PYMNTS : AI + Consumer Behavior Data = Sales Growth

“With the popularity of Amazon, Pandora and Netflix, today’s consumer has the expectation that retailers will know them personally,” said Craig Alberino, president of Grey Jean Technologies. “When they walk into the store or browse online, they expect an experience tailored to their unique needs, desires and wants.”

Grey Jean Technologies marries artificial intelligence with customer and payment data to generate accurate predictions of consumer behavior that ultimately focus on driving sales figures. The New York City-headquartered company uses its proprietary AI-powered recommendation engine called Genie to quickly merge existing data sources with more than 500 consumer behavioral attributes to identify what patterns connect consumers with specific products. These consumer insights are used by marketers and executives to engender ways to enhance interaction with consumers and ultimately improve sales.

“The right recommendations serve as a valuable discovery mechanism that connects customers with the content and products they actually want,” said Alberino. “ECommerce personalization helps retailers meet their customer’s needs more effectively and efficiently, making interactions faster, easier and more satisfying — encouraging repeat purchases and creating loyal customers.”

Back in 2015, conversations about the business began to swirl between the two founders, Cosmas Wong and Alberino, as a way to apply their combined experiences of working with Big Data in the financial and retail sectors. Grey Jean launched in May 2016 at the Shoptalk Conference, where Genie was unveiled, along with the work the firm was doing with its first two clients: Hiro Sake, a handcrafted premium spirit company, and Namco Pool, one of the largest dealers of swimming pools.

“Predicting what shoppers will buy next has long been a dream for retail marketers, but very few have had the technology to pull off that vision,” said Alberino. “While many advancements have been made over the last 20-plus years, true one-to-one, real-time personalization is still a rarity in retail, due to its historical struggle with lack of data, followed by too much unorganized data.”

Alberino said that, through advances in artificial intelligence and machine learning, retailers have the ability to harness the power of their data to predict consumer purchase behavior and more effectively target those customers based on their unique preferences and behaviors. And, over time and through more data, Genie’s predictive algorithms inspire better recommendations, which, in turn, is an increased benefit to retailer clients that are looking for a slew of outcomes.

“Genie has demonstrated a 72 percent accuracy rate in predicting a next likely purchase at the category level. This accuracy enables our customers to improve personalization and micro-targeting but also take actions that will be the most effective with each consumer,” said Alberino. “This results in more conversions, higher redemption of coupons and promotions; increased visit frequency, foot traffic, time in store and basket size; and greater brand affinity.”

Alberino gave the example of one of its new clients that has already achieved some quick success. Pure Green, a New York City-based juice bar, had Grey Jean provide insights into its customer base as it sought to expand the number of its retail locations. First, there were three locations, but the hope is for 30 by the end of 2018.

“Leveraging Genie has allowed Pure Green to have a greater understanding of their consumers’ demographics, preferences and behaviors, helping them identify who their target audience really is and, subsequently, where they would have the most success with a new storefront,” said Alberino.

Ultimately, more time and more data only helps this type of technology. The company said that, each passing year, Genie’s algorithms get stronger and more refined. And it helps that more people are shopping digitally, where AI can easily live.

“We plan to stay ahead of the curve by continually integrating new attributes and data points that keep Genie the most accurate recommendation engine for predicting consumer purchase behavior,” said Alberino. “We also want to help shape retailers’ understanding of AI, which is currently used in many ways within the market.”

While some retailers currently use AI technologies — such as natural language processing — for sentiment analysis, Alberino said that this application alone will not give the full picture of what retailers are looking for in terms of consumer shopping behavior. While it may be helpful to know how a consumer feels about an ad or product, retailers also need to understand what drives purchasing.

“Focus on using artificial intelligence to drive actions, rather than just delivering insights,” said Alberino. “We actually help retailers take an action with their customer, rather than just presenting their data in new ways and leaving them wondering what to do with it.”

Convenience Store Decisions: Boosting Your Data IQ

From analyzing customers to assessing their inventory, more convenience stores are using operational data to their advantage.

By Howard Riell, Associate Editor, Convenience Store Decisions

Today, collecting data is clearly essential for convenience store retailers. But capturing and making use of the most pertinent information—from customer analytics to competitive intelligence—can spell the difference between operating in the black or not.

Data solutions aren’t always black and white, however. Often, it requires operators to assess category gross margins and related variables—especially in the form of inconsistencies or problems—and figure out how to address them.

