Retail Pulse Report: Consumers Decide to Rearrange Their Wallets

In the wake of rapidly falling confidence, consumers don’t pull back so much as re-evaluate their expectations

Adobe Stock’s AI response to “consumers rearranging their wallets”.

As we wait for the economic impact of trade wars to show up in numbers, one number that is getting everyone’s attention is falling consumer confidence. The new thing I learned this week is that consumer confidence has not necessarily predicted a pullback in consumer spending – at least not since the pandemic. The vibecession continues. However my new favorite phrase – as I’m sure you’ll be able to tell – is that consumers are not pulling back so much as “rearranging their wallets”.

What does that mean exactly? Let’s dive in!

Retail Economic Indicators

Since everyone seems to care a lot more about what’s going on in Canada these days, I thought I would open with an update on the Canadian Consumer Price Index provided by Statistique Canada. February CPI rose 2.6% year over year and 0.7% month over month (seasonally adjusted).

This does not yet reflect any impact from tariffs or trade wars – the increase was mainly attributed to the end of a tax break on some products that fell halfway through the month, where the impact was felt most keenly on restaurant food prices. Everyone is twitchy looking for trade policy impacts, but we’re just not there yet. It'll be April before we really have any chance of seeing it.

Meanwhile in the US, the US Census Bureau released its take on US February retail sales. The headline that ran in the press what that sales were “only” up 0.2% month over month, when the consensus expectation was 0.6% - those nameless surveyed economists.

The bigger deal is that sales were up 3.1% year over year, and retail sales excluding motor vehicles and parts as well as gas stations were up 3.5% year over year. Total sales for the 3 months December through February were up 3.8% year over year. Those are all solid numbers, at least relative to past uncertainty. The vibecession continues.

Speaking of vibes, the BRC-Opinium survey on UK consumer confidence was up slightly. Personal spending on retail moved from -5 in February to 0 in March, and personal spending overall went from +4 in February to +11 in March. The increases were attributed mostly to Gen Z’s willingness to spend. The BRC read the numbers as showing that consumer confidence appears to be stabilizing after a low in February.

Meanwhile, a competing index from Nielsen-IQ and GfK ticked up one point to -19 in March. Views on personal finance for the past year were slightly down from -7 to -9 in March, but looking ahead, expectations for the economy were up two points from February to -29. A choice quote from Neil Bellmay, the Consumer Insights Director on the report had a bit less positive spin than the BRC: “If consumer confidence were a patient languishing in a hospital bed, a doctor would say there is little evidence of a recovery as yet.”

UK’s Mother’s Day tends to fall in March while the US Mother’s Day falls in May, so indicators for the March holiday can sometimes give insights (or early warnings) about the bigger spending day in May. GlobalData estimates that UK Mother’s Day will grow 5% year over year to reach GBP2.4 billion. More consumers are planning to participate, and they plan to spend nearly 14% more year over year, on average. Again, more vibecession than actual spending pullback – so far.

And while this last one is not directly an economic indicator, I thought it was worth covering in this section. It’s a great question, and very relevant for the moment: how much could declining consumer confidence influence spending? The answer turned out to be very interesting: the pandemic broke the tight tracking between confidence and spending, and it has not come back. It’s an important reminder, as we’re trying to parse low consumer confidence against steady (if not extravagant) spending. Some researchers interpret some of these breaks between confidence and spending as more reflecting a “rearrangement of consumer wallets” than an actual pullback on spending – in other words, they are changing their priorities on what they buy, but not net pulling back on the total spend.

Retail Tech & Research Data

ThredUp is on its 13^th annual resale report (3,034 US consumers over the age of 18 and 50 top US fashion brands and retailers, in surveys conducted by GlobalData), which has tracked the rise of resale and its underlying trends. With consumers theoretically “rearranging their wallets”, this year’s survey comes at an interesting time, and the results don’t disappoint.

59% of consumers say if tariffs make apparel more expensive, they will turn to more affordable options like secondhand. 48% of consumers say personalization, and improved search and discovery have made secondhand shopping as easy as shopping new. 39% of younger generation shoppers (Gen Z and Millennial) have made a secondhand purchase on a social commerce platform in the last 12 months.

Overall, resale outpaces retail sales growth, and online resale outpaces both – the apparel secondhand market grew 14% overall and 23% online in 2024. Gen Z and Millennials say they’ll spend nearly half (46%) their apparel budget on secondhand in 2025.

On the retailer side, 94% of retail execs say their consumers are already participating in the resale market (up 4 points from 2023 and an all-time high for the response). 32% of consumers who bought secondhand in 2024 bought directly from a brand – which has a lot of interesting retail implications, especially for stores. But one potentially troubling sign: 46% of consumers say if they can find an item secondhand, they won’t buy it new.

