JUST INOur 2024 eCommerce SmartReport is here!

Learn More

12 Things That Won’t Happen in Commerce in 2024

Cover Image for 12 Things That Won’t Happen in Commerce in 2024

As we head into 2024, it’s common to make predictions — but what we thought we would do here is turn it around and take some commonly held beliefs and “unpredict” them — try to unpack why some of these assertions that often dominate a popular narrative may or may not frame the future correctly.

Bryan Gildenberg, CEO of Confluencer Commerce, breaks these up into 3 main (and obviously interdependent) sections: 

1. Technology — With an emphasis on the transformative nature of AI, machine learning, and ChatGPT

2. Marketing — With an emphasis on trends that impact retail media

3. Retail — Some basic truths about the US retail marketplace


AI's evolution mirrors other tech and will help us be better marketers.

To start, predictions around AI should be grounded in a solid definition of what people mean when they talk about AI. Too often AI is conflated with ChatGPT – a branded version of text-based generative AI, which is a subset of generative AI, which is a subset of AI… which is, in and of itself, a fluid computing concept on a continuum with machine learning, advanced analytics and basic predictive models.

AI in a variety of forms is being used, and has been changing critical workstreams, for years. Analytic and predictive models continue to transform quantitative analytics, and even within generative AI the most important use case for it may not be text-based, but image-based generative AI through platforms like Midjourney. Increasingly, major platforms like Adobe are incorporating AI generated imagery into their internal workflows, and Amazon is enabling first party sellers today to use its own tools to generate imagery for its product description pages.

All of this automation seems, at first, as though it replaces an enormous amount of work today done by people – notably copywriters, designers and visual production teams. This, of course, ignores what the quantitative side of AI will allow, which are more meaningfully differentiated behavioral and contextual segments or audiences of consumers. These small-bore segments – historically unprofitable to reach or develop campaigns for – will be targeted and served content more effectively by AI powered engines that don’t see making 1,000 variations on a theme to be significantly harder than making 2 or 3. Once this power is harnessed, we will return to the paradox of technology – what at first appears to be removing work is in fact creating more. 

As far as it being the metaverse – there is one area where AI could run the risk of turning into that – which is where companies who are given a “check the box” type assignment from senior leadership on “develop me an AI strategy” in much the same way “develop me a metaverse strategy” was a board level request in the past. The base case for AI vs. the metaverse is that AI both has and replaces existing practical applications whereas the metaverse was more often trying to invent use cases from whole cloth…that’s infinitely harder. But, our second “unprediction” tackles this misconception about the metaverse.

Focus on marketing to the multiverse: personas matter more.

Well, again – definitions are critical here. The Oculus-based, Meta-powered version of the metaverse is certainly off to a slow start but there are two critical things to remember here:

1. Other hardware devices (like Apple’s Vision Pro) will change the game with this type of metaverse experience.

2. The metaverse doesn’t need to only be this type of VR – broadly speaking it may make sense to understand the metaverse as any ecosystem a shopper spends an extended amount of time in an immersive way in which they can, if they choose, adopt a persona that is defined by their choices not their inherent attributes.

To that end, the idea of digital personas will become critical – in a conversation with an Amazon executive this week he referenced one of his colleagues calling Amazon’s Marketing Cloud “the multiverse” – i.e. the number of behaviorally and attitudinally based stories you can develop around an individual within that data set are almost limitless. 

“Machine teaching” will be more critical than “machine learning”.

OK this one is true, but with a crucial caveat – some of those new skills are new to organizations but they shouldn’t be. The truly critical skill with AI is the same critical skill with any technology, which is ensuring that the technology and the teams responsible for it understand the business problems it is required to solve. In the same way that Amazon wins in the marketplace because its people are better at teaching development teams how to build digital products that power the business (i.e. they have better “business requirements document writers” that other companies, so they build more relevant technology faster) winning companies in AI will be ones that excel at the “Training” part of AI as much as the ”Inferencing”. Retail expert Andy Murray has gone so far as to say that the “new 5th P” of marketing should be the “prompt” – how good your people are at asking platforms like ChatGPT what to do.


