Retail Pulse Report: GenAI is the Free Shipping of Productivity
Plus, Walmart wins and Target loses – again.
Adobe Stock: “flaming GPUs” - I had to work harder this week as there were no results for “a GPU incinerator burning cash”
This week saw the first real, measured US consumer reaction to the execution of promises made by the new political regime, and it wasn’t promising for 2025.
GenAI doubters are starting to get louder and more pointed in saying “Show me the money!” – the only money seen so far goes into a GPU-powered incinerator.
And Walmart and Target seem to demonstrate a distinct shift in how to compete in 2025: investing in caring about causes doesn’t pay if you don’t have the will to stick to your investments. But investing in technology to build compelling experiences seems to attract – and keep – shoppers of all kinds. At least so far.
Let’s dive in!
Retail Economic Indicators
The Conference Board released its Consumer Confidence Index results for February and the news is not good. The index took its largest monthly drop since August 2021, falling 7 points to 98.3. The Expectations Index fell 9.3 points to 72.9, putting it into an indicator for a future recession, as TCB measures it. Pessimism about future employment prospects worsened to a ten-month high, and future expectations for inflation (12 months from now) went from 5.2% to 6.0%, reflecting current inflation stickiness, high food staple prices like eggs, and fears of the impact of tariffs. The Present Situation index fell 3.4 points to 136.5 in February.
This pretty much sums up the vibe in the US at the moment. The current situation is pretty much unchanged. In a barrage of disturbing political news, it feels almost surreal to go about daily life. Yeah, sure, the egg section of the grocery store looks like a pandemic-era toilet paper aisle. But unless you are a government worker or you know one, life goes on at the moment. It’s where life goes on to that is becoming a larger and larger concern.
The economic hits keep on coming, too. In a week of relatively scarce economic news (the CCI was the biggest one by far), it turns out that any good news has bad news underneath it. Moody Analytics revealed in their latest consumer study that the top 10% of earners – those making $250,000 per year or more – now account for nearly 50% of all consumer spending in the US, which is the highest it has ever been, going back to 1989, when it accounted for about 36% of all spending. Keeping inflation in mind, the bottom 80% of earners increased spending by 25% over 4 years earlier, versus price increases of 21%. The top 10% spent 58% more.
Anything that shocks wealthier consumers - typically, stock market declines or big drops in house prices - will have an outsize impact on the economy if they pull back spending in response. This cannot possibly be the foundation of a healthy economy.
Retail Tech & Research Data
While not an economic indicator per se, Circana released a report on the state of private label globally and found that the category continues to grow. However, while it is true that private label is benefitting from consumers looking for value and willing to trade down to get it, some of the growth of private label has come from innovation and differentiation.
As has been the case for a long time, the EU dominates regions in private label adoption, capturing 39% of spend and 47% of units. The US market is the fastest growing, but holds 22% of spend and 24% of volume. Private label growth always required “threading the needle” – striking a balance between catering to consumer tastes and preferences vs supporting brand manufacturers. Consumers seeking value put a lot more pressure on that relationship, especially as retailers and brands battle over who takes on the brunt of inflation in consumer pricing.
AI & Retail
Someone pointed out this rant by Ed Zitron, a media relations expert, about how OpenAI and Anthropic are a giant scam that have “used a compliant media and braindead investors to frame unprofitable, unsustainable, environmentally-damaging and mediocre cloud software as some sort of powerful, futuristic automation."
Applying critical thinking skills, I will say that the article is not the most coherent (thus, “rant”). To be fair, I have written my own rants that seemed perfectly reasonable and reasoned when I wrote them and seemed utterly unhinged when I went back and read them later, so take it with that rather large grain of salt. But underneath some of the messiness, he raises some valid points that are worth surfacing. Focusing specifically on OpenAI, he points out that:
Unit costs don't scale - every query costs. So every user added costs more. For Meta or Google, adding a new user has scaling advantages built in that OpenAI in particular cannot boast. It's the grocery adage, I know I'm losing money on every unit but I'll make it up on volume.
The infrastructure that OpenAI runs is very high-end and it's running full tilt all the time, so the replacement rate (and cost) is going to be much higher than other tech companies. And, because of OpenAI's relationship with Microsoft (which he points out seems to have soured), it's not OpenAI who bears all these costs, it's Microsoft. And Microsoft has given OpenAI a deal so they're using all this capacity and they're not paying market rates for it, so it's sort of costing Microsoft twice. I had not seen this pointed out before, and I think it’s an interesting point.
