Electronic Shelf Labels have been in production in Retail Stores since the 1990s – they are not a new technology. First widely adopted in Europe as a means to automate price changes at the shelf edge and allow associates to spend more time with shoppers, the first wave of ESL screens were segmented (think Casio Digital Watch) and displayed only numbers, single letters or symbols.
The current generations of Electronic Shelf Labels are fully graphical (think Apple Watch or a TV screen) and use either LCD or increasingly ePaper screens. AS Electronic Shelf Labels have evolved, they have become more functional and less costly, and in corresponding fashion, retailers have expanded the ways in which the technology is used.
The earliest generations of Electronic Shelf Labels used either Infra-Red communication protocols or sub 1G radio systems which required wired antennas to be run throughout the store. At the time, these were the only options available – back then, cellular technology ran on 1 and 2g networks and data was rarely transmitted over cellular networks, SMS was barely used, Wi-Fi networks were a relatively new invention and data rates were measured in megabits not gigabits per second.
Between the 1990s and 2020s much has changed in wireless technology. The cellphone in the 1990s was used for voice only, today: Smartphones are the most widely used computing devices, they are our web browser of choice, we run applications, watch shows and movies, listen to music, send email and shoot movies with them. If we tried to run all of these applications on the 1990s 2g network, we would crash it in seconds.
The use of ESLs has changed in very similar ways to the way we use our cellphone. Back then – ESLs were segmented, and the amount of data transmitted to each label was relatively small, rather than a fully graphical image, simple metadata was transmitted over the early Infra-Red and Wireless networks. The number to be displayed on each segment was sent to the tag, relatively small amounts of data and even the basic network protocols could transmit it quite well.
Today’s Electronic Shelf Labels are fully graphical and now Retailers transmit multiple image files over the network – think about the difference in the size of a text file with three or four lines of text versus a JPEG file. Advanced users of Electronic Shelf Labels use up to seven different full screens of information on the same label with information that may change multiple times each day. Users globally use the screens on their Electronic Shelf Labels to convey information to shoppers and associates concerning price, promotions, inventory, enriched product information, allergy and food preference information, reviews and many other kinds of information. Today’s Electronic Shelf Labels use Near Field Communication Technology to enable connectivity to shopper and associate mobile devices and enable payment solutions at the shelf edge. In short, the growth in use cases for Electronic Shelf Labels has grown as exponentially as the ways in which we use our smartphones.
It would, then, be reasonable to expect that the networks which support Electronic Shelf Labels would have evolved in the same way that cellular networks have evolved to support cellphones – the way wireless networks have evolved to support computers and tablets – in many cases not!
Broadly speaking, the networks used for modern Electronic Shelf Labels fall into three categories, Infra-Red (Pricer, Eldat), Guaranteed Time Slot Zigbee or Modified Zigbee (Vendors in this category claim they don’t use it) and Spread Spectrum, Synchrous, Time Intelligent (Store Intelligence Spectra).
Back in the 1990s and early 2000s, Infra-Red was the most popular system for ESL’s. IR was also used in TV Remote Controls and PDAs (does anyone remember the Compaq iPaq?). The Remote Control is still the most popular use case for Infra-Red – sadly the iPaq suffered the same fate as the Dodo!
Modern applications of the IR technology to electronic Shelf Labels can be successful, but there are notable tradeoffs. On average, IR networks require an access point for every 1500 Sq. Ft of floorspace, sometimes more if there are multiple obstructions to the line of sight which cause problems, this would mean in a Super Center type store with 140,000 square feet, you would need 90-100 access points. For retailers prepared to invest in this much infrastructure – one-way data transmission is relatively speedy and works well. AS a result, some retailers who had already installed Infra-Red infrastructure to support Segmented Electronic Shelf Labels enjoyed some success in transmitting price data to fully graphical shelf labels over the same IR network. Fewer retailers are prepared to make the huge investment to install all of these access points, but when they do so, the tradeoff is in bi-directional communication, IR networks are somewhat effective in sending data from the network access point to the client (label) but the return journey is a lot more problematic. IR networks are only capable of simple, binary communications back to the infrastructure in the form of a ping – a 0 or 1 to send a simple yes or no back to the network. This becomes problematic as use cases become more sophisticated as shoppers and associates interact with the ESL directly. So even if you are prepared to pay the high price to build the infrastructure highway, the traffic only goes in one direction.
