Many types of sensors may be involved across a retail ecosystem, including RFID on items we buy, Near Field Communication (NFC) and Bluetooth Low-Energy (BLE) on smartphones we possess, location-based technologies—such as RFID, real-time location systems (RTLS), Wi-Fi and GPS—electronic asset protection at the point of sale (POS) and point of exit, bar codes, application-specific sensors and other emerging technologies. The challenge is to optimize data from multiple sensor types, at scale, and realize the potential that the IoT can provide.
Most in-house devices, such as RFID readers, readers for loyalty cards, POS scanners or electronic article surveillance systems, remain disconnected from each other, even when they may share the same network. Because of this, they can’t aggregate data for smarter decision making. Moreover, not all of them are set up to provide real-time alerts at the point of use or contextualized data for further analysis.

Integration for Point-of-Use
Online retailers know what products you browse and research, how much time you spend on a product webpage, what attributes you use to compare one product against another, what’s been sitting in your shopping cart and a plethora of other data points. They likely also know what you or your friends have “liked” or “pinned” on social media. These touch points provide a gold mine of data for finely grained marketing offers, personalized to the individual shopper.
Leveraging RFID, beacons, NFC, BLE and other location technologies enables retailers to use similar data points in the physical world: browsing behavior, dwell times by product category, check-ins and walking paths.
A shopper may check her phone in-store to view what’s on sale today. Without sensor integration, she may see a list of generic promotions. But when sensors are integrated, this same promotional information can be integrated with real-time inventory and customer-loyalty data, so she is served up offers based on her purchasing history and the items that are physically in-stock at that store at that time—perhaps her favorite brand, style and size of jeans. Compelling, personalized, relevant offers like these are most likely to convert to a sale.
Integration for Future Analysis
Sensor data collected in-store can also be analyzed to generate predicted buying behaviors and patterns based on a shopper’s history, including search, discover, purchasing and returning items. This predictive analysis goes beyond traditional demographics to indicate which shoppers have a higher propensity to buy certain items, and how goods should be merchandised to maximize sell-through. For corporate planning purposes, this data can help predict the best place to locate a new store or distribution center, product assortments by region or optimal distribution networks.
Building a Data Foundation
As devices proliferate, retailers need to put an IoT foundation in place to manage high volumes of sensor data. In order to provide availability of operational data at the point of use, and to capture enterprise data for future analytics, retailers need to revisit their data-warehousing strategy, consider a standards-based approach to sharing information with trading partners, consider a single platform for inter-device communication across multiple modalities and open application programming interfaces (APIs) to correlate data between different sources.
Leveraging the Data
To make massive amounts of sensor data from multiple sources useful, the information must be managed and correlated. As such, retailers should consider where and how to store the data, and how to make correlations between different sources of inferences with data sets gathered from multiple IoT devices, processes and individuals. They also need to apply business rules to obtain useful analytics in order to differentiate correlation from causality.
For example, if an apparel retailer sought to leverage IoT data to improve shopper conversion, it would need to create structured data feeds that capture the number of people who pass through the door, as well as the number of shoppers who enter the fitting room, try on items and then purchase those products at the point of sale, in order to create an ongoing flow of usable data for further analysis. Then the company would create modular data sets to enable the integration of this disparate data.
Taking the Next Step
Much is possible when devices, processes and people are integrated to form a complete picture. Today, loyal shoppers are receiving special offers on their smartphones after communicating with beacons, while manufacturers and distributors are receiving alarms at dock doors when outgoing shipments are incorrect, based on RFID readings, for example.
The next step is to use standards, open APIs and integrated data sourced from different sensors, and to then set both operational and analytical objectives. In that way, retailers can make better immediate and longer-term decisions that lead to increased revenue and improved margins.
Anurag Nagpal is the director of RFID solutions at Checkpoint Systems.