Retail Pulse Report: Discovery Commerce, Our AI Overlords, and More Divergence
I don’t know how you have increased concern over tariffs but decreased concern about economic impacts, but who said consumers are rational?
Source: Adobe Photostock. Prompt “social media addiction”
Here’s what to look for this week:
Continued divergence between consumer sentiment and consumer spending (though not always in the way you expect)
The Chinese economy absorbing tariff-abandoned goods, so far
Consumers both more worried about tariffs and less worried about the economic impact
More questions about the long-term impact of using LLM’s – apparently, ChatGPT is not going to be our AI overlord, Sam Altman is
Whether “discovery commerce” should replace what happens in stores – and whether consumers really trust social media-driven commerce
One way to make it up to consumers after you got hacked
You know, light reading for the summer! Let’s get into it.
Retail Economic Indicators
In the UK, three key measures dropped – consumer sentiment, retail sales, and inflation. The UK appears to continue the trend of divergence between consumer confidence and sales results, but not in the same way as the US. And in the meantime, the UK’s inflation readings show that tariffs or not, getting out from under inflation is still as tough as ever.
Let’s start with consumer sentiment, a survey by the British Retail Consortium, which improved from -36 in May to -28 in June. Consumer personal finance outlook improved from -12 in May to -5 in June. Personal spending on retail jumped to +2 from 0. The survey showed the second month of improvement, and the BRC attributed some of that improvement to minimum wage increases that came through in April (I have to note that these were part of the same Budget package that they complained about as driving too many costs for retailers).
That sounds very rosy and hopeful. But then May saw the “sharpest monthly fall” in UK sales since December 2023. This is one of those annoying headlines. Yes, monthly retail sales fell 2.7% in May. The ONS reports it this way, and you have to go looking to find any year over year comparisons. When you find it, you see that at least over the three months to May, sales increased 1.7% against last year, which is a far cry from as bad as the headlines make it seem.
However, when you see the inflation results below, it will show that this number is still not very impressive, but it is more meaningful and has better context than the month-over-month one. And retail sales are still 2.7% below pre-pandemic levels of February 2020 – that one has more useful context than month over month.
So, inflation. UK inflation came in at 3.4% in May, which was slightly better than April’s 3.5% (which was then revised downward to 3.4%). Rising food prices made the biggest contribution, offset slightly by falling airfares. That’s almost twice the rate as the sales increase year over year in May, which is why even a sales increase doesn’t look that great.
Inflation is still pretty high, retail spending is not awesome, but consumer confidence is high? What gives? It is partly an issue of timing – consumer confidence is more a June measure than May, but even May confidence was up over April. It just has not really shown up in spending – yet.
Meanwhile, in the US, May sales were up 3.3% year over year, which is less than the last 2 months’ increase year over year. Total sales March to May are up 4.5% year over year. This falls in line with NRF’s earlier call that any pull-forward spending by consumers is starting to taper off.
At the same time, imports are resuming at ports as retailers recover from the initial tariff panic and uncertainty. With new deadlines looming and no certainty of an extension or resolution, retailers are scrambling to bring in everything for Back to School ahead of whatever might happen next. NRF’s port tracker forecasts that import volumes will fall off in the last four months of the year, no matter what happens with tariffs, just because retailers have already brought in so much so early.
Will retailers’ rush to bring in goods ahead of tariffs run headlong into a consumer spending dip – consumers who also rushed to bring in purchases ahead of expected tariff-driven price increases? If that happens, then retailers may be stuck with a lot of excess inventory that they will have to discount to move. That seems to be top of mind at the Fed, where Fed Governor Waller predicted that a rate cut could possibly come in July, mostly because he doesn’t expect tariffs to increase prices.
On the other side of the mirror, Chinese retail sales rise, even as tariff reactions take a toll on factories. Retail sales rose 6.4% year over year in May, helped along by promotions on products stranded in shipments that never left due to tariff-driven order cancellations. Manufacturing output also rose, to 5.8% year over year in May, but that reflects a downward trend – March’s year over year output was 7.7% YoY. Exports to the US fell 35% in May YoY. China offered a program to subsidize trade-ins of household appliances, cars, and other goods that had big tariff impacts, which also helped to boost sales.
Retail Tech & Research Data
More survey companies are gearing up with better tracking of consumer sentiment with regard to tariffs specifically. Numerator has launched a Tariff Sentiment Tracker. The survey runs every two weeks and surveys 5,000 US consumers. The first survey launched during the week of April 21. Findings so far:
77% of consumers surveyed are aware of the new or proposed tariffs.
88% are concerned about the impact on their personal finances or shopping habits.
