Camera Ready

JULIE GALLAGHER

Retailers and manufacturers are gaining valuable insights with ceiling cameras and analytics

“Shop as if no one is watching” seems like a simple enough command. But place a recruited consumer in a simulated grocery environment and bias almost always comes into play.

That’s one reason why brand marketers are trading time in labs fashioned after specific chains, for real-life observations in actual stores.

If consumers appear to be shopping as if no one is watching, it’s because they’re unaware that someone actually is. Tests are discreet since they involve ceiling cameras and video analytic software that gauges demographics such as gender and age range, while tracking consumers’ traffic patterns, dwell time in front of products, and whether or not they make a purchase.

“It’s almost like real-time feedback about how many people clicked on a page. This is the equivalent in a real world setting,” said Rajeev Sharma, founder and chief executive officer of VideoMining Corp., provider of video analytics to Food Lion, 7-Eleven, Circle K, AM/PM, Sheetz and Cumberland Farms.

As part of one study, tens of thousands of shoppers were observed interacting with a CPG company’s packaging, both pre- and post- redesign. In another, a brand marketer tested three versions of a new package to see which led to the most purchase conversions.

Sharma was not authorized to share the name of the tested brands, but said that VideoMining is conducting similar research with 26 manufacturers including Procter & Gamble, and over a dozen grocery, convenience, mass and drug chains. As part of the model, manufacturers pay VideoMining to conduct tests. Retailers like Food Lion — which has two stores with 120 ceiling cameras — don’t get any direct financial benefit from brand marketer’s research. But participation is worthwhile since they gain access to results that can help grow sales.

As part of another VideoMining test a brand marketer explored whether a display positioned beside a complementary category brought incremental sales. It turned out it did not. “We were able to very precisely tell how many people picked up the product,” said Sharma. VideoMining also analyzed how many people who picked up the product, visited the category positioned beside it. “But even if the consumer didn’t pick up the product, we tested whether the display was driving more traffic to the adjacent category.”

Similar research will be conducted across a wider geography beginning in August with the launch of VideoMining’s Grocery Shopper Insights program. It will involve 120 grocery stores across the country.

Fifteen stores will be fitted with 100 cameras, while consumer audits and intercepts will supplement video collected by a fewer number of cameras in the remainder of stores. VideoMining would not disclose the names of retailer participants, but at least 12 grocery banners are said to be in the study.

“Retailers will get access to data from their stores and the aggregate data from the study so they can benchmark their performance along key shopper metrics,” said Sharma.

VideoMining conducted a similar program with several convenience chains. It found that over a third of visitors shopped for some kind of beverage immediately after entering the store; nearly one in four (24%) stopped to shop the food service category; and stores generated more than $3 for every minute consumers shopped the store.

Independent tests that use video analytics to study shoppers are also cropping up.

p>Green Hills, a one-store retailer based in Syracuse, N.Y., recently began working with brand manufacturers to test displays, shelf signage and planograms in its store. The retailer has been conducting similar research independent of its trading partners since 2009. Its store is equipped with more than two dozen ceiling cameras that capture video at its entrances, end caps, and within certain aisles and categories.

Unlike the VideoMining system, Green Hills’ video analytics from BVI Networks, San Jose, Calif., do not gauge shopper demographics. The technology simply measures traffic by hour and day, to inform things like labor scheduling, and analyzes shopper interaction with different products and displays.

“The video is measuring traffic flow down the aisle; it’s measuring if I stopped and stood before this section or this brand; and then it’s recording if there was a purchase made at the POS,” said Gary Hawkins, CEO of Green Hills Market and president of the Center for Advancing Retail & Technology.

The cameras do not track a specific shopper through the store to determine whether a purchase is made. Rather, information related to the number of people who walked down the cereal aisle is compared to POS data that shows how many boxes of cereal were purchased within a specific time frame.

“With this data we can understand who made the purchase. Was it a gold level shopper or a lower level shopper?” said Hawkins.

In one test involving Nestlé products, Hawkins studied the effectiveness of two-foot-square off-shelf candy displays. He found that dwell time was 30% longer than at category shelves; overall product exposure increased almost 500%; and the number of shoppers engaged grew by more than 300%.

Something Hawkins may have been unable to tell at the time, was how many Nestlé products were purchased from the primary display and how many could be attributed to the secondary. That’s changed now that Green Hills is using Shoppergauge — a tool that ties BVI’s video analytics to a shelf system from Rock-Tenn Co. that detects when one or more units of product is taken from the shelf.

Shoppergauge helps determine whether items have been taken from the shelf, only to be abandoned elsewhere in the store. The tool will also alert store managers to a possible theft when several high-priced items are picked up at one time, and can determine how many sales were lost due to an out-of-stock.

“We can tell you if a high-end cosmetic is out of stock, how many people walked by the out-of-stock, and if 10% of those who usually encounter the product buy it, how much money was lost,” said Alexei Agratchev, CEO of BVI Networks. Shoppergauge is also helping Hawkins test the effectiveness of different signs, package designs and displays. One piece of research showed that shelf signs used to highlight a product actually reduce dwell time rather than prolong it. “We’ve seen across several categories that if there is some type of signage on the shelf — whether it be a price promotion on a brand or simply information being called out — it’s helps the shopper make a faster decision about what to purchase,” he said.

Green Hills is also exploring product adjacencies and their impact on incremental sales. The retailer found that a cereal promotion helped boost sales of juice, positioned beside it in the same grocery aisle. “We are beginning to leverage this information for promotion planning and [to rearrange] our category sets and aisles,” Hawkins said.