Published on April 16th, 2014 | by Travis Korte0
5 Q’s for First Insight CEO Greg Petro
The Center for Data Innovation spoke with Greg Petro, co-founder and CEO of Pennsylvania-based retail analytics firm First Insight. Petro discussed which retailers he thinks are doing analytics right and how good data can help retailers avoid investing in products consumers do not want.
Travis Korte: Can you introduce First Insight, what you make, and who uses it?
Greg Petro: Over 50 percent of new products fail in the retail industry. The result is over $250 billion lost each year due to excess inventory and markdowns. Also, many companies miss the upside and experience stock-outs because they don’t buy deeply enough on winning products.
I founded First Insight in 2007 to address this problem. We are a solution provider that empowers retailers and manufacturers to drive new product success by introducing the right new products at the right price. Our cloud-based software solution gives retailers and manufacturers insight into expected product performance and optimal entry price points for new items that have no sales history, all within 48-72 hours.
First Insight delivers this through a scalable platform that enables thousands of consumers to evaluate hundreds of new products through online games that are presented to them via social media, websites, emails, and on mobile devices. Our predictive analytic solution filters and weights consumer input, ensuring that our clients are listening to the right consumers. The result has been three to nine percent gains in sales and margin dollars for all of our customers.
Retailers and brands use our solution—specifically designers, merchants (buyers), and planners. We help them develop, select, and price more winning products. Often we can enable a company to double its win rate on new products.
TK: What retail companies would you say are doing a good job with analytics, and why?
GP: Many retailers have embraced the value that predictive analytics can bring to their businesses. The first wave of analytics was adopted on the “sell-side” of the business, and of course the first applications here were in e-commerce where the data is structured and is easy to mine. Amazon is famous for monitoring customer search and purchase behavior, applying analytic models, and making recommendations for additional purchases. The company is expanding its analytics capabilities with a new, patented algorithm referred to as “anticipatory shipping,” which places items close to the consumer in anticipation of an order, to reduce shipping time.
Now, companies like Burberry are integrating online consumer behavior with in-store actions to present targeted offers while the customer is still in the store. And Macy’s constantly mines its customer data and presents targeted offers via email campaigns.
The “second wave” is the application of predictive analytics to the “buy-side” of the retail business—making product design, buying, and pricing decisions. Our client Abercrombie & Fitch, as we announced in a press release in February, is “now testing new candidate products, in every product category, every week, throughout the merchandising organization and the product development lifecycle (design, selection, pricing, and buy depth).”
TK: How does having highly responsive information about customers’ tastes change companies’ behavior? For example, do you find that a better understanding of a new product’s chances to succeed makes companies more inclined to go for the sure thing, or does it allow them to be experimental?
GP: The proper implementation of analytics can have a major impact on a companies’ behavior as it relates to their customers. Traditionally, retailers have used intuition to make decisions on which new products to bring to the market. But as reported by Gartner, MIT Sloan School of Business, and others, more than 50 percent of new products fail. Some would put this number as high as 70-80 percent. By engaging with consumers, retailers can use data, in conjunction with merchant experience, to gain more confidence in their decisions on new products. In this way, retailers are adding science to the art of selecting new products, which in and of itself requires a change in behavior throughout the organization.
It’s all about reducing risk. We give the merchant the confidence to buy deeper on a winner, but we also give them the confidence to select a new product or even enter an entirely new category. Just as important is the ability to eliminate poor performers early in the process, before the company spends a lot of money making samples and investing in inventory.
With an accurate forward view of how the market will respond, merchants sleep better at night because they have data to back up their decisions.
TK: You conduct a great deal of data collection to support your analytics. Is this collection mostly ad-hoc, or do you try to re-use data? Have you made any serendipitous discoveries from analyzing older data?
GP: The data that First Insight collects is unique to each of our customers. We focus our solution on each customer by first ensuring a complete understanding of the challenge they are facing and trying to solve. The companies we work with are varied in the types of products they offer. We work with companies that sell apparel, footwear, jewelry, accessories, sporting goods, electronics, and home goods, to name just a few.
Our system has been running for over six years, and we now have tens of millions of data points across all of these product categories. Our analytics engine continues to learn and improve as additional data points are collected.
One of the biggest discoveries we have made from the data we have collected is the missed upside on new products. We collect nearly three million data points a month, and from that data we have uncovered the fact that more than 11 percent of new products can bear a higher price than originally thought. In a time when many retailers are in a “race to the bottom” to meet or beat a competitor’s price, we have shown them how to avoid this race and instead lead the pack when it comes to new product success.
TK: Have you gleaned any insights about how retail trends work generally? Do different products have different expected lifespans, for example? What factors influence people moving on to a new style?
GP: Retail trends, as you may already know, are very difficult to predict. Consumer preferences can change very quickly and there are many factors that can affect their preferences. But the key is to ask the customer—because she knows what she wants.
We work with fast fashion apparel retailers which are making decisions every week on new assortments. But we also work in categories such as jewelry where product lifecycles are a year or more. Regardless of the lifespan, implementing a disciplined, predictive analytic approach that effectively incorporates the voice of the customer yields far better results than relying on the judgment of single merchant.