Ethical Personalization in Digital Commerce

(and the future of Meeting Consumer Intent)

Imagine this scenario: a man named John is getting ready to spend the day with his 7-year-old daughter Emma. As he gets situated in the morning, he uses his computer to check the weather, browse the news, and find directions to a local playground. Unbeknownst to John, a man named Terry stands just outside his window.

Terry is employed by the local chamber of commerce and part of his assignment is to follow John and his family around and gather information on them. With that information Terry reports back to the chamber who then sells that information to local businesses who might be interested in certain activities or interests of John and his family. By purchasing that information, the local businesses are now able to personalize their promotions, offerings and services directly to the needs of John’s family.

Let's return to John's day.

As John uses his devices, Terry peers through the window, meticulously noting John's every action. Once John wraps up his preparations, he and Emma head out for their day's adventures.

On the drive to the playground, Terry tails John and Emma closely, documenting their every move. Pulling up alongside them, Terry spots Emma gazing out the window at some digital billboards. He immediately alerts the chamber about this ad placement opportunity, targeting a young child's profile. The chamber swiftly contacts local toy stores, auctioning off the nearest billboard space. A local bike shop wins the bid, and an ad for a scooter instantly appears. Terry watches intently to see if Emma notices. When she doesn't, he informs the bike shop about upcoming billboards along their route, ensuring their scooter ads are displayed. Terry diligently records Emma's reactions, making a mental note to follow up on whether John eventually buys Emma that scooter.

Finally, John, Emma, and their unseen companion Terry arrive at the playground. As John and Emma head for the swings, Terry finds a cozy spot behind a tree and pulls out his binoculars. While observing them, he notices other individuals hiding nearby. Upon inquiry, he discovers they're also monitoring John and Emma. They exchange notes and data, building a more comprehensive profile of John's family. This understanding, they claim, helps them better serve their customers.

Eventually, John and Emma pack up to leave. Terry and a couple of other watchers gather their belongings and continue their assignment. On the way home, John and Emma stop for ice cream. Terry notes the shop's location and the amount John spends. This information will help determine what ads John might see and how much he might be willing to spend on advertised items.

After ice cream, John and Emma visit a toy store to window shop. Terry, busy relaying the ice cream information to the chamber, doesn't worry about missing this stop. Another watcher has already noted which toy store they entered, its exact location, and product offerings. This watcher passes the precise information to Terry, who then reports everything he's gathered to the chamber of commerce.

With today's data haul, the broker now possesses an extensive amount of personal information about John and his family—and that's just from one day of surveillance. They justify collecting and retaining this data by claiming it allows them to better curate their offerings and ads, presenting items John might actually want to buy or engage with.

Alright, let's pause this scenario.

If this sounds unsettling, that's because it is. In our society, we have a term for such behavior: stalking. The tech industry, however, prefers to call it "personalization".

This scenario is my adaptation of a digital situation described in a paper titled "A Day in the Life of Your Data" by Apple—ironically, a company that provides the means for much of this digital data capture (or stalking, if you will). In their paper, Apple uses the fictional story of John and Emma's device usage to illustrate how, in reality, companies can amass thousands of personal and unique data points from your devices, often without your knowledge or consent. This data is then sold to third-party companies and data brokers, who have compiled thousands of data points on millions of people. And this happens every single time you interact with your devices.

Digital stalking... er, personalization.

In their paper, Apple claims to offer two options: first, to "Be aware that this is happening," and second, to "Identify that there are ways for you to opt out of this." While these points are helpful, they fail to acknowledge Apple's role in creating the infrastructure that enables much of this digital stalking. The potential for data tracking is built into the apps in their app store and the browsers on their devices. They also subtly suggest that if John had used their products for his digital activities, he would have had an extra layer of protection. It's a clever pitch: "Use our products, our apps, and our online services, and you'll be protected. Use our products, but not our apps and not our online services, and your personal information is at risk."

But this underhanded scenario isn't unique to Apple. Consider Google: of the $305.6 billion in revenue they generated in 2023, $175 billion came from Google Search. That means over half of their revenue stems from selling advertising space (and your personal information) on their platform. The list of tech companies profiting from personal data goes on and on—it's safe to say the issue is widespread.

What's intriguing is that at this point in history, this information shouldn't come as a shock to any of us. Perhaps the extent to which these companies use and sell our personal data is news, but by now, we're all somewhat aware that our attention and data are being bought and sold en masse.

Before we delve into the 'whys and whats' of all this, let's explore how this data collection paradigm came to be in the first place.