Within a c-store today, being able to track customer tendencies, the workforce, fuel and product sales and inventory is vastly important to a retailer’s competitiveness.

“Typically, they are not properly accounting for variations in inventory,” said Steve Montgomery, president of b2b Solutions LLC , a Lake Forest, Ill.-based consulting firm, referring to typical scenarios. “The second reason is that something is amiss at the sites. This is true if the margin is higher or lower than is normally achieved.  In either case it should be investigated. Are the changes following the same pattern as their other locations as far as sales, margins, etc.? If not, why not?”

For many retailers, knowing how to gather and assess information from fuel inventory and customer sales is an important aspect of any store operation that sells fuel.

“Variations can be reported in both spread sheet and graphical format. Specific items can vary by chain. However, everyone should be monitoring sales (dollars, units and gallons), margins, customer counts, transaction size and expenses,” said Montgomery.

C-stores must also track fuel pricing processes and procedures.

“Obviously being in a c-store, fuel pricing is a huge deal and so you want to make sure that you are keeping up with your competition on that. It’s an easy question but a hard answer,” said Karla Grimes, director of operations for the Kent Cos., which runs 40 Kent Kwik Convenience Stores.

Being in the fueling industry, Grimes continued, many operators should ensure that they are staying up with that crucial parts of their business, as well as the competition.

“Of course, as far as consumer analysis goes, you’ve got to get into your competition and look at their stores and see what they are doing; how they take care of their customers,” Grimes said. “That’s a big part of it.”

“We do a lot of our data in-house,” added Alex Garoutte, marketing director of Kent Cos., based in Midland, Texas. “There are a lot of (data analysis) companies out there that are valuable. You just have to determine if it’s worth spending the money on.”

Inside sales, of course, deserve just as much focus.

“What are you doing, and what differentiates you from your competition?” Grimes said again.

STRATEGIC INITIATIVES
A good strategy for a chain of stores, according to Montgomery, is to look at the same type of data gathered from all locations at one point in time, and the same locations over time to determine patterns of activity. “The first allows the retailer to compare the results of his stores. The second will show whether changes are following the same pattern as their other locations,” Montgomery said.

Deciphering the data and then putting the results to use are the next steps. Disseminating the necessary operational information to key store personnel such as managers is vital and too often overlooked, Grimes added.

“You’ve got to decide what data you want to use and then figure out how to push that out. It could be a group of store managers, it could be employee leaders,” Grimes said. However you are going to do it, I think you have to go about getting that pushed out to every level of the company. And what you need is to have key players who are going to do that for you.”

SALES INFORMATION
First, a retailer must determine what data to share.

“Our experience tells us that the most important data that convenience store operators use is their own sales data,” said Brian Nelson, co-founder and chief operating officer of NewsBreak Media Networks Inc., a Knoxville, Tenn.-based programmatic merchandising platform for the convenience store industry that converts fuel-only customers to multi-product purchasers.

“The average c-store completes around 1,100 transactions (or more) per day. That is 1,100 unique customer surveys on what products customers want, and what products they will likely purchase together,” Nelson said. “A c-store’s proprietary sales data is a constantly updating data set that can be used to map and predict market trends, purchase patterns and inventory expectations. Ask any researchers what they would give for the ability to consistently collect 1,100 surveys daily.”

Data analysis and predictive analytics are areas that are experiencing rapid growth and innovation. If operators don’t seek help from experts, Nelson said, their category managers, marketing directors and executives are going to need to become data experts, which takes those individuals away from their proper roles and responsibilities.

“Seeking support from experts can help operators get to actionable information faster,” Nelson said. “Experts often have the infrastructure in place to parse, analyze and model historical and dynamic data quickly, instead of building it from scratch.”

Each time a customer completes a transaction, Nelson explained, he or she is giving the c-store operator direct feedback.

“Taking that feedback across multiple customers and multiple locations can give operators great insight into how they promote their products and even what products they carry,” Nelson said.

Market-basket data can be analyzed to help operators decide what products to promote together and what time of day they should be promoting those combination deals. For example, by promoting a discounted salad with a high-end water product, a c-store could increase unit sales for the promotional product while simultaneously driving up the total market basket.

C-stores can utilize data analytics to drive overall marketing initiatives, promotions and even inventory management and product selection. Analysis can include integrating key consumer buying variable data and demographic data into a store’s existing sales data in order to build location specific, programmatic merchandising campaigns.

