Using data to strike a balance between company and customer needs

Across Australia, retail theft is on the rise and, due to their narrow profit margins, supermarkets have been particularly hard hit. In the year to June 2024, Australia’s most populous state, New South Wales, recorded a 47.5 percent increase in theft from retail stores, including supermarkets.

As one of the country’s two biggest supermarket chains, Coles is not immune to the problem and when the software vendor that provides its self-checkout technology offered a new way to deter thieves, management decided to try something new.

Major retailers across the country including Coles added front-facing “cameras” to self-checkouts which showed the customer an image of their own face as they scanned goods. They weren’t the first to implement the technology – it is used by major grocery retailers around the world because it is assumed that increasing surveillance will deter theft.

It’s important to note that the technology was a “live stream” of activity and was not recording. However, from the quality of the picture to viral social media claims that facial recognition technology was being used for advertising purposes, customer sentiment around self-checkout cameras was largely negative. Furthermore, for Coles, the use of the cameras was changing how its customers interacted at the tills.

“When we implemented the checkout cameras there were a lot of customers who immediately did not like it, but we weren’t able to see the impact that it was having on various functions on the self-checkouts,” says national CX manager for Coles, Claire Cunningham, who is also MD and lead consultant for The Customer Connexion.

Cunningham knew there was a customer sentiment problem to investigate and sought approval to investigate further by looking at data from across the organization to prove her hypothesis. By collecting and combining multiple data sets she was able to disprove the efficacy of a new technology that was being introduced in stores by retailers around the world.

“Theft is a big problem for retailers across Australia, so it was never going to be a case of ‘people don’t like this, let’s turn it off’. We needed to understand a whole host of different things to prove if the technology was doing what we intended it to do, and if it was working, did that outweigh the customer sentiment,” Cunningham says.

Once Cunningham started to investigate the available data, she found the “cameras” were not only not deterring theft, but they were negatively affecting the customer experience and the speed of the self checkouts. 

“We went through various trials and tests, we collected a lot of different data, and we did a lot of different research to show that when the camera is off, not only did customers like that better and attempted theft was unaffected, but it also actually improved the function of the self-checkout,” she explains. “Having all those different components really brought us to the end decision of not just turning this off for customers but showing that actually, this technology doesn't have the capability everyone thought it did.”

Meeting the needs of the customer and organization

Cunningham’s project draws on many of CX’s biggest pain points, from balancing the needs of both customer and company, to securing the required top-level buy-in to get such a controversial – and significant – project off the ground.

“There were two parts to getting the buy-in: we recognized right away that this was an intense pain point that every customer experienced. Everybody uses self-checkouts, not only customers, but also our staff do their grocery shopping here and nobody liked it. Having that anchor point of a shared experience really helped to get support. The other part was canvassing stakeholders and not only finding out their opinions but also asking what data points they would need to see to evidence the need for the change. In short, what would convince them that this is not working in the intended way,” Cunningham explains.

The project touched on another major CX pain point: the need to humanize automation and self-service technology.

“Every person shops for groceries and they all want a different type of experience. A retailer can’t say ‘this is who we are’ and that’s the end, a supermarket really does have to be everything to everyone,” Cunningham says. “The core root of humanizing something, particularly something like self-checkouts, is trying to make it as easy as possible for your customer base and if that customer base is really broad, you have to give them the options that cater to their needs.”

Finding numbers that speak to the human experience

For many in CX, the idea that it is possible for a practitioner to secure buy-in at the highest level while simultaneously advocating for the customer is ambitious – to achieve such a feat with data is nothing short of a dream.

CX Network’s 2025 research into the Global State of CX found that data pose challenges for practitioners in all markets. From the ability to extract actionable insights to siloed and insufficient customer data, practitioners are largely failing to embrace the full potential of the numbers available to them.

Cunningham holds a bachelors degree in business administration, but is not a trained data analyst. However, by working in customer-facing and CX-related roles she has gained a strong enough idea of customer sentiment to understand the key drivers of the human experience. Furthermore, she knows the answers can be found in data, if the right data are interrogated in the right way.

“People don’t always understand what data can speak to in terms of customer behavior,” she explains.

“Sales data, website data such as time spent on page and bounce rate, technically these things are not customer sentiment data, but they do prove customer behavior and if practitioners are a little more creative in terms of what they interpret to be customer data, there are endless insights to be discovered.

“We can get a little too narrow minded sometimes, thinking that if we haven’t directly asked the customer about their feelings on something, we don’t know how they feel. But we do know because they didn’t buy the item, something in the experience or journey deterred them,” she adds.

Cunningham says the mindset required to collate, analyze and utilize customer data in this way demands both creativity and curiosity. “It’s a case of asking what touch point does the customer have with us and what are we measuring there, then really finding a way to turn that into a story,” she explains.

You can hear more from Claire Cunningham in this video interview from All Access: Customer Data and Analytics APAC

Quick links

• Using sentiment analysis to drive customer satisfaction at Toyota Finance New Zealand
• How Energy Queensland is using generative AI to transform customer billing
• How to prove your CX program delivers results