The practice of merchants gathering customer information to inform selling strategies is as old as commerce itself. However, the large-scale collection of personal (i.e., non-anonymous) information can be traced back to the US government in the early 20th century. FDR's New Deal gave birth to Social Security, and administering benefits on such a scale required the government to collect certain information about every American citizen. This necessitated efficient record-keeping. Who better to assist than a company whose origins lie in the invention of a tabulating machine designed for another massive data collection effort: the US Census? This up-and-coming company was none other than International Business Machines, or IBM.

As technology advanced and the need for information grew, the US government began using mainframe computers to store and process vast amounts of data on nearly every American. By the 1960s, a substantial amount of government-sponsored research was being conducted by social scientists who lobbied to consolidate hundreds of siloed databases into one National Data Center under President Lyndon Johnson's administration. Privacy advocates pushed back, and the proposal never materialized. However, the potential value of consolidating this data didn't go unnoticed.

Soon after, corporate America recognized the potential in this massive trove of information and gained access to these databases. By the 1970s, they were using this data for various purposes, particularly to "sophisticate" their direct marketing campaigns. Regulations intended to govern personal data use were passed from the mid-60s to the mid-70s, but according to historian Margaret O'Mara, "all of them focused on individuals' right to know about the information these databases held. None addressed the question of whether this information should have been gathered in the first place."

Fast forward a couple of decades to the age of internet proliferation (aka, the '90s). The internet allowed anyone with a connection (soon to be almost everyone) to exchange information—whether through purchases, communications, or simply browsing. As with any medium that attracts a large audience, the advertising industry saw an opportunity and paid handsomely to display ads on sites whose owners were eager to monetize their growing audiences. Initially, these ads were disconnected from any understanding of the viewer's personal identity, but that changed in 1999. DoubleClick, a digital ad company, attempted to merge with a major data broker to "de-anonymize its ads". This move sparked outrage, with privacy advocates petitioning the FTC, which ultimately blocked the merger. This led to a trade group establishing the first standards for online advertising. However, as we've seen before, just because it didn't happen then doesn't mean it couldn't happen later. The potential for monetizing this information was clear to those who stood to benefit.

Nine years later, that potential was realized when DoubleClick was acquired by none other than Google. In 2016, Google quietly dropped its ban on linking browsing data to personally identifiable information. Almost imperceptibly, we entered our current paradigm: one where we allow thousands of personal data points to be collected, stored, and used by an unsettlingly large number of organizations.

The "what" and the "why" behind this data collection are actually quite simple in concept. Much of it is an attempt to better achieve what we started out discussing: personalization. Improved personalization has been commerce's ultimate attempt to marry the two major forces that drive and influence all trade: Merchant Intent and Consumer Intent. Or, in more familiar terms, Supply and Demand.

Merchant Intent (i.e., Supply) can be distilled to the goals and objectives of any given business. It could be a quota that needs to be met, a specific product that needs to be sold, or shelf space that needs to be leveraged. It encompasses whatever the merchants intend to sell and how they plan to do so. But Merchant Intent is meaningless without a buyer on the other end who would actually make the purchase. This indicates that transacting from the buyer's side requires a very specific element: intent.

Consumer Intent (i.e., Demand) is what drives a person to engage with and ultimately purchase a good or service. This intent is typically tied to a need and thus is definitive, but it can also be manufactured through sufficient influence on the buyer. That influence has a name: marketing. Most of us have experienced succumbing to a purchase we knew we didn't actually need but still wanted to make. However, for the most part, whether intent is expressed as a felt need or is simply an interest being monetized through clever marketing, that intent being leveraged must resonate with the buyer. Simply put, Demand cannot be met with just any Supply that the merchant chooses to offer.

So, to identify what types of intent a consumer may have and then meet it at precisely the point where that intent leads to a transaction (i.e., a purchase), merchants must get to know their consumers better. And in recent years, the primary tools for doing so have been—you guessed it—data collection and personalization.

…the primary tools for doing so have been data collection and personalization…

Is that necessarily a bad thing? Well, not entirely. For starters, personalization in the technical sense allows us to maintain a high level of continuity and consistency on digital platforms, which can greatly enhance our user experience. For example, many of the sites and platforms we use store information that makes logging in and checking out effortless, as we no longer need to input the same information every time. They also store information about our search history so that future results become more aligned with what we'd like to see or buy—or rather, what our intent might be. This same concept can backfire in certain scenarios, particularly with search and social media algorithms, where a person's ideological leanings are reinforced with supportive rhetoric and bias that goes unchecked and unchallenged. But that's a topic for another discussion. This continuity and "search curation" can be helpful in many ways, but ultimately the benefits cease when we move from one platform to another, as each has its own method of tracking our information.

