Conflicting messages about consumers and retail continue unabated. Consumer confidence is starting to reflect the fact that inflation is easing up, which is good news. But there is still a lot of “fragility” in the numbers, where even as confidence increases, spending seems poised to pull back, as evidenced by Mother’s Day numbers. That perception is not helped by weaker GDP numbers either.
AI takes a bit of a beating this week, with questions around how good it really is – not as good as the hype – while also taking hits around sustainability, especially data center water consumption.
Redefining customer service as well as experiential retailing, more hits for SCO, and the new Foot Locker store concept round out this week’s high points in retail. Let’s dive in!
Retail Economic Indicators
This was a week for reports on consumer confidence, with one from Savanta that surveyed consumers across Europe and one specifically for the UK. Both showed improved consumer confidence in the economy, with notable improvements for the Nordics and the UK. The Savanta survey is still in negative territory, but increased confidence in perceived levels of disposable income and greater gains among younger people got levels almost into single-digit negative territory (yes, that’s the bar we’ve been set at this point).
In the UK report, personal finance confidence rose to positive territory (+2) the first positive and the highest score since December 2021. Overall confidence is still in negative territory, but up from March and up significantly year over year. Savanta commentary notes that consumer confidence remains “fragile” and GfK, commenting on the UK results, questioned whether these are actual improvements or a temporary pause before gains are lost once again. Not exactly the kind of commentary that would give any retailers confidence.
As a specific example of how challenging it can be to read consumer behavior, NRF released spending predictions for US Mother’s Day (May 12). Depending on how you read it, the estimated $33.5 billion spend is either the “second highest spending level” (beat by last year), or it is a sign of some pullback by consumers. 2023’s spend was estimated at $35.7 billion. With inflation still at play, that means 2024’s expected $33.5 billion is a sign of consumers spending less. The participation rate is as high as ever, but the spend per person is estimated to be down in 2024, to $254 vs. the estimated $274.02 per person in 2023.
And you know you’re not out of the woods for consumer confidence when shoplifting hits 20-year record highs – which is exactly what happened in England and Wales in 2023, according to the Office of National Statistics (ONS). This new watermark was also an increase of more than a third over 2022. Retailers and governments have taken steps already in 2024 to combat the increase, so it may be that 2023 or early 2024 is the peak and it will start to come back down. But the fact that it’s still very high reveals its own insight into consumer stresses.
Finally, the advanced estimate of US GDP came in at a tepid +1.6% for Q1 2024, down from +3.4% in Q4 2023, and below analyst expectations. Is this the flutter and dip of the plane’s wings before it gently settles on the runway in the long hoped-for soft landing? Breaking down the components, increases in consumer spending and housing investment were offset by a decrease in inventory investment. That’s maybe not as much of a negative as it looks on the surface.
Retail Tech & Research Data
Power Retail surveyed 1,000 Australian shoppers about click & collect (the more elegant name for buy online, pickup in store). Their survey has seen a consistent decline in the use of click & collect since 2022. In 2024, 63% of surveyed consumers said they had used the service option in the last 12 months, down from 72% in 2022. Of the 63%, 52% say they use it because it’s a cheaper option than delivery. 45% also say it’s faster, though that is down from 52% in 2022. Power Retail guesses that having wider availability of faster delivery makes click & collect less popular. I think there’s also something to be said for coming back off of pandemic highs. It will be interesting to see if the decline continues or levels out. It would also be interesting to compare to pre-pandemic levels.
AI & Retail
JLL made news with a new virtual spokesperson for 13 shopping centers across the US. It sounded like the intent was for use both in the mall as well as online, but the benefits cited were more focused on online, with site visits and web sessions up 500-1000% year over year. I’m not sure that this isn’t really more than a visual face and voice on what is otherwise still a chatbot, but answering questions about store directory or open hours seems a relatively harmless use.
In what is not that harmless of use, cybersecurity provider Thales reported that nearly half of all global internet traffic (49.6%) was produced by bots in 2023. This is a 2% increase over 2022, and the highest level since they started tracking it in 2013. The report includes a breakdown of the different kinds of “bad bots” and how that traffic differs by country.
If you think that sounds bad, consider that bots aren’t even that sophisticated yet. The Neuron’s Noah Edelman and Pete Huang came out strong against over-hype in AI, using quotes from Sam Altman and AI-driven revenue results from some of the biggest players (which is currently pretty anemic), to argue that the hype has left reality far behind. GenAI has a long, long way to go to match the hype. And I find myself more often in conversation with other software providers who are scratching their heads over what comes after the bots.
I have had it suggested to me from two different sources that GenAI would “really turn invoice matching on its head”. This is more of a head-scratcher than a head-turner for me. First of all, I frequently use invoice matching as an example of why you don’t do technology for technology’s sake. Do we need microservices to do invoice matching? Does it really bring something new to its efficiency? No. So, we definitely don’t need GenAI for invoice matching. I mean, using it write product descriptions – arguably a value-adding use-case – is already being compared to using a Ferrari to deliver pizza (as in, a bit overkill). I can’t imagine what the analogy would be for invoice matching. Sure, the use case could benefit from image to text. Sure, GenAI could probably make it single-digit percentage better at matching without human intervention (and I’m being generous there), but how many people’s lives are going to be impacted by this in any given company? Three? And you’re going to use multi-billion dollar foundation models to make those three people’s jobs 5% better? Color me a skeptic.
