How Vera Bradley Profits From Big Data

Big data was undoubtedly one of the most buzzed about topics in the retail industry last year. What's most interesting about that is that there's not one universally accepted definition for the phrase — it means different things to different brands. For Vera Bradley, an omnichannel retailer of quilted handbags, accessories and luggage, big data means insight into what its customer wants.

In a session yesterday at the National Retail Federation's Big Show in New York City, Scott Steever, senior director of strategic initiatives at Vera Bradley, discussed how the brand uses big data to make its marketing and merchandising decisions more profitable.

Email Marketing
Vera Bradley wanted to improve the productivity of its email program, Steever said. After years of declining effectiveness, in large part to a "spray and pray" strategy of sending the same message to everyone on its list, the retailer began to use big data customer insights — e.g., purchase history, average order size, types of products purchased — to send targeted messaging to its subscribers.

The results were predictable. In a A/B test where the same offer was sent to Vera Bradley's full list and a segmented list based on the style of bag purchased, the segmented list performed substantially better despite 63 percent less emails being sent. The segmented list recorded a 43 percent increase in open rate, 101 percent increase in clickhrough rate and a 275 percent increase in conversion rate. Overall, the campaign sent to the segmented list generated a 14 percent increase in total revenues compared to the unsegmented list.

Pricing Optimization
In addition to improving its email program, Vera Bradley is using big data to optimize its pricing. With the help of First Insight, a product testing platform provider, Vera Bradley has revamped its pricing strategy to focus on pricing products based on consumer value and competitive positioning. Big data insights that Vera Bradley now factors into its pricing include competitive set analysis (i.e., what are its competitors charging for similar products), margin analysis and "what would they pay" online consumer engagement campaign data. Within that consumer testing, Vera Bradley is also collecting the percentage of favorable and unfavorable responses to the MSRP, the model price the consumer would pay for the item, and qualitative data as well. All of this data helps Vera Bradley price a single item.

Leveraging big data for pricing has paid dividends for Vera Bradley. Steever cited an example from Vera Bradley's baby line, specifically a dress and bloomer set, that was originally priced at $49. After analyzing the data as well as the feedback it got from customers, the retailer changed the price of the product to $48. The move resulted in a 6 percent increase in sales in the indirect channel and a 2 percent increase in sales in the brand's direct channel.

Pattern Testing
If you're familiar with Vera Bradley, you know that patterns are a big part of its business. In fact, Steever said they're the lifeblood of the business. So it's not surprising that the company would use extensive testing — and big data insights — to help it determine what its most popular patterns will be. While the pattern testing that's in place now is rather new — previously Vera Bradley relied solely on the opinions of its designers and merchants to determine which patterns would be sold — it's been quite effective at predicting the "winning" patterns.

Vera Bradley works with market research company Ipsos to measure consumer sentiment on pattern designs before deciding which to put into the marketplace. The goal of the testing is to find the patterns that stand out and resonate with consumers, Steever said. Biometrics such as heart rate and sweat output are measured for consumers who are exposed to the test patterns to get their subconscious feelings. Then a survey is used to get the explicit opinions of these consumers on the test patterns. Finally, a select group of the survey respondents are chosen for a focus group and consumer panel.

Vera Bradley typically starts a season with 16 to 18 pattern designs, Steever said. After analyzing the data post-testing, that number is reduced to three to four. We have to make our bets earlier in the process because of the long lead times with our products, Steever said. To verify the effectiveness of its pattern testing process, Vera Bradley back tested older products on which it had sales data to determine if it's a good predictor of future success. It was.

The toughest part of the process was convincing our designers that the pattern testing was effective, Steever noted. Now they've come to realize that it's another tool to help them do their jobs more effectively.