However, companies have found a way to mitigate this loss of consumer intent signal by gaining additional access to information about their customers that they didn't have to collect themselves. These efforts are precisely what the opening story highlights from Apple's "Day in the Life of Your Data" paper. Most companies manage this information storage and access using a tool known as a Customer Data Platform, or CDP. According to the informational site CDP.com, "CDP platforms were created to manage first-party, second-party, and third-party data from multiple disparate channels, unite customer-centric efforts across marketing, sales, and customer service."

In this paradigm, entities can collect your data not only to store and use in their own CDPs but also to share access to this information with other companies. In some cases, they sell it to data brokers who "gather information about a customer's behavior across multiple interactions with various entities — the credit card issuer, car dealership, online shopping site and others," according to a paper on data privacy written by Penn and Wharton experts. This practice of data collection to enhance personalization undoubtedly benefits both the top and bottom lines of these companies. But what do consumers stand to gain?

Living in a paradigm where information is used to create consistency and better-curated experiences may seem preferable to the alternative, but is there a better approach? And is there a way to accomplish these goals without leveraging people's information against them?

At Syntheum AI, we believe there is a better way, and we're convinced that by leveraging AI and machine learning technologies, we can achieve it. Revisiting a point made earlier, consumer intent is often definitive, meaning it typically stems from some felt need on the consumer's part. These felt needs are often not random. So what does this mean? It means that patterns of behavior are likely to emerge in the process of identifying consumer intent, and the more patterns we can recognize, the more insight we gain into predictive outcomes.

Thanks to machine learning, we can anonymously track consumer behavior across a given site and then detect patterns that point to more precise outcomes as the site is used more frequently. To give a simple example, if you land on a clothing website, click on "Men's," scroll for a bit, and then search for "brown shoes," chances are you're looking for men's brown shoes, and the displayed results should reflect that reasoning. To delve deeper, let's say you're on that site intending to purchase a suit for a wedding. You've browsed a few different suits and even added one to your cart. Now, when you search for brown shoes, instead of being shown an arbitrary selection of men's brown shoes, you might be presented with brown shoes that other customers bought when they purchased or added to their cart the same suit that's in your cart.

These are just basic examples, and the list could go on and on, but pattern recognition tied to click stream data collected anonymously on a site can lead to some incredibly accurate predictive models. These models ultimately benefit both merchant and consumer while preserving the consumer's privacy. The question we should ultimately be asking isn't how much privacy should be enabled, but rather how much value can be provided regardless of the specific knowledge about the individuals.

We call this approach Ethical Personalization, and we believe that building tools to enable it is the future of meeting consumer intent. By focusing on outcomes rather than profiles, merchants are compelled to drive value and not just seek "easy wins" via clever marketing or hyper-targeting. It puts the onus on learning how consumers express intent and then building strategies that require the merchant to deliver what the consumer wants. In this paradigm, the traditional Supply and Demand model is flipped on its head, becoming Demand and Supply.

The approach outlined above is just the first step toward a more ethical way of handling personal information. We believe that the next stage—a more powerful and efficient advancement in ethical personalization and meeting consumer intent—involves becoming far more personal, but from the customer's perspective rather than the merchant's.

This vision will be realized when customers have centralized ownership of their own data. This shift makes the process truly Personal, allowing customers to leverage their data for their own benefit. In contrast, the current Personalization paradigm puts data ownership in the hands of merchants and brokers, who, as discussed, use it to drive transactions that primarily serve their own interests.

Doc Searls—an author, journalist, prominent blogger, and advocate for open source and customer agency—has long explored this concept of "Personal vs. Personalization." He aptly describes the current situation: "All that stuff is data. But most of it is scattered between apps and clouds belonging to Apple, Google, Microsoft, Amazon, Meta, phone companies, cable companies, car makers, health care systems, insurance companies, banks, credit card companies, retailers, and other systems that are not yours. And most of them also think that data is theirs and not yours." (emphasis mine).

Revisiting John's story, we must acknowledge that John and his family already possess a clear understanding of their desires. They have inherent intent. This intent is acted upon every single day through information unique to their family. We believe the technology they utilize should facilitate meeting this intent based on their personal information without compromising its privacy.

Our mission is to develop this technology, as we believe it's the right thing to do and it’s the future of meeting consumer intent.