Self-checkout: balancing cost efficiency with customer experience

Retail’s love-hate relationship with self-checkouts

Self-checkouts could be seen as the modern marmite / vegemite – either loved or hated, they make the news and instigate wide debate. First introduced in 1986, self-checkouts were designed to reduce the wait times for customers, increase efficiency and deliver cost savings for retailers.

The heated debate is like that of motorways and highways. Many hate them as they get congested and they’re boring to drive, yet without them it takes much longer to get from destination to destination. The argument for self-checkouts is similar. By the time customers place the last item in their basket or shopping trolley, they are mentally changing gears to what happens after they exit the store. They have a new destination in mind, and the checkout is a congestion they must get through to reach it.

Implementation challenges and successes

Like any new technology implementation, the introduction of self-checkouts has brought both challenges and successes. Rising retail shrinkage rates, for example, are often linked in headlines to self-checkouts, with ‘20% of people who used self-checkouts saying they accidentally took an item without paying for it‘ often quoted while the fact 80% of shoppers when prompted will self-correct mistakes is overlooked! So, are self-checkouts on borrowed time, or should we expect them to stay, but with changes?

Retail giants such as Walmart are removing some self-checkout lanes, or tying their use to paid for membership schemes. Target is also trialling changes to in-store processes, with a 10-items-or-fewer policy for self-checkout use which they say is improving efficiency as it has both reduced checkout times, and curbed missing inventory, with Dollar General and Schnucks following suit.

Are self-checkouts more efficient?

Self-checkouts offer retailers opportunities to improve both efficiency and customer experience in store. Recent market data demonstrates that self-checkouts offer valuable efficiencies in retail – reducing checkout times by up to 50%, with the ability to manage up to 50% more transactions compared to traditional manned checkout lanes. The key, as many retailers have found, is striking the right balance.

Some retailers are reaping the rewards of self-checkout efficiency, for example Sainsbury’s recently went on record to declare that their customers love the self-checkout, and their successful rollout has been achieved by offering self-checkout options while maintaining traditional checkout methods. Also in the UK, Booths a high-end supermarket chain, has revised their self-checkout strategy based on customer feedback, recognising that their customer base valued the face-to-face interactions and opted to utilise self-checkouts only in their busiest tourist footfall stores.

This is further supported by recent Newsweek polling on the sentiment behind self-checkout usage, which found that 62% of US respondents reported not liking the technology because it takes away another person’s job, and 40% dislike self-checkout because they prefer to speak to a person.

The key to achieving the best cost efficiency and customer experience lies in having a workforce augmented rather than replaced by AI. According to SeeChange CEO, Jason Souloglou, “It’s a misunderstanding that AI and other tech means that stores can hire fewer workers. AI frees up employees to conduct higher-value tasks that improve customer satisfaction.”

This sentiment is echoed by Gerry Hough, of McKinsey,  “Retailers are seeing the need for a multipronged strategy. AI is a part of every box, but it is not the same in every box. We are seeing modification and stratification.”

Lessons from frictionless store experiments

The journey towards frictionless stores offers a compelling vision of seamless shopping, yet it presents significant challenges in implementation. The news of Amazon stepping back from just-walk-out technology demonstrates some of the hurdles faced when introducing customers to new shopping methods. The futuristic technology relied heavily on human oversight, with limited ways to improve efficiency, raising concerns about the real-time aspect of transactions, billing accuracy and unsustainable labour costs.

Deloitte suggest the just-walk-out model has not achieved widespread adoption due to the substantial initial investments, and uncertainty among customers about the shopping experience. It was also noted by Forbes that the retailer had failed to first capture a grocery audience before transitioning them to a frictionless model, resulting in a lack of motivation for shoppers to change their habits.

A sharp contrast to this is seen in stadiums and arenas, where just-walk-out stores are enjoying remarkable success. Fans rate it as the top initiative in 2024, appreciating quick access to food and drink which allows them to return to their seats promptly.

Good tech implemented badly: unexpected item in the bagging area

Self-checkouts offer lessons in navigating the complexities of technology implementation in stores at scale. While designed to enhance efficiency and reduce costs, self-checkouts can introduce friction instead.

The phrase “unexpected item in the bagging area” has made it the punchline to many a joke, yet the issue isn’t with the security scales themselves rather how they’re integrated into the self-checkout process. While the scale serves to confirm the expected weight of products, its effectiveness relies on accurate SKU data and predetermined weight parameters along with on-going accurate calibration, such as when customers use their own shopping bags.

Focus therefore has to go beyond the technology itself to the implementation process, with the aim of maximizing optimal functionality and enhancing customer satisfaction.

Computer vision-powered self-checkout: delivering on the customer promise

Computer vision is the use of artificial intelligence (AI) and image processing techniques to analyse visual data from cameras installed, in this context, at self-checkouts. Together with input from other sensors such as barcode scanners and weigh scales, computer vision adds intelligence to the self-checkout process so that errors such as mis scans can be prompted and resolved, while identifying fresh produce on scales can be automated.

For the shopper whether checking out at a manned or self-checkout the experience should be positive. For retailers looking to adopt computer vision-driven solutions, there are three key considerations:

Reducing interventions
Errors, such as ‘unexpected item in the bagging area’, can be vastly reduced, lowering the number of times an employee is needed to assist a customer. This not only improves the customer experience, but it also alleviates stress for employees who often struggle to manage their workload during peak times.

By minimising interventions, retailers enhance customer experience, gain a happier workforce, and foster greater employee advocacy for the technology.

Your rules, your way
Every store is unique, and no store remains the same throughout a day. You need to be able to adapt to the varying demands of your store in real-time according to configurable rules.

What may be effective during quiet periods may not at peak times. At certain points in the day your aim may be to reduce shrinkage, at others it’s to shorten queues. You need control over this balance to enable dynamic adjustments to overall system sensitivity to shrink so that interventions at self-checkout are in line with your store’s operations and customer needs.

Intelligence that shoppers love

Let’s take the example of searching for a ‘fresh item’ at checkout. Typically, it takes a customer between 10-15 seconds to select a single non-barcoded item from the picklist. This leads to 3 in 10 customers avoiding self-checkouts due to perceived complexity and its time-consuming nature.

In contrast, an AI-powered solution such as fresh produce recognition eliminates the need for manual searches through on-screen catalogues, reducing identification time down to just 0.5s, and the end-to-end time to recognise and add items to the till roll to just 6 seconds. This delivers significant time savings – up to 40% increase in transactions per hour – as employee interventions for struggling customers are minimised. More importantly, 92% of shoppers when given the choice, opt for the new AI-powered approach over manual look-up.

And why stop at self-checkout? The same recognition system can be deployed on scales in the aisles to deliver a consistent experience for customers and a cohesive, connected approach for retailers.

What’s next for self-checkout?

So, what’s next for the self-checkout? The same survey that reported fraud at self-checkout also highlighted that 41% of Americans almost always use self-checkout if it’s an option and in the UK, the Association of Convenience Stores recently reported that while only 16% of local stores currently offer self-service the ACS expect the number of self-checkouts to double over the next 5 years.

By leveraging computer vision and AI sensitively, retailers can navigate the self-checkout landscape efficiently and effectively, fostering a seamless shopping experience while minimising shrinkage concerns.

Join the discussion on computer vision AI

Discover how computer vision AI can empower retailers to combat shrink and enhance customer experience in our Intel webcast.