Welcome to The Martech Weekly, where every week I review some of the most interesting ideas, research, and latest news. I look to where the industry is going and what you should be paying attention to.
👋 Get TMW every Sunday
TMW is the fastest and easiest way to stay ahead of the Martech industry. Sign up to get the full version delivered every Sunday for this and every TMW, along with an invite to the TMW community. Learn more here.
Some news: I’ve partnered with WARC as their spotlight editor for Australia on a new series investigating how recent developments in privacy regulation in Australia is changing the marketing and advertising industry. Link
A new frontier for personalization
Personalization is one of those things that has been around since the inception of the internet. Since the early 2000s, the approach to increase customer relevance with content, products and experiences has gone through a variety of iterations.
But every once in a while, a totally new paradigm hits the internet and causes a series of uncontrollable second-order effects. That new paradigm is Generative AI and Large Language Models, and it is pushing how we understand personalization to new limits.
Up until this point, personalization has meant a couple of things in the context of marketing technology, but the thing that ties personalization together is the company’s efforts to harness data, tech, and content to make it happen. Usually construed as “right person, right time, right message,” companies have sought to bring greater relevance to a customer, an end user, or an audience. The end goal is to drive greater value for customers and better commercial outcomes by making experiences, products, and marketing messages more engaging, helpful and compelling.
To do this, marketers rely on complex data structures, identity resolution, and data activation tools to personalize something for a person. In a lot of cases, pricing and product offerings are dynamically changed to suit customer tastes, browsing history, or propensity to convert. Content changes based on what we know about a customer and ads are relevantly targeted to users based on a variety of data points.
Things like Send Time Optimization dynamically sends email content to users based on their best time of day to engage. Apps can be used to send you a notification based on your geolocation. Spotify Wrapped is an example of mass personalization as a brand campaign, showing the listener that Spotify deeply knows your preferences and listening habits. Over the years, personalization has branched out to a wide variety of experiences for consumers and has grown exponentially.
The demand for personalization has also grown beyond the technology category. Consumers are now wanting to have personalized experiences. Gartner suggests that 71% of B2C and 86% of B2B customers expect companies to utilize their personal information when interacting with them. Marketers that do well with personalization programs can see more than a 40% increase in revenue compared to most other companies.
In almost all cases, brand investment into personalization has created a positive sum outcome for both marketers and consumers – marketers grow their brands and consumers get a better experience. And the list goes on of strategies and tactics to make the web a more personalized place for the everyday person. The silent revolution inside the four walls of enterprise brands has been the increased focus on harnessing data to communicate to customers in increasingly relevant and personal ways.
But here’s my rationale –there’s a good chance that our current way of thinking about how personalization works will change in the coming years. Let me introduce you to something I call volitional personalization.
The frontier of choice
The idea of volition is the choice-making capacity of a person. Enabling a person’s ability to make their own decisions as opposed to reacting to stimuli is the next frontier for personalization and an interesting behavioral model for how consumers interact with content, products, and services.
The assumption that many people working in Martech carry with them is that personalization is about what we think the customer wants. Marketers infer that through explicit and implicit data collection. And that assumption has worked well over time; brands build data, experience, content, and process capabilities to show customers a product recommendation carousel, a personalized offer, or a tailored message that has overall improved experiences.
But underneath all of these things is complex data orchestration, alignment of teams, advanced analytics, and a wide variety of technological and strategic changes to get to a personalized experience for a user that’s meaningful and valuable. As one example, a study into telecommunications companies by McKinsey suggests that only 5% of companies are fully utilizing their data for personalization. And companies that attempt to do personalization but go no further than your simplistic and superficial tactics end up costing brands more than it's worth.
The next frontier – what I’m calling volitional personalization – is the idea that when it comes to interacting with a brand, a customer should be able to customize not only their experience but what product they want to buy. Going beyond the data brands collect, customers may in the future be able to construct their own products and experiences that suit them, right down to the most granular details. Rather than have personalization based on what a company thinks a customer wants, customers can use their own volition to tell brands what they want.
This gets us to true personalization – a product, service, or experience that matches the specific preferences of a single person. That’s why it’s called volitional personalization – it’s an experience based on the consumer’s true, explicitly stated preferences. In other words, the shift is a movement from generative search to generative products – customers using AI and new tools to totally personalize their interactions with brands in the way they want it.
