BEIERS_1170x120

Big data, big impact: maximizing in-store technology

Print Friendly, PDF & Email

Editor’s Note: In this article, the third in a series on all things digital retail, the authors examine how advanced analytics and machine learning are creating new possibilities for customer-centricity.

Our first article outlined the evolving competitive and consumer landscape driven by Walmart and Amazon. Our second article discussed how food, drug and mass (FDM) retailers are upgrading in-store technology to increase efficiency and offer improved, differentiated in-store experiences that reinforce their value ­proposition.

Since then, Walmart has begun exploring cashierless stores to compete with Amazon Go and is piloting a personal shopping service in New York City. Kroger is expanding its self-checkout program “Scan, Bag, Go” to 400 stores, and JD.com (perhaps the most advanced retailer) has announced plans to open hundreds of cashierless convenience stores. The digital and physical lines continue to move at a rapid pace — reinforcing the need for all retailers to act.

With this third and final article, we turn our attention to big data and machine learning, looking at how retailers can use these capabilities to achieve greater operational impact and faster, more relevant customer-centric offerings.

Making the most of in-store technologies

Eddy Fong

Investments in digital technologies are only meaningful if accompanied by the capabilities to aggregate, make sense of and act on the data the technologies collect. Take beacons, for example — essentially they are just isolated sensors. To provide a complete view of a store, they must be networked with many other sensors to capture big data. But big data is just that: data. Without insights and action, it is useless. This is where advanced analytics and machine learning come into play — software that can aggregate, cleanse and make sense of data in real time to create tangible value.

Forward-thinking retailers are focusing on advanced analytics and machine learning. Consider Shoppermotion: The company works with retailers such as Carre­four, SuperPrix and Aldi to implement an IoT (Internet of Things) solution to track shoppers’ in-store movements. This data is fed through machine-learning algorithms to deliver live insights so retailers can better position in-store displays and promotions and reallocate shelf space.

Big data and an array of other digital technology terms are often used with no clarity about what they are or how they interact. To shed light, we offer the following definitions:

• Big data — Data sets commonly defined as high volume (quantity), velocity (creation speed), and variety (fields and types).

• Artificial intelligence (AI) — The simulation of intelligent behavior by technology: essentially machines thinking like humans.

• Symbolic AI — The classic AI execution, with static algorithms such as sophisticated if–then statements being created to calculate determinations, using human ­expertise.

• Machine learning — The newer AI execution, which uses algorithms to parse data, learn from it, and predict outcomes.

• IoT — The concept of connected devices across platforms and locations, enabling broad data collection, typically to create big data.

Areas of opportunity

Brent Duffin

In today’s new digitally physical landscape, retailers have access to many areas where they can apply big data and machine learning to create a sustainable competitive ­advantage.

Established opportunities

Because data is a prerequisite for analytics and machine learning, retailers have traditionally focused on three areas that offer easier access to it. These areas tend to be the starting points for retailers to establish big data capabilities and create value:

• Inventory management — Early capabilities were typically symbolic AI: human experts creating rules on when to reorder. However, advances in IoT and sensors have enabled big data and machine learning to transform this space by creating dynamic replenishment rules. For example, Blue Yonder is partnering with Natsu and Morrisons to combine internal sources such as point-of-sale data, products, locations and promotions with external sources such as weather, holidays and events to create big data and then forecast demand using AI to provide daily replenishment ­recommendations.

• In-store labor scheduling — While in-store labor benefits from automating routine tasks, several companies also offer smart scheduling to optimize labor by aggregating relevant data sources and creating a schedule tailored to the needs of shoppers, retailers and employees. For example, Percolata is working with 7-Eleven to create big data from in-store traffic, marketing calendars, events and employee availability, which is analyzed to optimize workforce schedules.

• Pricing — Retailers have always had access to pricing, but aggregating and managing this data was extremely difficult. New technologies now enable fast, automated collection and analytical capabilities. Amazon is known for using machine learning to price products. Otto is using Blue Yonder to combine retail data (sales, promotions, products, product presentations and related products) with market data (competitor prices, holidays and events) to produce big data. Using algorithms that evaluate price elasticity, pricing policies and retailers’ objectives (sales versus profits), AI delivers pricing recommendations that can be executed in real time.