30% support the tariffs vs. 45% oppose – and 29% are strongly opposed vs. 12% who strongly support. Overall support is down 2 points in the latest survey vs. two weeks ago.
27% of consumers believe that tariffs will have a positive impact on the US economy over the next year.
77% are concerned about the possibility of a recession in the coming year.
81% of consumers surveyed expect to adjust their finances or shopping habits in response to tariffs.
However, economic concerns have declined in the past month – 33% of consumers surveyed are concerned about the impact on the stock market, and 29% are concerned about slower economic growth. This is versus 41% in late April for the former and 33% in late April for the latter.
How are retailers responding? A new survey of 500 eCommerce professionals, conducted by Sygnifyd, found that retailers are raising prices, moving production, and changing product sourcing. 76% of respondents report that they have raised prices, and that they are passing along 51% of tariff costs on average, choosing to eat the rest. 75% have reduced the number or size of their discounts and promotions to lower costs. 71% have switched suppliers from higher tariff to lower tariff countries. And 71% report they have limited their inventory or number of SKUs that are subject to tariffs.
Changing focus a bit, WalletHub also has a new survey out. This one is more generally focuses on social media shopping. Nearly 1 in 5 Americans would describe their social media purchases as scams. 4 in 5 say they would spend less money if they deleted their social media accounts. 54% report regretting purchases they’ve made through social media.
And who can resist a tariff question? WalletHub found that 60% of Americans are concerned about tariffs making their social media purchases more expensive.
Hold on to the points in this study – we’ll come back to them in two more places.
AI & Retail
Here’s more evidence that LLM’s, and specifically ChatGPT, is not good for your mental acuity, this time from Cornell University and a whole slew of researchers. They studied essay writers divided into three groups – one aided by LLMs, one aided by search engines, and one with no aids (just your own brain). After 3 sessions writing essays, they swapped the LLM group to brain-only, and swapped the brain-only group to using LLMs. They also scanned everybody with EEG’s to assess cognitive load for essay writing.
They found that those who use LLM’s showed much less brain engagement. They struggled to quote their own work. And four months later they consistently underperformed at neural, linguistic, and behavioral levels. My words: they learned less, they retained less, and were less able to use what they did retain. Oof.
But if you think that’s bad, just wait! This paper will make you seriously paranoid. MIT published a paper arguing that the ultimate aim of OpenAI is to generate and then sell consumer intent, and that models today are already manipulating user intent in various ways. Apparently, it’s not the AI overlords we need to fear, it’s the people running the companies that build these AI’s. I mean, this is MIT saying this.
What does that mean, to “generate and then sell consumer intent”? Well, let’s go back to that WalletHub social media survey, where consumers said they have bought things that are scams, they have regretted purchases, they think they would save money if they deleted their social media apps, and yet, they keep spending.
The MIT researchers define the intention economy as a “digital marketplace for commodified signals of intent.” In their words:
"[A] digital marketplace for commodified signals of ‘intent,’ would differ from our present-day attention economy. Today, advertisers can purchase access to users’ attention in the present (e.g., via real-time-bidding [RTB] networks like Google AdSense) or in the future (e.g., buying next month’s ad space on, say, a billboard or subway line). LLMs diversify these market forms by allowing advertisers to bid for access both in real time (e.g., ‘Have you thought about seeing Spiderman tonight?’) and against possible futures (e.g., ‘You mentioned feeling overworked, shall I book you that movie ticket we’d talked about?’). If you are reading these examples online, imagine that each was dynamically generated to match your personal behavioral traces, psychological profile, and contextual indicators. In an intention economy, an LLM could, at low cost, leverage a user’s cadence, politics, vocabulary, age, gender, preferences for sycophancy, and so on, in concert with brokered bids, to maximize the likelihood of achieving a given aim (e.g., to sell a film ticket)."
But they take it one step further, and put out evidence that this is exactly what OpenAI is shooting for. They point to several emails, public comments, etc. from various OpenAI execs, but quote Sam Altman saying that the company believes that chatbots work explicitly to gather the user’s intent.
I’ve seen this in the behind the scenes of other AI platforms – this gathering of user intent – and heard retailers talk about it in terms of understanding what their employees are asking. That’s cool when your objective is to improve the user experience and identify new ways to better answer the questions your employees ask. That’s not so cool if your objective is to sell the ability to influence the people asking the questions into buying things – sell it to the highest bidder, no less.
Humans who are uniquely susceptible to digital influences already, and who lose mental acuity the more they rely on LLM’s – how do we all stand a chance against this level of influence? I’m actually a bit surprised that it’s taken this long (6 months) for this research to make it to someone like me.