In the same newsletter, the duo also turns attention to “AI Agents” as the way to truly unlock revenue-generating opportunities. As in, AI Expense Agent, check against my AMEX bill for any expenses generated while traveling and update my expense report with the correct amount and category (including foreign exchange effects). While that’s not life-changing (I might argue otherwise), it is certainly something that delivers tangible value to a very large number of people, as in people who would be willing to pay for that. So I agree, that kind of use-case is interesting. The hard part is the complexity of the business model – you have these foundation models that say they need many many billions of dollars (at least) to get better. They have to figure out how monetize those models, but that means a vast number of agents being invented by a vast number of companies who take small pieces and pass back to the foundation model creator an even smaller piece. Is there enough VC money (and patience) to get to where the trickle becomes a steam becomes a mighty river? We’ll see!
The water analogy is no accident, because before we get there, there’s probably going to have to be a reckoning over the environmental impact of AI in general, but GenAI in particular. Your mind, like mine, might immediately go to power and whether that power is green or not. Power is a concern, but the real issue is water. Yale Environment 360 publisheda thoughtful piece on both questions. One could argue that the power of AI could be applied to reduce our carbon footprint and thus not just negate but overcome the footprint of its own power consumption. That remains to be seen. But one thing you can’t question is that AI is thirsty – and it drinks the same water we do.
According to the article, Google’s data centers used 20% more water in 2022 than they did in 2021 – before GenAI really really hit – and Microsoft’s water use rose by 34% in the same period. That’s potable, clean drinking water, at the same quality (if not better) that comes out of your tap. In The Dalles, OR, a city with 3 Google data centers and two more on the way, activists had to force the city to reveal that just the 3 existing data centers already consume more than 25% of the city’s drinking water. Even comparing to reducing the amount of carbon in our atmosphere and reversing the effects of global climate change, the water problem is not going to be easy to solve.
Retail Winners and Losers
I’ve been thinking a lot about customer service lately, because in the industry we throw that phrase around a lot, but it can mean many different things. It could be specifically, how well a company helps a customer that has a problem. It could be general, like how well a company meets customers’ needs and wants. It could mean just looking at how well the company’s people helps its customers – its call center reps and its store associates. When you’re trying to figure out how to help retailers enable their “customer service strategy”, having an unclear definition of the focus is a problem.
Qualtrics is here to muddy the waters by publishing an aggregated ranking of customer service for 50 companies to find the companies with the highest customer service satisfaction. Chick-Fil-A topped the list, followed by Publix, then insurance provider USAA, then Trader Joe’s and Wegmans. The list highlights the problem. It’s much more likely that USAA is getting high marks for helping their customers in need, while Chick-Fil-A is famous for moving its line fast, and Trader Joe’s for its fun (and also fast) environment. I still don’t know what customer service really means, but consumers definitely good customer service when they see it.
In my favorite story of the week, Sur La Table has partnered with Academic Travel Abroad to offer educational trips focusing on “immersive cuisine experiences”. These are mostly trips to France, in honor of the company’s founder and her own culinary heritage. The service is called Culinary Journeys, and I expect the first two will sell out in no time. This also isn’t the company’s first foray into experiential retail, with reservations available for training from in-store experts and the company’s “Ask a Chef” program. When I talk about the concept of retail leverage, this is exactly the span of experiences that I’m thinking about. It doesn’t have to be as high-end as a week of food tours and cooking classes in the north of France, but any experience that expands the brand’s relationship with customers should be on the table (see what I did there?).
In less positive news, Walmart is removing self-checkout (SCO) from two stores, one in St. Louis, MO and one in Cleveland, OH. However, in Progressive Grocer’s coverage, they highlighted all of Walmart’s recent SCO moves, and a few of them stand out. Some shoppers are reporting that some locations have turned their SCO lanes into access for Walmart+ subscribers and/or Spark delivery drivers only. And last fall, the retailer rolled out item quantity limits to 200 locations, capping them at 10 items or less. Treating SCO as a ‘premium’ experience is interesting. For sure it reduces the temptation to steal since every transaction is identified in order to get all the “+” benefits.
Store Innovations
Finally, Foot Locker got lots of coverage for its store of the future concept, which opened this week in New Jersey. It’s one of 5 planned stores for this year, but will shape the redesign of an additional 900 stores over the next two years. The intent was to be more immersive and community driven, with a “drop zone” for featuring new sneaker drops, a communal try-on area, and a sneaker hub for customizing shoes with options like specialized lacing.
It’s an interesting balance of the need to serve the customer vs. the need to cater to brands who want distinctive treatment. Is it better to have a Nike zone, or better to have a running zone with Nike featured? This is the dilemma for any retailer who sells other companies’ brands. We’ll see how much of the concept actually makes it to the 900 additional stores – props to the company for a very aggressive timeline for the redesign, at least. It’s easy to be tempted by a long, slow and drawn-out remodel. And doing it fast is expensive. Two years is fast. Hopefully it pays off.
The Bottom Line
What did we learn this week? There’s no use waiting for consumers to figure it out. And there’s no use dreaming of a more stable, predictable future. Much as consumers have decided to “embrace the mess” of the post-pandemic world, retailers just have to do the same.
That means trying new things even in an uncertain future, like Sur La Table and Foot Locker. It also means embracing the things that AI is good at – like summarization and explainability, or natural language interactions about product selection or recommendations (aka, chatbots), while not losing our heads about the extent of the use-cases that are possible – or profitable. Ultimately, it means moving forward without losing sight of the risks – and preparing for as many of those as you can.
We talked about disruption in the past, meaning technology disruption and consumer expectation disruption. Now we mean economic uncertainty. The hardiest retailers will learn how to navigate this uncertainty to thrive. Everyone else will wait for a future that might never materialize.
It’s a mess out there. Embrace it.