Nelson encouraged operators to innovate and try new product offerings and marketing strategies, but warned that caution is called for. “Use data analytics as a tool, but don’t get stuck in the data.”

While a c-store operator will likely never get sales data from his competitor, operators can get a sense of how they stack up against the market, state or region from their vendor representatives. The companies that represent the brands that are sold through the c-store industry, Nelson said, can be a great resource to operators and category managers for target demographic information and customer analytics.

IDENTIFYING BEHAVIOR
Craig Alberino, the CEO of Grey Jean Technologies in New York City, said his firm generally works with behavior and identity data. “That’s things that you buy, places you go, even the weather. It could be various influencers.”

Knowing why consumers buy what they buy helps drive engagements and forge deeper relationships, resulting in more predictable revenue.

“It’s our view that all retailers, convenience stores included, could be serving customers better by knowing them more intimately,” he added.

Taking transaction and loyalty information inventory numbers, promotional calendars and more, and enriching it with elements like census statistics, weather and social media can provide a sharp focus on buying patterns for small geographic area around the store.

“Even if there is somebody in my dad’s age category, which is 70-plus, he’s still carrying an iPhone,” Alberino said. “He’s still looking for deals, and he is hypersensitive to gas prices. He knows if there is two cents less per gallon down the road, and he’s going to go to that other pump. Even though he might be spending more, it’s the trigger and it’s kind of a win, that he beat the system. And so, you are almost ‘game-ifying’ it a little.”

COUNTING CORRELATIONS
Retailers naturally want to drive up customer visits and average sales per transaction. More convenienc stores are evaluating customer sales over time to gauge future buying tendencies.

“This data is far more informative that just total sales,” said Montgomery. “It can tell why sales are up or down. Is it because fewer customer are coming to the store, or because they are buying less each time they come or both?”

Another tool is the correlation between gallons sold and store sales changed, he added. “Are they still buying your fuel but shopping more or less frequently inside you store? The data will not solve your problems, but should tell you where they are occurring.”

One of the easier analytics that can be checked is the correlation between fuel sales and in-store sales, Montgomery said. “Does an increase in gallons result in increased store sales? If so, does it impact certain categories more than others? These types of insights can help retailers determine what role they want a category to play in their overall marketing strategy.”

Montgomery suggested retailers benchmark their data with sources such as the National Association of Convenience Stores (NACS) because it might show opportunities to better sales and margins or control expenses.

“However, every store (or) market is different. The most important benchmarking retailers can do is against their own data,” Montgomery said. “That the industry averages ‘X’ is not as meaningful as whether you are making steady improvements in sales, margins and expense control.”

MORE DATA
In retail, data and data gathering comes in many forms and the benefits are just as varied.

“In any retail setting, convenience stores included, we generally work with behavior and identity data,” said Alberino. “Identity data would be things that you buy, places you go, even the weather. It could be various influencers. So generally when we are working with a retailer we will take transaction information and any loyalty information they might have, inventory, promotional calendars, and we will enrich it with census, weather, social media, etc. What is happening in the vicinity of your physical presence, within a mile of that convenience store?”

Among the data he suggests is scheduling of local events.

“There could be a rock concert happening, for instance, and in fact one of the convenience chains we’ve spoken to was in charge of sponsorship for a very large concert venue,” Alberino said. “We were talking to them about how they could increase traffic. Because many people were making long journeys across the West from California to the venue, how do you get them to stop, not only at the pump along the way, but also to stop in and get other things?”

 

Originally published on Convenience Store Decisions

Boosting Your Data IQ

Insights Are Overrated

Image courtesy of McKinsey&Company

With the amount of digital data in the universe growing at an exponential rate — doubling every two years by most accounts — it’s easy to see how recent years could be dubbed as “The Big Data Era.” Retailers and marketers played a large role in this, focusing on gathering data from every source, and squirreling it away for a rainy day.

From there, we came into the “Era of Insights.” Retailers began to take their data stores and analyze the information hidden within them, hoping to glean insights that will help improve their interactions with customers. How can we summarize what has been happening in the market? What guesses can we make based on what happened in the past?

Most recently, we’ve come to realize that insights are overrated. By simply analyzing data, we’re creating more data — but then what? How can retailers achieve personalization based on their insights?

Getting Actionable

In order to take advantage of our masses of data, the next phase is moving from insights into real-world action. Of course this is easier said than done.