OpenAI has to pay a premium for talent because everyone is sucking up AI talent, but at $1.5B in annual salary (before stock-based compensation) - salary alone is 37.5% of revenue, forget about GPU's.
There are only 15.5 million paying subscribers, and whatever service they're using, they are not paying enough to cover their costs (OpenAI doesn't provide a breakdown of its subscriber revenue).
They only report weekly users - 400 million - of the free service (I'm in there somewhere). Monthly subscribers is not the same basis of measure as weekly average users, which makes it more difficult to measure conversion rate of free users to subscribers. If you do believe that 400 million WAU is also the monthly user rate (why not just say that?) then OpenAI's conversion rate is 3.9%. If you think maybe it's double, 800 million MAU, then their conversion rate is half that. Zitron uses 500-600M MAU across web and mobile, which gets around a 2.5% conversion rate. And he points out, but even those that convert are losing money with every visit.
And then he goes back to my earlier question – is it really delivering life changing value? Sure, I use Copilot all the time at work. Work pays for that. Do they get $20/month extra productivity out of me? Definitely. Do they feel that in the top or bottom line? Very doubtful. If Copilot suddenly jacked its rate up to $100 or $200/month, would my company continue to pay for it? I'm very doubtful.
Basically, to use a retail analogy, he's saying that GenAI companies are the free shipping of productivity. Sure it's highly convenient and people will use it. But if you suddenly start charging for it, they're not guaranteed to suddenly agree that they'll pay for it. They'll go to someone else who offers it for free, and if no one else does, they'll start thinking about other ways to get it for free, like picking it up in store.
And to underline that, at least for retail, it is no guarantee that GenAI will be so useful that retailers will pay enough to support a real business model, Omnisend (an eCom marketing company) released a consumer survey on shopper attitudes towards shopping agents, now that we’re past the holiday season. And two-thirds said they would refuse letting AI make purchases on their behalf, even if they get a better deal.
It was interesting to find that it wasn’t that shoppers thought the AI would rip them off or spend all their money. Their biggest concern was privacy: 58% were concerned about handing over their personal data and 37% were concerned about their privacy in general. Even more interesting, 42% said they felt the AI didn’t have their best interests at heart, that they were really there to upsell. And 40% said there wasn’t enough human support in case they got stuck. In fact, 39% said they have abandoned purchases because of frustrating AI interactions.
With that kind of lukewarm reception, combined with OpenAI’s complete lack of even covering their fixed costs, let alone marginal, I can’t see retailers paying what GenAI really costs, when it inevitably comes about that they have to do so.
Retail Winners and Losers
I have a feeling that this WSJ article on how the internet makes in-store shopping “miserable” made the rounds across the industry. It did inside my company too. It comes as we round the corner on 5 years since Covid shut down the world (five years!!). I’ve seen some of the stats that the article quotes before – “nearly three-quarters of consumers prefer shopping in physical stores, but only 9% are satisfied with the store experience” is a common one.
But one I had not seen before came from AlixPartners: a study of 30 retailers found that on average only 9% of their online women’s clothing assortment was available in physical stores. I don’t know how they’re measuring that – if it’s at the style level, or color-size, which would make a huge difference to the number. If the style is available, maybe in the right size but not the right color, that at least is close enough to tell if it fits. Across format, the study found department stores had about 7%, mass merchandise retailers had 2%, and specialty retailers did much better, with over 30% on average.
I’ve encountered this, and worried about it. Last summer, I discovered the health benefits of running regularly (I had never run, voluntarily, before this). Wanting to be prepared for winter, I went to a couple different sporting goods retailers in the fall to see what they had for Gore-Tex running shoes. I mean, I live in Colorado. You’d think you’d want to carry cold weather running shoes in a state that has cold weather. But no, every place told me “we no longer carry those in stores, you’ll have to buy them online.” I ended up ordering 10 different pairs of shoes, kept 2 part and sent the rest back. I’m 100% certain the retailer I ended up buying from lost money on that deal.
The author pointed to smaller stores and thus a smaller assortment as part of the reason why stores are missing out, coinciding with my own anecdote. And, she also noted labor – or the lack thereof for in-store help.
Especially now that online has become noticably less profitable as a channel, retailers really need to consider whether they are closing the right sales in the right places. One thing that might help would be looking at products with high return rates and seeing if they would sell better (and more profitably) if they were stocked more heavily in stores.
