Is your phone listening to you?
55% of people think so - and marketing is only getting weirder. Value x Vibes Part 3.
Last week, I asked y’all to weigh in on an important question in my life:
I’ll admit, I’ve been writing “ecommerce” this entire time, in hopes that the time and energy saved from typing “-” repeatedly would lead to improve wrist health and allow me to achieve self-actualization. Unfortunately, I have been outvoted.
Google is still in my camp, finding both ways acceptable. Because I believe in listening to my audience, I will add a dash moving forward. Thank you for your feedback!
Today, we are moving on to the final quadrant of the Value x Vibes matrix for AI adoption in e-commerce. I predict that AI will be used differently for different categories of products, based on what we’re looking for when we shop for these products.
Part 1, covering “Boring” categories, is here. I covered how AI tools will enable automation and personalized research and product comparison to improve the customer shopping experience. These tools will be leveraged by brands and retailers to differentiate and steal share.
Part 2, covering Fun/Expensive categories, is here. I focused on how AI is less penetrated in these niche, expensive categories, but early movers are promising to democratize tools once only available to the wealthy - like the ability to source rare vintage luxury products, access a personal stylist, or even curate an art collection.
Some readers have taken issue with the products grouped into “Fun” and “Boring.” Several people seem to feel strongly that strollers belong in the “Fun” category. As someone who watched hundreds of Youtube and TikTok videos trying to understand what a “travel system” is and how that is different from a car seat, I beg to differ. But this feedback points out something important about the framework. Brands selling a product must understand how consumers shop for that product. Perhaps some customers approach a product as a “Fun/Expensive” purchase and are looking for the Brand to serve the role of Curator. Perhaps others categorize the purchase as “Boring” and look at the purchase journey as a way to maximize utility - seeking product comparisons and research. Many brands may have segments of customers inclined to shop differently - and must offer a suite of tools that enable multiple types of shopping journeys.
Fun/Cheap: Serendipitous moments
In the Fun/Cheap category, we’re looking for inspiration. We’re scrolling on our phones after work and shopping for a cute hair clip, a new book for our upcoming vacation, or a hostess gift. We’re open to buying right now…if we find something we like. In this category, we may not even “shop” in the conscious sense. The shopping comes to us via ads that seem to pre-empt our desires.
Brands are looking to present us with something that feels unique, special, and personal – a sweatshirt with the inside joke from my favorite podcast, or the candle that all my coworkers recommended, but I couldn’t recall the name of. Paid marketing has already advanced to the point where putting these things in front of us feels magical – 55% of us believe our phones are listening to us! And as AI sophistication continues to increase, this will feel even freakier.
AI in Paid Marketing
The serendipitous feeling of discovering the perfect product while browsing online is often powered by paid ads. Brands allocate budget to channels like Google, Meta, and TikTok to show relevant users paid “sponsored” content - in all, roughly 1 in 4 visitors to ecommerce sites come via paid traffic, with search ad spending alone topping $132B in the US.
In the heyday of paid media (2015-2019), brands used third-party cookies - small pieces of data stored on your browser when you visit a website - to track you across different sites and build a profile of your browsing activity. For those of us on Facebook during these wild times, we might remember the perennial print-on-demand ads for hyper-specific tees and mugs that would pull in your birthday and details about you from your Facebook profile. There was a sense of novelty in the early days of data-driven marketing that made products like this feel fun.
In a similar fashion, third-party cookies would compile a profile of you and data brokers and ad tech companies would sell it to Brands so they could serve you hyper-specific ads. In these peak years, paid media spend was growing at 49%, with 88% of ads dollars being spent programmatically.
In recent years, regulations have targeted third-party cookies, and Apple and Google are phasing them out. This reflects a value shift toward privacy-focused marketing. The idea that random internet brands could know everything about us no longer feels harmless - it feels creepy. For many Brands, paid ads have become less effective, and therefore more expensive. Increasingly advanced AI tools are promising to fill this gap with ever more sophisticated solutions.
Predictive Analytics and Targeting
In the world of third party cookies, a shopping experience might look something like this:
I visit Nike’s website and browse running shoes.
A third party cookie tracks my visit.
Later, I see Nike ads following me around while browsing the New York Times. This is called retargeting.
There are a few issues with this approach. Consumers might find it intrusive to have products they’d previously Googled show up all over the internet. I remember a colleague getting flustered when Victoria’s Secret ads showed up on a totally unrelated website while he was screen sharing during a meeting. He quickly protested that he was just buying pajamas for his wife - but we got ammunition for plenty of jokes.
Another issue is that I will keep seeing these retargeting ads, even if I already bought the shoes elsewhere. This is a waste of money for brands.
In today’s world, AI-driven targeting could look more like this:
I visit Nike’s website and browse running shoes.
Nike collects first-party data on my visit.