Guaranteed Time Slot Zigbee or Modified Zigbee
There have undoubtedly been some amazing developments in Wi-Fi technology since the 1990s. Speeds are now measured in Gigabits per second not Megabits – however the order of magnitude data speed increases have been achieved in 802.11 protocols largely designed to support persistent connections between the access point and up to 25 clients. Most network access points advertise a recommended maximum of 25 clients – typically on these kinds of networks, the clients are connected all of the time and if not connected, looking for a connection. This type of connectivity is called “asynchrous”, meaning simply that clients are looking to connect to the network all of the time not at selected times. If you’ve ever attended a conference in the basement of a building where there is little or no cell service, you’ve probably noticed that your smartphone battery runs down very quickly, this is because searching for a network to connect to uses a lot more power than staying connected to the network.
Zigbee in its various types was an attempt to create a wireless networking protocol that could support thousands of clients but sending and receiving much less data less frequently than typical 802.11 networks. Modified Zigbee or guaranteed time slot ZigBee networks function by allowing any number of the clients on the network to attempt to connect but then using a traffic cop like process to decide which of all of the contending clients will get 7 guaranteed slots to send and receive their data. In reality, almost all modern ESL vendors who are not using IR use some flavor of this Zigbee protocol. As a result, the published limits of their capacity in terms of tags per access point and area covered are largely theoretical, in practice for highly dense grocery environments, with high update frequencies, we see customers deploying one access point per 5,000 tags or so.
Spread Spectrum, Synchronous, Time Intelligent (Store Intelligence Spectra)
The Store Intelligence Spectra Platform solves all of the issues inherent in other solutions for dense environments, frequent updates of multiple screens, bi-directional communication and light infrastructure.
The Spectra Platform uses the full spread spectrum of the 2.4 GHZ WiFi capability but in Synchrous mode across the full WiFi spectrum.
Whereas Zigbee Solutions use almost than half of its 16 available time slots to decide which client should receive data and which should not, the Spectra Platform uses all 1024 time slots for data transmission and hops intelligently between 79 available frequencies in the spread spectrum, as a result no single channel is ever overcrowded and Spectra does not cause any significant interference with the other Wi-Fi applications running in a retail store.
When you compare the Spread Spectrum, Synchronous, Time Intelligent approach versus Zigbee or Modified Zigbee and Infra-Red, the comparison is stark; the numbers listed below are based on experience in real retail stores versus marketing claims made by various vendors:
Installing greater amounts of infrastructure is not only costly during the initial deployment, hardware has to be maintained, and typically, large retailers calculate that the cost of supporting hardware is 3x the cost of the hardware to install and another 3x to maintain over it’s useful life, so if an Access Point costs $1,000 on average, the cost to deploy infrastructure only in the largest format store is $42,000 for a Spectra Solution, $665,000 for Infra-Red (16x) and $140,000 for most Zigbee Solutions (3.4X).
It would seem that choosing a platform for ESLs is a relatively easy decision! However, the main challenge in doing so is that retailers evaluating these solutions typically think in terms of the volume of price changes and activity they had with a paper system, in this comparison, the Spectra Platform outperforms the alternatives, but once you move into a world where the retailer can drive information to shelf-edge whenever they want to, the usage of the system increases exponentially. In Europe where ESLs have been deployed for a decade, retailers see a 10x and greater increase in the volume of price changes when they move from a paper to an electronic system, or as one retailer says “When you have a Ferrari, you drive it, and you drive it fast!”.
Dr. Ken Cioffi, SVP and Chief Architect, Store Intelligence.
Rob Crane, President and CRO, Store Intelligence.