One example is the collision between generative AI and fashion. Shien is one of the world’s fastest-growing fast fashion brands, with the ability to rapidly turn around new styles and operationalize them into its supply chain, adding 5 to 10 thousand new products to its site every day, Shien could have a vertical integration with a generative image tool that allows customers to design their own styles for them to be shipped the next day. The point here is that the entire experience of shopping for fashion changes in an instant by giving customers the ultimate choice over a personalized product. Instead of asking customers to browse aisles or rows of SKUs on a category page to find something they like, why not ask customers to create what they want?
There’s an opportunity here for brands to give all choices over to a consumer. If you’re looking for a dress with polka dots and frilly shoulders, it’s not hard to see how using text prompts and image prompts to design what you want could be something that is turned into a physical product quickly and at scale.
Another example of this is the Arc browser’s ability to easily customize any website to your own liking, called Arc Boosts. Don’t want to see a feature in a product? Or want to remove all ads from an article? How about changing the flow of how something works on a website? New browser concepts put low-code tools into the hands of anyone to use. In this example, there could be literally millions of variations of everyday websites that are personalized to the tastes of each individual consumer – and this has nothing to do with the efforts of the brands.
The other edge is customized customer service, with LLMs and intelligent customer service agents, people don’t need to jump through pre-defined hoops or stay on the phone forever to talk to a real human, they can have a realistic conversation with a real-seeming person with all the nuance and intonation needed to provide a good customer service experience. A newly launched app, Shop Agent, is attempting to plug into the Shopify ecosystem to provide this kind of service for millions of small to medium-sized e-commerce platforms.
This week, Open AI launched plugins for its wildly successful text-based generative AI product Chat GPT. Plugins integrate with other services and websites to run highly complex tasks for people. In the example below, a plugin with a recipe website and Instacart gives the customer the ability to bring up a recipe using a text prompt, customize the recipe as they see fit, and pull up a customized ingredient list from a grocer. Clearly, this is a superior experience for meal planning and grocery shopping, but it has nothing to do with the actual companies that provide the services.
This goes beyond the ability to take a customer’s available history and data to show the customer that they are known. This kind of capability doesn’t need to orchestrate huge volumes of data or tools to activate it. It just needs an intelligent agent to act on your behalf. Talking to a customer service representative who is looking at a computer screen is no match for a generative AI that can add suggestions and sympathies at the moment it matters the most.
As personalization continues to mature, the importance of volition will become a step change in how consumers are increasingly thinking about relevance. Last decade something relevant would have been a tailored ad, product recommendation, or a personalized email subject line. This decade it will be products, content, and designs that are created based on exactly what you asked for.
Don’t we already have this?
You can argue that volitional personalization is something that already exists. More than ever, customers have a choice over how they shop and where they shop, and to customize their experiences.
The primitives of generative products are everywhere in the market. Sephora’s Color IQ tool scans a customer’s face and gives personalized recommendations that attempt to remove the trial and error of makeup shopping. Another one is Uniqlo’s in-store uMood kiosks that show customers a variety of products and measure their engagement to filter products. Both bring together a variety of data points to make the shopping process easier for consumers without having to overly rely on collecting a lot of data about them.
While these examples are interesting ways to do personalization, the leap forward to volitional experiences is about removing the constraints that have typically characterized how brands offer products to consumers. You can only still select what’s in stock with Sephora’s Color IQ – it does not open up the possibility of entirely new color combinations based on a customer’s unique complexion.
Chatbots could traditionally only handle a small amount of pre-defined inquiries, and website personalization strategies were limited by the types, quality, and quantity of data the brands could collect. Allowing the customer to really choose something unique for them is letting go of all the constraints, pre-defined pathways, and assumptions made about what a person wants.
In this way, volitional personalization technology becomes an open canvas where what a person wants at that point in time can become a reality in a relatively short time. In very similar ways to how ChatGPT and Bing work, there are limitations to the tools across ideological, harm reduction, and access to real-time information, yet it’s the most sophisticated example we’ve ever seen of a technology that can serve information to a person across almost any conceivable domain of knowledge.
Aaron Spinley captures this progression of personalization as a pathway into individualization, suggesting:
“Think of all those customer-facing or enabling technologies that we have. From systems of record like ERP or sales and service CRM, to marketing automation and transaction engines like digital commerce, all the way to our ad-tech and beyond. They all represent an 'inside-out' operating model. It is marketing’s digital expression of the industrial era. This is all about optimizing a process toward an internal objective, by imposing that process on the customer. As a result, brands have had to be channel-specific, and tactical. But the world has changed. The demands of modern and connected society are a disruptive force on that traditional mindset. It is the customer, not the company, that is the sole arbiter of whether something is 'personal.'"
