SAP Retail Innovation Lab on an IoT Mission

Great article from the SAP retail team helping to bridge the gap between the theoretical and the possible. 

The SAP Retail Innovation Lab is using design thinking, quick-start implementations, and a miniature model of an IoT-enabled store to promote its IoT innovations for the retail sector. We take a closer look at the current new approaches in this field and at the obstacles to their adoption in the retail industry.

Oliver Reulmann, SAP Strategic Archetect for Retail Solutions, is on an IoT mission. He travels to trade shows and conferences to present the “IoT Store on Wheels,” a sensor-equipped, connected model of a retail store. The sensors in the model send information to the cloud about every movement a shopping cart or product makes within the store. This data is analyzed on an IoT platform based on SAP HANA Cloud Platform to provide new insight into shoppers’ buying behavior.

“Our aim is to make the technologies of the future attractive to retailers,” says Reulmann, a member of SAP’s Digital Business Service unit. “And what could be better than a hands-on model that lets you run analyses yourself.”

IoT for Retail: The Main Hurdles

Reulmann and his colleagues have a fair bit of convincing to do. Because, as Michael Osthof and Hendrik Hilger, who together head up the SAP Retail Innovation Lab in St. Ingbert, Germany, explain, “There are still hurdles to overcome before the retail industry is ready to embrace its technological future with both hands.”

1. Spot solutions

Many retail companies’ approaches to the IoT center around spot solutions supplied, say, by small start-ups that address very specific aspects of the sales process.

“But how does it benefit a company to analyze visitor streams if it can’t synchronize that information with current revenue data?” asks business information engineer Hilger. “Even now, the business context is still too often ignored completely or not taken into account soon enough.” 

2. Too many technologies

“There are plenty of technologies out there that support IoT scenarios. But there’s no single technology that can do everything,” says SAP specialist Osthof.

For example, when it comes to monitoring in-store visitor streams, you can use Wi-Fi sensors to anonymously track customer movements, analyze data from video cameras, or deploy Bluetooth beacons. Each of these technologies has its own particular strengths and weaknesses.

What the SAP HANA Cloud Platform can do is enable retailers to combine data from each for analysis. And, crucially, the consolidated insight from these different sources can be linked to data and figures from ERP, inventory management, logistics, and finance. “We unite the “as-is” world with the business context,” explains Osthof.

3. Data protection

In a pilot project at a new shopping mall, visitors’ movements are captured by a video camera and displayed in the form of a heat map – providing an anonymized analysis that will be used to add value for customers. Of course, complying with data protection regulations is essential when conducting analyses of this kind, and these can vary considerably from country to country.

“It’s important to be open from the start about what you’re using data for and what added value it has,” says Reulmann.

You also have to bear in mind that there are often considerable discrepancies between what is legally permitted and what society is willing to accept.

For example, 25 students taking part in a recent hackathon came up with the idea of a smart fitting room in which customers could take a picture of themselves in a new outfit they had chosen and have their picture displayed in the store window. “The customers who took part were clearly not put off by the idea of acting as ‘advertising icons’,” says Osthof.

4. Rival projects

“Omnichannel” ‒ the business model that focuses on a consolidated view of online, mobile, and in-store shopping ‒ still occupies the top spot on many retail companies’ agendas. “This means that many retailers currently still lack the motivation to address the emerging topic of the Internet of Things,” says Osthof.

“Quick-Start Implementation” Dispels Fear of Large, Costly IoT Project

Getting started with the IoT is often not as complicated as many assume. “In our experience, with a quick-start implementation, you begin to see the initial success of deploying IoT projects in as little as eight to ten weeks,” says Hendrik Hilger, one of the SAP specialists* working on new technological approaches for the retail industry. The procedure is straightforward: Use design thinking to specify requirements, build a prototype, address the question of linking to the IT infrastructure, and develop version 1.0 of the IoT application on the SAP HANA Cloud Platform.

“An entry-level project is predictable in length and does not require extended planning periods,” says Hilger, who lists the preconfigured IoT platform as one of the aspects that enable user-centric approaches to be implemented relatively quickly.

Three Compelling Examples of IoT in Retail

1. Clothing Life Cycle

A major international fashion retailer that followed a procedure similar to the one described above began working with an “innovation stream” for SAP HANA Cloud Platform after just 10 weeks. The main advantage of this approach for the retailer in question, as Osthof explains, was that it did not initially require a link to the inventory management system.

The initiative involved fitting RFID chips to items of clothing in order to track them through each stage of their life cycle – from their manufacture in the Far East, to their storage and distribution, to their arrival in store. Once the items of clothing have reached their destination, the so-called “micro-movements” they make continue to provide vital information. For example, they show the retailer which items of clothing customers frequently try on but rarely buy.

A major benefit of this initiative from the customer perspective is transparency. “It’s vital to give consumers the option of finding out exactly where the item of clothing they are considering buying was made,” explains Osthof.

2. Machine Learning in Reverse Vending Machines

A grocery retailer recently began using algorithms to predict fill levels in a reverse vending machine for recyclable bottles. The machine identifies when customers come to the machine and correlates this information with the fill level to accurately predict when human intervention will be required. The system records the times and days when the machine is particularly busy and the periods during which very few customers come by to return their deposit bottles. Shortly before the point at which the machine has predicted it will be full, it sends a message to an employee at the store to notify him or her that it needs emptying. If customer behavior changes, the system adapts accordingly. In short, the machine learns as it goes along.

3. Biometric Advertising

Currently in the early stages of development is a new approach to personalizing store-based advertising. Osthof explains the idea behind it as follows: “When, say, a young person enters a store, a screen positioned a few meters further along his or her path automatically displays an advertisement aimed at teenagers.” A video camera – so the idea goes – makes an assumption about the customer’s age and type so that the retailer can show a commercial that might appeal to that particular customer.

“The concept still needs some fine-tuning,” admits Hilger – particularly now that a retail pet store chain has expressed interest in the solution. In this scenario, if a Golden Retriever enters the pet store with its owner, the camera will need to be able to distinguish the dog from a Dwarf Poodle, Wire-Haired Dachshund, or other breed – so that the correct dog food commercial appears on the advertising screen at the right moment.

The following presentation from SAP is also a useful reference deck.