AI & Retail

When it comes to AI, we seem to be living in two different worlds at the same time. In one world, AI is practical and eminently useful. It has targeted use cases that provide measurable benefits. That seems to be the story of Wally, the GenAI assistant that Walmart has rolled out to its merchant teams.

In this world, Wally helps merchants identify why certain products are under- or over-performing, “generate insights from complex datasets”, answer operational questions and raise tickets for unresolved issues, and automate complex formulas and predictions. (All of this sounds like par for the course from any advanced analytics solution, perhaps trained and tuned specifically on Walmart’s workflows and datasets).

In the other world, AI is vacuuming up immense amounts of capital, and while use cases can deliver value, that’s not on a cost basis that reflects what the AI powering those use cases actually cost. This is getting cheaper, but the performance of cheaper is also getting muddier. In this world, a big vision for the future of all of this investment is important, because it keeps everyone focused on a willingness to pour more money into AI investments and wait for the “real value” to be delivered.

Andrea Rosales, the lead data scientist at Colibri Digital, covered comments that Microsoft CEO Satya Nadella made about the future of SaaS in an AI world, and it definitely strikes me as falling in the second kind of world. She’s got a diagram there that’s worth checking out. In some ways, it’s not much different than lots of other diagrams that people have been putting out about what an “agentic organization” will look like. There is a demand for something (a “task”), and an orchestrator that is designed to break the task down into discrete steps that it then parcels out to specifically trained AI (in her diagram, there’s a coder AI, a terminal AI to run the code, a web surfer AI, and a file surfer AI).

But it’s also probably a more “complete” version than other diagrams. You could interpret it in the smallest sense of a task that human worker assigns, like “research the market potential for a product that does X”, or more aggressively, “create a software application that does X”. You could also interpret it in the largest sense: “create a product that will grow my company’s sales”.

There is a world of difference between these three types of tasks. The first task is one that can – almost – be done today with today’s AI tools. You might need to add a step or two to get there, but you can get pretty close. The second one, you can do if all you want is a cute little mobile app. There are a ton of mixed reviews out there about coding and its ability to deliver much more complex code bases (not to mention whether it will eventually kill off programming skills in humans). But that’s all you interacting directly with the coder. Still some manual steps in there. And the last one, that’s a huge leap.

I’m trying to live in both worlds at the same time. I see the value of these narrow use cases. I definitely want to keep an open mind about the future of business and the future of software and that kind of thing. And I always keep in mind that people tend to over-estimate the impact of big changes in the short term, and under-estimate their impact in the long term.

All ingredients that are going into the stew, I guess.

Retail Winners and Losers

Two more stories to finish out the week, both picking up on past themes I covered that I think will continue to shape retail in 2025.

The first follows a plus-sized retail consultant as she went shopping at American Dream. It was an exercise in frustration, where she was searching for an outfit for a special occasion, and found one store – Torrid – in the whole mall that actually had her size available to try on, despite several retailers there promoting that they carry extended size ranges. They do – online.

So back to how retailers are turning stores into frustrating experiences. This reminds me of the classic assortment planning problem of dealing with how to make sure that the smallest stores don’t just get black and brown and no other colors. If you have to spread a wide range across a lot of stores, many of which don’t have the space to hold a lot of inventory, if you don’t actively prevent it, the algorithms recommend the highest volume colors, which are often the most boring.

If you’re not careful, it’s pretty easy to sub-optimize yourself into a corner. Nobody may buy plus sizes in your stores simply because you don’t carry them, not because they don’t want them.

The second article is more a thought-generating piece on an AI-driven future. The title is what caught me: “What jobs are you doing for your customers?” It was intended in the spirit of making sure that you know what your customers want from you. But what it made me think of was “what jobs are you making customers do for themselves?” – because that seems to be a lot more of what’s happening in retail these days. Consumers are pushing back against the jobs that retailers are making them do – like checking out, or finding the right inventory or the right fit. As consumers “rearrange their wallets”, retailers need to keep in mind that the value equation might very well expand into how much work you make your customers do.

What Did We Learn This Week?

Consumers’ lack of confidence has not translated to a pullback in spending for at least the last three years. The idea that they are still trading down, however, seems alive and well – and resale is poised to take advantage of that particular wallet rearrangement. Full price retailers could take better advantage of that than they do today – but that’s a post for another time.

AI – and it feels reductive to specify GenAI, though that’s what everyone is fixated on – is both a focused use case and an overly expansive vision of the future. Trying to keep both extremes in mind is difficult but necessary. The key is the adage, people tend to over-estimate the short term impact of change and under-estimate the long term impact. That’s how both can be true at the same time.

And finally, while retailers can survive the future by focusing on what it is they actually do for customers, they’d be better off focusing on the things they’re not doing for customers – but should be.

This week I’m at Shoptalk, Aptos has a booth so come by and see me there (or in sessions whenever I can be). And program note, next week will undoubtedly cover what I learn from that event. Until then!

- Nikki