Emerging areas

Bryson Waterman

As IoT becomes more pervasive and big data gets richer, new areas for AI and machine learning are emerging. As highlighted in our previous article, two digitally enabled opportunities are especially ­powerful:

• Customer-centricity — Historically, data and analytical constraints have limited retailers to a “macro view” of customer segments. But with big data, retailers can now get insights into an individual customer’s needs to deliver tailored pricing, relevant promotions and personalized assortment recommendations. In addition, granular data and sophisticated techniques enable customer-centric merchandising. For example, Ocado uses machine learning to identify missing store items and can suggest pre-populated grocery baskets based on a shopper’s preferences and recommend alternatives, such as products with less sodium or sugar. Instacart uses machine learning on a massive data set — more than 3 million purchases from more than 200,000 users — to recommend products that users are likely to try or rebuy based on the time of day. This new focus on granular, targeted personalization is also seen in AI-enabled shopper interactions. Ahold’s PeaPod collaborated with Store­Power to create a text-to-order option called Chat-to-Cart, and Fresh­Direct partnered with Mastercard to create collaborative shopping lists and ordering through Facebook Messenger. As AI-powered smart home devices become more popular with consumers, retailers will need to integrate them to reap the benefits. Tesco is already connecting to Alexa through IFTTT, and Walmart, Whole Foods and Walgreens can be accessed through Google Home. As technology evolves, retailers will need to connect with a broader ecosystem, collecting additional data that can be curated into more relevant customer-centric offerings.

• Frictionless interaction — As evidenced by Amazon Go and Walmart’s move toward cashierless stores, computer vision and machine learning are driving the future of frictionless in-store interactions. Frictionless commerce also includes omnichannel. By combining shoppers’ purchasing data with delivery data such as store locations or GPS coordinates and then applying machine learning, Instacart optimizes route deliveries, improving speed and the cost of delivery.

Requirements: big data

Big data is essential to AI and machine learning. Capturing this data requires networked sensors to collect the data; sophisticated algorithms to consolidate, cleanse and index the data; and secure protocols to ensure safe storage of the data. Then data usage requires refined techniques to mine and analyze the information and tools to visualize, export and act on the information.

Implementing AI and machine learning can occur in two ways:

• Partner with third parties to quickly add best-practice capabilities — Think Morrisons with BlueYonder for replenishment, Whole Foods with Conversable for chatbots, Meijer and Food Lion with Eversight for in-store promotions, Loblaw and Foodland with Manthan, and Carre­four with ­AntVoice.

• Build internal capabilities to create a differentiated competitive advantage. Retailers are adding data scientists and machine-learning engineers to their organizations. For example, to enable its cashierless store vision, Walmart is leveraging its in-house WalmartLabs team for machine learning. And Kroger owns 84.51°, formerly the data analytics group of Dunnhumby.

However, it is not necessarily an either/or path. In fact, many retailers blend the two. With the right technology investments in infrastructure, processes to collect and analyze the data, and the people and partnerships to drive action, big data can be a true ­differentiator.

Combine data with insights and action

Investments in digital technologies must include both hard technologies such as beacons, sensors and cameras, and soft technologies such as analytics, big data and machine learning. Only by combining data with insights and action can retailers gain the full value of the technologies they invest in. Although big data has historically focused on inventory, store labor and pricing, advances in AI are creating new possibilities for customer-centricity and frictionless interaction. In this evolving landscape, retailers must evaluate their big-data and machine-learning strategies — or risk falling further behind.

In this series, we have looked at digital retail and its impact on FDM retailers. Walmart and Amazon are upending the competitive and consumer landscape, but the vast majority of consumer shopping will still be in stores. Consequently, FDM retailers will need to reinforce their value proposition by upgrading their in-store digital technologies and big-data and machine-learning capabilities to drive efficiencies and offer a differentiated customer experience. By doing so, they can stay relevant in this evolving ­landscape.

Bryson Waterman is a principal in the Topline Transformation practice of A.T. Kearney, a global strategy and management consulting firm. He focuses on consumer goods and retail and can be reached at Bryson.Waterman@atkearney.com. Brent Duffin is a manager and Eddy Fong is an associate in the firm’s Consumer Goods and Retail practice, and they can be reached at Brent.Duffin@atkearney.com and Eddy.Fong@atkearney.com. In developing this article series, they collaborated with Randy Burt (Randy.Burt@atkearney.com) and Eric Gervet (Eric.Gervet@atkearney.com), partners in the consumer practice.


McClane_728x90

BMG_728x90_10-23-17

Comments are closed.