On another note for retail winners and losers… For at least the last year, I’ve been saying that Walmart is becoming a winner and Target a loser. The trend continues into 2025. I have to point out that I did predict in 2024 that another brand would find themselves between a rock and a hard place over politicized topics, a la Budweiser. I missed the timing by only a few months.
Target seems to be aiming to become the new Budweiser. They had long been a passionate supporter of diversity and equity, and with the new regime in political power in the US have dropped that support like a hot potato – except in their social media posts. And they are getting burned for it, fairly relentlessly over the month of February. My favorite response, to a post promoting a pretty benign soda brand: “I’m surprised ‘orange’ isn’t the only flavor.”
At the same time, Target has seen a decline in foot traffic. For the week of 2/17, Target traffic was down 7.9%, according to Placer.ai, which was about double the 3.9% it fell in the previous week. Walmart’s traffic fell 5.2% in the same week, so Target wasn’t alone, but Costco’s traffic has been up – it rose 4.8% the week of 2/17. No one is willing to go on the record that politics has played a role, but Target had long ago picked a side, by centering the brand on diversity values. To cave to political pressure is to risk both losing loyal customers without doing anything to win detractors at the same time.
While Walmart’s traffic may be down in February, on the earnings front, they continue to kill it. I stopped trying to cover sales results from retailers because it got a bit overwhelming. At some point in my future, I think there's a valuable angle in looking at some kind of "IT value index" that could rate retailers on how well technology is contributing to their sales results (or detracting). But not today.
However, this story was worth covering, because WMT's remarkable results have come in part from wealthier shoppers trading down. That's happened before, but in the past they haven't been able to keep them. This time around it looks like they may stand a chance - 36% of high income shoppers (making more than $100k per year) said this year that they have a "very favorable" impression of WMT, up from 27% in 2019, according to a Morning Consult survey.
Some of their results are coming from the leverage they get in monetizing their assets and services, very much like Amazon does. But what I'm most curious about is whether some of those tech investments - their app, for example - are helping them keep those higher income shoppers that they might have lost again in the past. It's also possible that the economy has not improved enough for those shoppers to confidently shift their spending away and may yet do so, so we’ll see.
But Walmart is not stopping at all on the innovation front – they also announced this week a 3-episode gamified miniseries on Spatial. And, in what feels a bit like a shot at Target and their historical support for diverse and innovative suppliers, the miniseries focuses on discovery of unique Walmart suppliers, including taking gamers to a commerce hub in the game to purchase in-real-life goods.
Store Innovations
Finally, I will cover one new store this week that might possibly be innovative. It’s the new Cross Store being opened by Bandai Namco, featuring several of its properties, including Pac-Man. There isn’t necessarily anything innovative about the experiences in the store – there are arcade games you can play, for example, I guess. But two things are notable: one, the Brooklyn location they are opening came after testing things out with a few pop-up stores. And two, they are not localizing the store in the sense of conforming to local tastes. They opted to be “genuinely Japanese” down to the Gapachon vending machines in the back, and featuring slightly less generally well-known properties like One Piece, that are meaningful to anime fans.
What Did We Learn This Week?
US consumers don’t like their future prospects, and if real numbers like inflation or the labor market start going downhill to hit those future expectations, we could find ourselves in a consumer spending crunch pretty quickly. For at least 80% of consumers we’re apparently already there.
GenAI really seems unlikely to save us from ourselves, and it’s back to costs as the reason why. It seems like the emperor has no clothes, and we’re all just milling around waiting for the little kid to pipe up and say it out loud. I also think that even if the people asking to see the money are wrong, it’s still not a certainty that GenAI delivers sustainable value. If our cell phones suddenly stopped working, people all over the world would notice immediately and revolt. If GenAI stopped working, a lot of the people using it right now would frown, shrug, and go back to the old way of whatever they were doing. Even more important, if GenAI companies started charging everyone a sustainable price, people would drop them like a hot rock.
And the tide continues to turn on what is required to become and remain a winning retailer. The interplay between Target and Walmart confirm that. Building a brand on passionate support of specific values can work – but it locks you in. Because the minute you step back from that support, you lose all your supporters without gaining any detractors. And more and more, retailers are proving that it’s technology and how they use it that is the differentiator.
Now if they’d only put that to use for enabling stores…
Until next week!
- Nikki