Marketers must adapt to a broader skill set to stay in the game.

In this case the prediction isn’t an “unprediction” because ChatGPT won’t change search – it will. The biggest shift however is the shift that’s already taken place in the commerce world, which is that Google has been largely disintermediated as a primary tool for shopping by a wide range of platforms that already exist at scale:

· In 2022 56% of all shopping related searches started on Amazon, vs. 21% for Google (and 12% for Walmart).

· A Search Engine Land survey revealed that 51% of Gen Z women prefer TikTok as a search engine to Google, and TikTok’s own research suggests that over 50% of the new product discovery done in the US is done on TikTok.

· Walmart.com has passed Google as the 2nd most widely used source for shoppers for both price and review comparisons.

Media teams will face the dilemma of choosing from emerging but imperfect ecosystems.

This notion, which the retail media networks are working hard to disprove, is fundamentally driven by the assumption that there is no waste in the media ecosystem that could easily be repurposed to something more productive. According to WARC, the global media landscape is fairly simple for a $1 trillion ecosystem – 85% of the spend goes to 4 places:

1. Search, social and TV are about 23% each. TV is still 83% linear, 17% connected.

2. Retail media is about 15%. The overwhelming share of that is Amazon in the US and the Chinese commerce platforms.

Today there are two distinctly vulnerable pieces of this:

3. Linear TV is the obvious one — and the degree to which this spend is liberated is the degree to which major brands can unwind or alter the assumptions their Marketing Mix Models are based on.

4. Search – Put simply, search in the retail/commerce ecosystem is broken. Some brands have abandoned thinking about search as a media tool altogether, and simply refer to it as an aid for “physical availability” – it puts products in front of shoppers.

Seamlessly capturing attention and driving action is key for marketers.

Personalization is a tool, not an outcome. 

The outcome may be something like relevance, but increasingly what anyone is trying to capture in media today is the scarcest resource imaginable in marketing — attention. Using attention as an advertising currency has been promoted by wise leaders like Chrissie Hanson at OMD for a couple of years now.

This becomes particularly critical when we realize how much human attention is shifting. Platforms like TikTok, Pinterest, Twitch, YouTube, Connected TV, Spotify — all of these share something in common from a media perspective. They’re all a measurement challenge for media teams. As a result, for most brands in most markets, these platforms are all hard to measure. One of the roles for off-platform retail media and the promise of closed loop measurement may be to try and improve conversion signals from these types of hard to measure gatherers of attention.

On vs. off platform segmentation: Retail media will recognize their distinct roles.

The problem with this prediction isn’t that it’s directionally wrong but that the counting for retail media is somewhat poorly formulated. There are two challenges with the metric, both largely related to Amazon:

1. Over 75% of Amazon “media spend” is on sponsored product ads — the degree to which this is “media” vs. “merchandising” is certainly up for debate but the role of a sponsored product ad is often way more similar to the role that, say, an end aisle display, would perform in conventional retail. The notion that Amazon collects more money than, say, Walmart’s total ecosystem, for this type of “merchandising” is simply absurd for most brands. And the effectiveness of retail media needs to be evaluated within the comparative framework of other tools trying to do the same thing – end aisle displays, for instance, are among the most profitable promotional executions in the world when done in the right categories and executed well. 

2. Amazon owns a variety of fairly conventional media properties today — notably Thursday Night Football. Treating this as anything different than a television program makes very little sense. 


Contextualization will replace personalization.

Personalization is not the key, but this time for a different reason. Not only is personalization just a tool, but in retail’s case, it’s the wrong tool.