AI-driven targeting software uses Nike’s data about me and broader shopper data to predict what my future purchase journey will look like and inject the shoes at relevant touchpoints.
If I Google anything else related to running shoes, I might see the shoes I previously browsed (targeting based on relevant content)
Whether I buy the shoes from Nike or not, the prediction engine will phase out their marketing of those specific shoes and evolve to show me products that may be more relevant, based on my ongoing interactions
In short, old-school targeting was focused on what the customer did in the past. AI-based targeting is focused on predicting what customers will do in the future.
Everyone from data science startups to the Googles and Metas of the world are investing in capabilities for AI-driven targeting, as finding a unique edge to reach customers at the time of purchase is extremely valuable. While all brands are already employing some of this technology by virtue of running ads on major platforms, the level of sophistication will continue to increase.
Creative, Copywriting, and Personalization
Imagine you have a small healthy snack brand. You might have a few different groups of people who you want to advertise to:
Diabetics who need to control sugar intake for medical reasons
Health-conscious men who listen to the Huberman podcast
Crunchy moms who want healthy snacks for their kids
These groups will all need VERY different marketing materials when you create ads to speak to them. You and your team will need to create different visual “assets” for each group and different advertising text (“copy”). Now imagine you are running these ads across multiple channels - Google, Instagram, Podcasts - the amount of content you need to create will keep your team busy for weeks!
The promise of AI creative tools, like Adobe GenStudio or Google Performance Max for performance marketing, Canva, Synesthesia for video, and Jasper for content marketing, is that they can adapt your ad creative and copy dynamically for different audiences.
Google’s Performance Max (part of Google Ads) adapts a single set of ad inputs - like assets, copy, keywords, and headlines - into different ad formats best-suited for each Google channel. This saves work for creatives, and from Google’s point of view, makes it easier for Brands to spend money on Google.
Tools like Performance Max aren’t new - budget optimization and targeting capabilities have been around since 2021. Some of the creative updates have come more recently since 2023. Paid marketing has been deeply data-driven for years and this naturally evolved into AI-driven capabilities as the technology advanced. Many of these updates were largely “behind the scenes,” powering things like targeting (who sees your ads) and how spend is budgeted across a digital ad portfolio. Many of us are seeing these capabilities more now that they are maturing on the creative side - affecting the day-to-day work of creative and design teams.
What could the future of paid ads look like?
Since the majority of us already feel like our phones are listening to us - what’s next? This scenario illustrates how advances in predictive analytics, targeting, and personalization on the creative side could come together to power the next phase of digital marketing.
Returning to the Nike example - perhaps Nike has done a great job through their loyalty program, first-party site data, and community-building initiatives and has a bunch of data points on me:
I’m a runner
I’ll be running the Berlin marathon in the fall
I run in NYC and have preferred routes in Brooklyn - in the past, Nike had a partnership with Strava, though I don’t know the details of how data was collected and shared
Based on my purchase history, they know my general size and style preferences
Knowing all this, Nike could target me with personalized ads leading up to the Berlin marathon encouraging me to buy new shoes. None of this is novel, though few brands really nail occasion-based triggers.
Taking it to the next level, Nike could generate personalized copy speaking to the upcoming Berlin race. It could generate unique video assets to show a female runner in NYC and Berlin highlighting the product in my environment. Without crossing a line into “too creepy,” it could add a level of personalization to the ad that could help me envision myself in the shoes.
Going further, these ads could include chatbot-type functionalities that could let me learn more about the shoes and other related styles within the ad itself. It could let me swap colors on the model or see the key specs summarized.
Taken all together, an ad like this should drive significantly more conversion than a static picture of a shoe.
The bottom line - it’s all about data
One question underscores this scenario: How will Nike get the data to power this ad experience? These AI-driven marketing experiences are only as good as the data powering them. While Nike could theoretically get all of the data above via real, proven touch points - an online quiz, conversations with a store associate, a Nike-sponsored run club - the likelihood of piecing this all together under one clean customer profile is slim. Large marketplaces - like Amazon and Walmart - with millions of transactions per day across categories - are at an advantage. Smaller brands may struggle to glean detailed customer insights from infrequent visits and site interactions with a smaller product catalog.
I expect we will see a lot of innovation in this space to expand access to quality first-party data for brands, perhaps through partnerships. Brands are already focusing on loyalty programs and other sources of valuable customer data. Connecting real-life touchpoints in stores where possible could add another level of richness to the picture brands can paint of customers - but many brands operate stores entirely separately from their online businesses. Streamlining data collection efforts to create a “single view of the customer” will be more important than ever.
It’s ironic that in the “Fun/Cheap” category - the space where we shop on a whim for fun little treats - the situation behind the scenes is anything but whimsical. The seemingly magical ads that pop up on our phones are powered by cold, hard data - and getting that data will be a major strategic priority for brands hoping to optimize their approach to AI-driven marketing in the coming years.
Stay Curious,
Melina