Really Want to Stop Shrink at the Point of Sale (POS)? Rethink What a ‘Scan’ Should Look Like – and Look For.

There are a lot of ways to reduce shrink at the point of sale (POS) without degrading the customer experience. There are also many ways to stop losses stemming from accidental misreads or mis-rings by associates as they’re hurrying to queue bust, do inventory, or restock shelves. All of them start with having the proper barcode scanner.

I know you probably think you’re set with your scanners. But are you really? Are you sure you have the right scanners for the job?

I ask because one of the easiest ways to stop both malicious acts and mistakes in your stores, and particularly at the POS, is to take advantage of things such as 2D barcodes, cameras embedded in bioptic scanners, hyper-sensitive scales, and other loss detection technologies that are only available in newer scanners.

In fact, I was just telling colleagues the other day that the biggest issue I see right now in retail is that stores are overstocked with technology that’s quickly becoming obsolete, especially scanners.

Store associates are trying to manage shelf inventory, shelf tags, pricing markdowns, and more with consumer smartphones that must be pointed perfectly at a 1D barcode to get a read on the data. And both associates and shoppers are having to work around the issues caused by using older single plane and bioptic scanners that don’t have camera technology – or have really slow, poorly performing 1D scanners and cameras – at checkout lanes.

We need to work together to fix this problem ASAP, especially at the POS.

Whilst we may think of them as simply “scanners”, there is a big difference from imager-based scanners (using elements of camera technologies) and linear-based laser scanners. Imager-based scanners can extract data from 2D barcodes – such as data matrixes that don’t require straight-on aim – which will help with productivity, and they will naturally replace 1D barcode scanners that age out. (Check out the latest progress with 2D barcode technology here.)

However, scanners that ‘see’ more than just a barcode and can identify items better than people are going to be necessary at the POS moving forward, particularly in light of the GS1 project Sunrise 2027. This is additionally true given so many shoppers now expect self-checkout (SCO) options, which need to be simple to use.

You need to be totally comfortable opening more SCO lanes rather than feeling compelled to retreat due to negative feedback or concerns over misuse. As my colleague mentioned a few months back, ‘Now Is Not the Time to Retreat from SCO’.

So, how do we get you there? How do we build trust among your operations, leadership, and asset protection teams that both associates and shoppers are doing the right thing and that potential losses can be stopped at the POS?

Train Your Scanner to ‘See Something, Say Something’

There’s a brilliant (and quite extensive) report that was published in ECR Loss Design that is worth the read if you’re serious about stopping intentional or accidental losses specifically from an SCO perspective:

SELF-CHECKOUT LOSS: THREE WAYS TO RETHINK SCO DESIGN

The full report features ‘Key Insights, Three Ways to Rethink SCO Design and 20 Design Concepts’. However, the main takeaway that pulled through for me was that you must take steps to ‘improve scan accuracy’ at all checkout lanes, no matter who’s doing the swiping.

This goes back to my earlier question about whether you’re using the right scanners at the POS. Not just SCO lanes, but all checkout lanes.

Your associates are just as likely as shoppers to make an innocent mistake or be involved in fraud or theft activity. So, if you really want to stop losses, then you must have mechanisms to both detect the loss and act on it in real time at every register. You can’t do that if they’re only equipped with handheld scanners, single plane, or barcode scanners only capable of looking at 1D barcodes.

These types of 1D barcode scanners can’t…

• extract product information from every barcode on every item. (That’s why so many items are being picked manually – and incorrectly – by associates and customers.)
• identify a product or read a label even when there isn’t a barcode to scan. (It’s why a lot of losses are tied to produce, bakery, deli, meat market, seafood, or bulk items.)
• detect ticket switching. (Even if you have well-calibrated scales attached to your scanners, a store-brand 9V battery is going to be packaged similarly and weigh the same as a name-brand 9V battery. Scales won’t catch mismatches related to organic vs. conventional produce, either.)
• tell you if the scanner is overly sensitive, not sensitive enough, or working perfectly. If the system detects an item on the scale and all it is capable of looking for is a barcode, then the only thing you’re going to know for sure is that there’s a weighted item sitting on that scale. You won’t know if it’s a customer’s bag, wallet or phone; the handheld scanner that was left on the scanner; or if it’s a product that the customer intends to buy unless someone walks over there to visually inspect the scale and help the customer clear the item off there.

That’s why you’ll always struggle with loss detection and loss prevention until you install POS scanners that are not only imager based but also have built-in vision systems and advanced intelligence technology that can be trained to recognise what is right in front of its ‘eyes’.

With the right type of scanner at the POS – the ones that pair 1D/2D barcode scanners and AI-based computer vision systems – you will…

• take the guesswork out of selecting the right SKU either during label creation or checkout. If you have scanners that can visually recognise an item, you don’t put the onus on an associate or customer to guess what that item is as they’re scrolling through the pick list at the label-making station or checkout lane. You also reduce the odds that they’ll just pick something that looks ‘close enough’ because they’re either in a hurry and don’t want to scroll through a pick list or they don’t really know how to describe the item. If you have scanners that incorporate high-quality cameras and machine learning algorithms (i.e., AI), then you can be more confident that there isn’t any funny business happening at the labeling station or checkout lane, either intentionally or accidentally. You can see – the scanner can see – exactly what’s happening. If someone selects the picture for a plain donut but it’s really a fancy donut in the bag (which may be more expensive), the intelligence within the vision system can flag it to the associate.
• confirm that the item sitting on the scanner or in the bag matches what the barcode scan reported it to be. It won’t just accept the scanned barcode information or on-screen pick list selections as accurate. This helps to prevent those ‘mislabeled item’ scenarios as well as ticket switching. If the item that crossed in front of the camera and barcode scanner don’t match up with the item details put into the system, an associate can be alerted to come see what’s going on (at SCO) or the transaction can be halted until a manager comes to verify an associate’s action.