Retail Winners and Losers
I’m always on the lookout for frameworks that can help us navigate the current world of digital and physical retailing, so when I saw this headline (“A new way to shop is integrating the best of physical and online worlds”), I immediately jumped in. However, it only took me a couple of paragraphs to realize this was an opinion piece with an agenda (most opinion pieces are, which is why I don’t cover them very often).
Whose opinion? Accenture. Opinion about what? TikTok Shop. “Discovery Commerce” – that new way to shop – is about embracing TikTok Shop and influencer-driven marketing/direct selling. The initial setup was good: “Shopping in physical stores creates serendipity, with consumers relying on their environments to be inspired. It’s also entertaining, social, and fun.” The conclusion is not so good: “Discovery commerce takes all the things that people love about going to shops and replicates it online.”
Does it? And, even more important, do we need that? Why can’t online (and social specifically) have its strengths and weaknesses, and stores have their own strengths and weaknesses? The point is not to replace one with the other, but to handle customer journeys that flow seamlessly across all and any touchpoints.
Any time someone says something like this: “Brands must go all-in for maximum effectiveness.” – that’s a red flag. Yes, there is a point to “discovery commerce” – using influencers to reach consumers, and then making it easy and frictionless for them to buy it if they decide they want to try it. But you’re not going to sustain your success with those new shoppers if you rely solely on the influencer to generate sales. You need to build your own relationships, or you’ll always be one spat with an influencer (or one influencer meltdown) away from losing contact with your customers. This was something that Amazon sellers learned about how to make the most of that commerce channel – you can find customers there, but you’re not going to have a sustainable business if that’s all you depend on. Discovery commerce does have a role, but it’s not going to be the only way people buy products.
And why is Accenture promoting all of this? Because they have partnered with TikTok Shop to create a playbook for leveraging the commerce channel. And it’s designed for larger companies who are still mostly sitting on the sidelines. Agenda (which is fine, everyone has an agenda. Just don’t bury it towards the bottom and clothe it in thought leadership at the top).
This partnership is interesting, because even as recently as April at the Retail Technology Show in London, big retailers were dissing TikTok Shop. They said it was fine if you wanted to try out commerce in countries where you don’t have an established eCommerce presence, just to see if you get any traction or to help prioritize which countries to go after first. But they also said if you’re going to do it, you should only focus on low-price-point items because those will be the only ones that really sell.
And, as the WalletHub survey found, consumers are aware that there are a lot of scams. They’re aware that they spend money on things they don’t need on social media. They know if they spent less time there, they would spend less money. Does that sound like a great foundation for a strong relationship with a new customer? It doesn’t sound that way to me.
One more story from me: Co-op’s “thank you” initiative after recovering from a cyber attack. As you may have heard, a pretty large number of UK retailers got hit with cyber attacks. The most prominent and most damaging seemed to be Marks & Spencer, but Co-op got hit too. Everyone talks about the costs of an attack, and these attacks had a lot of costs because they were successful in shutting down eCommerce for a period, among other things. But you can’t ignore the costs of making it up to your customers. Most don’t really publicize it, so I have to give a shout out to Co-op who took it head on, with a thank you sale for a week, that offers £10 off a £40 purchase to its 6.5 million members. Yes, a cost, but an important outreach, and I like the tone – not “sorry this happened” which just sounds sad and lame, but “thank you for standing by us as we fought this off and recovered”. That sounds like a pretty good playbook to me.
What Did We Learn This Week?
I learned that this divergence between consumer confidence and consumer spending is for real. In some ways it doesn’t matter – the numbers are the numbers. But it does make it even more difficult to understand what could happen next.
Case in point: the divergence that continues between how consumers increasingly view the tariff impact and how retailers do. This could lead to an over-supply, an under supply, or who knows what. A mess, either way.
I learned that the idea of “The Algorithm” influencing and even distorting consumer behavior is more real than I was giving it credit for. I’ve read take-downs of TikTok, offering evidence that the Chinese government is using it to manipulate public perception – US public perception. And they can be pretty compelling. But regardless if you buy into that or not, I think there are enough studies and enough commentary from leaders of these companies out there to warrant much greater caution than “go all-in.” In the typical Gen Z fatalist world view, they urgently believe they should not make themselves susceptible to “The Algorithm” and also happily scroll through, engage, and buy on TikTok. Retailers should be slightly more cautious than that.
Program Note: No newsletter next week, June 30. I will be joining a RetailROI group going to Honduras to support Casita Copan and COMFE. RetailROI organizes donations and grants for orphan-focused charities around the world. Their biggest fundraiser is Super Saturday, the day before NRF kicks off in January. If you’ve never heard of it or heard but wasn’t sure how to get engaged, please check out any of the links above.
I’m not sure if I’ll be able to pull a newsletter together for July 7, but I will try!
Until then!
- Nikki