Indeed, if regular people continue to use generative AIs and if the category’s growth rate is any real measure of progress, then the mindset of what is personalized will not be what a brand decides it will be, it will be what someone asks an intelligent agent like ChatGPT, Google’s BARD or Microsoft’s BingChat. The goalposts of what actual personalization is are shifting in real-time, and all the questions become less about how companies can collect data to facilitate these experiences and more about what kind of LLM you want to deploy to meet consumer’s demands for better choice and customization over their experiences and products.
Personalization without data
The power over the lateral mastery of knowledge that new generative AI technologies represent can easily be migrated over to brands of any size and scope. GPT-4 is already being integrated into a wide variety of preexisting and new products. This means that in the near future, it’s possible that brands won’t have the imperative to collect customer data for personalized experiences, more relevant ads, or even for customer insights that might inform product or marketing strategies.
The way marketers think about opportunities is usually about what makes sense for the mass – where the largest volume of behaviors, preferences or purchasing histories are. But with volitional personalization, what makes sense for the masses no longer matters. As consumers can choose what they like without constraints, everything is brought down to an individual 1:1 level where billions of combinations of experiences, messages, and products are served to customers with little to no control from the brand’s perspective other than hosting the platform in which consumers and intelligent agents can connect. Imagine for a moment giving the customer the ability to create a new feature in a product that’s unique to their needs. In this kind of future, the experience or even the product itself be entirely directed by the customer.
Of course, this kind of thinking, even if it hits 20% penetration in most consumer interactions, has the potential to implode a large chunk of the marketing technology industry. About 59% of the marketing technology landscape is made up of data-intensive companies that rely on the premise that brands will want to continue to collect data to enable the wide variety of personalization that we see today.
Everything you’ve ever wanted
The elephant in the room with volitional personalization is that giving people exactly what they want all the time does not lead to good outcomes. Introducing friction to things like being able to buy guns or hard drugs is good for society. And while I’m not advocating that the next horizon of personalization will be making those things freely available – we will forever live with societal-level constraints – I am arguing that within the realms of what’s legal and possible, it’s hard not to see that giving consumers more control over what they consume and the products they buy can make us more impotent, stupider and more invested into things that we probably shouldn’t get into.
Social media is a good prior for what happens when algorithms are pointed at people with the express goal of giving people more of the content that they want. Expressed through everyday people’s behaviors and engagement, you can see all kinds of anti-social, detrimental, and harmful behaviors. One example is my wife teaching a student that identifies as a Furry – an online community of people that genuinely believe they are some kind of made-up animal. This is arguably the kind of dysfunction you get when people are increasingly pulled into algorithms that constantly feed people the content they're interested in, for good or ill.
Yet people time and time again signal that they want content and information to be personally relevant to them. Dealing with irrelevant information used to be a normal part of human life, but now it’s something to avoid like the plague. When it comes to interacting with brands, what we want is for those interactions to have some kind of personal connection with us and what we want.
There’s a strong argument to say that product discovery will remain one of the main ways consumers interact with brands. That’s why grocery stores put the most regularly purchased items at the back of the store, or why the retail storefront displays products for people to find. This is because people don’t know what they don’t know. Inspiring consumers with new products and experiences - things they would never choose - will continue to be a pillar of all marketing activities.
But as AI-powered choice machines come into the market and start interacting with the marketing and CX departments, it’s easy to see how people will easily adopt this kind of technology to make their own experience more convenient, interesting, and creative.
The age of infinite choice
Not every brand will qualify for volitional personalization, nor will every consumer want to use it. As a society, we’re used to having things handed to us with very little choice or control over what those things are.
But if consumers can design or modify something they want and purchase it, or if there’s a better way to experience something online, then there will be people who will line up for that. It’s only a matter of if there’s enough value to capture and in what industries it makes the most sense.
In other areas, outsourcing customer service to intelligent automated agents is a form of dehumanization, but if it gets them the answer they need at lightning speed with context, nuance, and empathy (although it’s not real), then is it really dehumanizing? It’s likely that to most people, it’s the difference between a high-quality service agent and a bad one. In most cases where speed and level of help matters, that a human exists on the other side is not a relevant factor.
If the next frontier of personalization gives consumers infinite choice, there will always be a few universal constants. People will always want to save money and time; they will always want to own things that are unique to them and they will want to have their problems solved quickly. If the shift to the next frontier in personalization becomes a reality, then this future will have to fit into these boxes. Everything else will be a waste.
Make sense of marketing technology.
Sign up now to get TMW delivered to your inbox every Sunday evening plus an invite to the slack community.
Want to share something interesting or be featured in The Martech Weekly? Drop me a line at firstname.lastname@example.org.