There are three key factors when it comes to enhancing personalization, generating leads and engaging consumers, and these can be the difference between effective campaigns and those that fall flat. Retailers’ actions must be:

  1. Timely: Any action must be taken in the right moment for each consumer — on the right channel, and while they are in the right location and receptive to the right message. While the definition of timeliness used to stretch over a day or two while the consumer thought through a purchase, today’s on-demand economy has changed our perception of timeliness to a matter of hours, minutes and even seconds.

  1. Forward-looking: In the Era of Insights, personalization was done based on your past activities and purchases. If you bought pants, you must like pants, so we’ll offer you more pants. Today, we’re taking a smarter approach. Looking at past information, we’re able to make better predictions about what consumers will purchase in the future, and what actions we can take to prompt them to make that purchase. So if, for example, you purchased pants during a pre-Fall sale at a 30% discount, we can predict that you might be interested in a shirt to go with it if we offer you 40% off.

  1. Strategic: Analyzing results is an important part of any campaign, but it is also often time- and labor-intensive. Simply quantifying success or failure is an underuse of your team’s time. Retailers should be using the analysis of their results — including weighting and scoring terabytes of data against actual purchase behavior — to inform future personalization efforts and improve the efficacy of their actions.

Technologies to Consider

Achieving timeliness for each individual, making forward-looking predictions and strategically scaling those efforts is all labor-intensive work. For retailers looking to convert their data insights into actions, technologies to help automate the process are an absolute necessity. Though the marketplace for these technologies is crowded, they generally fall into a few categories that work together to create a comprehensive solution:

  • Understanding Identity: There are many tools for gathering consumer data across touch points, including POS systems, social media, CRM platforms, mobile web, apps and more. The challenge to date has been finding a way to link these disparate data sources to create a clear, omnichannel view of the individual consumer. In order to do this well, martech (marketing technology) solutions you invest in need to play well with others.

  • Understanding Behavior: Rationalizing data and understanding the patterns within it can be done most effectively today through artificial intelligence (AI). AI technology can dive deep into data, and find links that could otherwise be overlooked. For example, you might guess that pool owners would need cleaning supplies when the weather started heating up, but perhaps they also purchase scrubbing tools every time they get a car wash. Using machine learning, a subcategory of AI, you can identify and use patterns like this to refine your targeting algorithms to better predict these purchase behaviors in the future. With companies like Google and IBM investing more deeply in AI technologies, the space is becoming noisier. The martech companies that will help retailers be the most successful with AI will be those that create the most accurate predictive algorithms — and the results will speak for themselves.

  • Understanding Location: Once you understand the person and their behaviors, a key component to targeting them at the right time with the right offer often comes down to location. The most popular technologies that can help with this include beacons, in-store devices that communicate with a shopper’s mobile device using Bluetooth connections, and geofencing, a software feature that uses GPS or radio frequency identification (RFID) to define geographical boundaries and identify mobile devices that enter those vicinities. While studies have shown that consumers like the hyperlocalized-based personalization you can achieve with these technologies, there is a fine line between helpful and invasive. Retailers should be careful about how they obtain and use location information, and ensure they are protecting consumers’ privacy.

  • Understanding How to Take Action: While personalization and targeting technologies have vastly improved, they still need human oversight in order to ensure your actions are mapping back to the overall marketing strategy. In addition, marketing teams need to have the skills to put these personalization technologies to use. An intuitive user interface or dashboard that pulls it all together and makes optimal actions easy to see and execute is increasingly vital for time- and resource-strapped marketers.

Big data and insights are still key components to successful personalization for retailers, but we also need to be able to take a step back and understand how turn those into actions that drive real, quantifiable results. There are still technological and social barriers to overcome, but by understanding the components that constitute a successful solution, and looking beyond just generating insights, we can begin to move forward and make the coming years into the “Era of Action.”


Craig Alberino is CEO of Grey Jean Technologies. An expert in consumer behavior and loyalty, Alberino has advised top agencies within the Omnicom, WPP and Publicis holding companies, where he defined the digital strategies for clients including FedEx, Kimberly Clark and Monster.com. He has been a speaker on the future of technology for iconic brands such as Chanel, Baccarat and the city of Beverly Hills. While at Accenture, he led the retail e-Commerce practice with clients including Chase, Citi, MasterCard, Visa, MCI, Digex, AT&T, and built the first e-Commerce site for Payless.