So, in my opinion, you can’t worry so much about what you see on an end-of-day or end-of quarter loss report or what an associate reports seeing at their registers every day. You need to worry about what you (and your associates) are not seeing…what the barcode scan or human eye is missing. You need a POS system that can be trained to say something when it quite literally sees that something is off, whether it’s the weight, quantity, product identity or something else that is questionable or confirmed to be a mismatch between the system input and the item put into the shopper’s bag.

What Does a ‘Good’ POS Scanner Look Like?

Now that we’re on the same page about why you need a visual recognition mechanism embedded in your POS scanners and how it can help you stop losses, let’s talk about what you need to look for when shopping around for this type of system. (The last thing I want is for you to spend a ton of money and effort replacing your 1D barcode-only scanners with newer scanners marketed as having visual recognition capabilities only to realise they don’t work as well as you need.)

First things first: you and I both know the average camera isn’t going to be able to distinguish a Gala apple from a Honeycrisp on its own. (Have you ever tried to use one of those plant, bird, or bug identifiers on your smartphone and received a list of a dozen species it might be?) There’s a lot that goes into designing a camera system that can accurately recognise anything and everything put in front of it.

Therefore, when looking at how well a vision system will work for retail use – whether it will be able to instantly recognise thousands of random items – you must consider:

• the quality of the lens
• lighting
• the level of intelligence in the image processing software within (or connected) to the camera

Multiple images of an item must be captured in a split second to create a 3D-type rendition that enables the (AI) brains of the system to figure out the item’s form, size, etc. The lighting must also be right and the lens quality good enough to see the item’s color, markings, and other distinctly unique (and possibly quite small) identifying details. It’s the only way the intelligent image processor can confirm which exact item is sitting on the scanner from that list of potential items.

Of course, the only way you’ll confidently know whether the scanner you’re looking at will be able to accurately collect and analyse all barcode and visual data alongside scale data is if you test it in real-world conditions. I can show you a demo at a retail expo or in our labs, but you really need to try it in your stores. Your IT and operations teams both need to understand how it works and how much work it will take them to get these newer scanners online and generating the results you want.

I’d imagine you’d also want to see what other impacts they’ll have on your operations. For example, is there a tradeoff by trading up in terms of sustainability goals? Will it be easier or more difficult to develop software for the scanner and vision system, integrate it with other POS system components, or maintain the scanners, cameras and scales?

I’ve worked with retailers who were worried that switching to newer POS scanners with built-in vision systems and AI would pull more power or otherwise hinder their ability to meet environmental goals. Depending on which scanning systems are selected, that is quite possible. However, I can tell you that there are multi-plane scanners on the market today that have the latest and greatest computer vision technology built into them to support shrink reduction efforts and have the lowest power consumption rating in the industry.  

Many of these more intelligent, vision-equipped scanners are also built…

• without the harmful chemicals and materials commonly found in other POS scanners.
• to last longer than other types of POS scanning systems, so you don’t have to replace them as frequently.
• to be recyclable when they do reach end-of-life many years from now (also reducing landfill waste).
• for easy servicing (to reduce the environmental and financial cost of frequent failures and replacements).
• to facilitate automatic performance improvements (via software updates and machine learning algorithms that continuously improve item recognition capabilities and other POS outcomes).

So, if you’re ready to get a handle on POS-related losses, I strongly urge you to check out the scanners that incorporate intelligent vision systems alongside data captured from the scales and 1D/2D barcode scans. They are the key to more accurate item identification at the POS to reduce the number of associate or manager interventions. They’re also the best way to ensure immediate intervention occurs when mismatches, ticket switching and other issues are suspected.

Perhaps start your research by watching this video: 

You may also want to check out this article:

Scan and Deliver

I can also put you in touch with retailers who are already using these intelligent POS systems with advanced vision and scanning technology. Many have told my colleagues and me how pleased they’ve been with how quickly and accurately the systems identify items and how much friction they seem to be removing at checkout.

While thrilled to hear they’re seeing such great results, I’m not totally surprised. I know how happy a good camera can make us – how much it can improve our lives. (It’s why people upgrade to the next-generation mobile device only because it has a better camera.) So, I’m hoping that many of the frustrations felt by shoppers, associates, and retail leaders these past few decades at the POS will finally go away once more scanners currently used at the POS or even label-printing stations throughout the store are switched out to the ones with cameras. It will be interesting to see just how much (revenue) you’ve been missing over the years.

###

Related Insights:

• Why – and How – Walgreens is Removing Friction to Stop Revenue from Walking Out the Door
• Shoppers Say It’s Easy to Swipe Items at Self-Checkout. Here’s How Retailers Should Respond.