, a retail sales data service based on scanner data, initially using a sample of about 150 supermarket outlets. Keep in mind that there weren't that many stores using scanners at this time. Scanner tracking by region and by top market areas soon increased rapidly. In 1986, Information Resources began providing their scanner data product, called InfoScan.
Originally, suppliers like Nielsen and Information Resources had samples that projected to markets, to regions and to the total United States In the late 1980s, retailers began allowing these companies to release "account specific" data, which consisted of projectable sample stores of a specific chain or wholesaler within a market, in addition to the entire market. All this information was provided in weekly increments (because that's when retailers pull data from their stores), and it wasn't available to manufactureres for two-three weeks after the fact.
More recently "census" or "all-store" scanning data has become available, meaning retail sales are available from each store in the cooperating retail chains.
When store-specific data became available as a concept, in 1990 Efficient Market Services took the concept a step further and developed systems that collect data at the stores, and make information available the next day, instead of the standard two to three weeks. Ems and Nielsen formed an alliance last spring to supply daily, store-specific data. In 1992, Information Resources and Catalina Marketing followed with a joint venture called Catalina Information Resources. This joint venture was dissolved only a few weeks ago, but the business relationship still stands, and IRI will provide daily, store-specific data as part of its offerings.