Personalization still traps most marketers into the trap of trying to know “who you are''. In the commerce ecosystem, often knowing “where” you are (either physically or in a journey), “when” you are (relevant day part, season, or mode), or “what” you’re buying are way more important than classic personalization. All of these are better captured by the concept of “contextualization” rather than personalization. Contextualization allows for the “multiversal” approach to an individual that was alluded to earlier — you can be as different as you want to be from you when you’re buying a sofa vs. running to the restroom in a roadside gas station trying to remember the kid’s snack orders to buy on the way out. And you can know that you’re probably more similar to a person doing that thing at that time, whoever they are, than to someone who “personally” is almost identical who is doing something different.

Instacart will become the hot spot for brands reaching shoppers.

There are 4 things true about Instacart that make this unlikely:

1. They have a better user experience for the act of shopping for groceries than any other platform in the US — its platform was built for buying baskets, not items.

2. They have an easy model to monetize charging a retailer’s shoppers who are willing to pay more for high end service — retailers struggle with this (Walmart’s yeoman efforts to get Walmart+ off the ground are a good example of this challenge).

3. They solve this problem for 2 extremely successful retailers (Costco and Aldi) who do not want to solve store-based eCommerce delivery themselves.

4. The service Instacart undertakes and monetizes is, for retailers, a massive loss maker for them today.

Instacart is also dramatically upgrading its ad products as it turns its commercial model into an advertising platform not a consumer/retailer service model.

Premium solutions? Thriving against the odds.

In the immortal words of Dianne Wiest’s character in Parenthood, “there are so many things wrong with this I don’t know how to separate them all”.

· Inflation in 2024 in most categories in the consumables commerce ecosystem will be back close to zero.

· Value-centric retailers struggle mightily when the economy gets challenging — even Walmart during this downturn is benefiting more from the efforts it made during the pandemic to improve their Online Pickup grocery model (allowing more shoppers to use Walmart without having to go into the store) than with their “value” proposition. The trade down benefit is never enough to offset the damage done to their core shopper’s purchasing power.

· A lot of the growth in categories during this difficult time has been in the premium end of categories — as prices have changed often the price gap between everyday and premium in both absolute and relative terms has narrowed, leading to more easy trade up opportunities for more affluent shoppers less worried about overall rising prices. 

eCommerce and brick and mortar are both here to stay.

Here it seems simplest to use Retail Cities 2023-2028 predictions as the key predictions:

· eCommerce growth will continue at about the same speed it always has if you look at it right — it will gain about 100–150 basis points of market share every year.

· This consistent dollar amount off of a higher base will yield a lower growth rate every year even though its share growth rate remains constant.

· Over 50% of the retail growth in America over the next 5 years will come from store-based retail.

· About 50–60% of all US consumables’ eCommerce will be distributed through a physical store and in all likelihood picked off a physical shelf.

Boosting eCommerce profitability might not involve category-level fixes after all.

This un-prediction is not to suggest that this does not need to happen, but simply to suggest that the best way to make eCommerce more profitable may not be to try and solve the problem at a category level at all. In eCommerce there’s no physical asset that needs to be paid out or evaluated. And there’s no reason why in an eCommerce world any one category should be “better” than any other since most of the interaction with a commerce site happens at the search level (where what comes back is, in effect, a contextual “micro category”) or at the product level. Categories are a legacy of a world that was defined by shelves and stores as the primary framers of choice. 

Therefore trying to improve a category’s performance may not be the most effective way to accomplish the objective of driving profitability in a digital world. In all likelihood the solution will lie in two pathways forward — shoppers and baskets. Knowing the shoppers that have the highest propensity for incremental profit (your “Profit Potentials” — shoppers that are underbuying high margin categories, private label or non-promoted SKUs) and, more importantly, the products that can be sold to them that drive incremental basket profitability will be the key to success. Also, having an activation platform that can enhance basket profitability becomes critical both for retailers and the brands looking to partner with them.

Hopefully over the 12 days of Christmas, you can enjoy these 12 “unpredictions” as food for thought for the holiday season. Enjoy the time with family and friends, and see you holiday shopping!

Be a SmartCookie. 

Sign up to be the first to know about new products, events, and our monthly newsletter.