You know the saying, “Well, now this is personal”? With advancements in technology and personal devices, there is a lot that’s now personalized. Sometimes people lament the popularity of Apple devices, complaining that we all now carry around the exact same phone, laptop, watch, etc., but even our generic-seeming phones are different from one another. Most people have backgrounds of their loved ones, their favorite views, or things that just make them happy. We can customize our ringtones (like my mom’s, which plays the Kim Possible theme song), our home pages, and the apps we download. 

Just a few decades ago, people purchased music exclusively by CD, record, or tape. This meant that you had to purchase an exact set of songs, and because CDs were expensive, you might not have had very many in your collection. Nowadays, with so many options for streaming services and the almost infinite number of songs available at your fingertips, it would be nearly impossible to find anyone in the world with the exact same music library as you. Even though we might choose to have many of the same devices, our technology use is incredibly individual, and the way we personalize those devices has become more and more elaborate. 

Personalization is a concept that revolves around the individual, highlighting the importance of cultivating an experience customized to the unique user. With more data available now than ever before, the applications of personalization have expanded to marketing teams and social media algorithms, improving healthcare, science, and athletic performance, as well as supporting education and learning. Although an excessive focus on personalization can foster a culture of extreme individualism, a thoughtful approach to personalization offers significant benefits.

Echo Chamber: The social media filter bubbles that reinforce existing beliefs, limiting exposure to diverse perspectives, and making it harder for people to take on a more nuanced view of complex concepts. 

Confirmation Bias: Our underlying tendency to notice, focus on, and give greater credence to evidence that fits with our existing beliefs.

Machine Learning: A subset of artificial intelligence (AI) that uses statistical techniques to enable machines to learn from data and improve over time, in ways that resemble human learning.

Reactance Theory: A theory that posits that when an individual feels that their freedom or control is being threatened by advice, they are motivated to protect their autonomy.

Market Fragmentation: A process in which a once relatively homogeneous market becomes divided into distinct segments, each with unique preferences, behaviors, or demand patterns. These segments often require tailored marketing strategies to effectively reach and serve them.

As long as people have been marketing to each other, there has been personalization. Fields like healthcare, education, business, and hospitality cater to a wide variety of people who all have different needs, driving the rise of personalization for custom solutions.¹ 

In the late fifteenth century, the German inventor Johannes Gutenberg invented movable type, a system of printing that used individual letters and symbols that could be arranged and rearranged to create pages of text. This allowed for the first trade catalogs to be published, which inadvertently targeted only the wealthy and elite, as only a small percentage of people were literate and able to afford such non-essentials as a catalog. Merchants ensured that their top customers received the catalogs, and personalized the selection of products based on what sold well and what didn’t.² 

Just as marketing has always been around, market fragmentation has existed as long as we've been able to group consumers together. Early on, personalization in marketing was extremely limited because retailers could only tailor their approach based on what they knew about local customers.³ At the turn of the 19th century, however, mass marketing took off when improved transportation systems like advanced railways expanded economies beyond just the local markets. As once-isolated regions became part of a larger economic network, personalization in marketing became nearly impossible. Companies lacked the means to collect individual consumer data, and standardization was the new priority. The result? A shift toward uniform, mass-produced marketing.³

Despite the push for standardization, elements of personalization began to emerge in the late 19th century. One early example was the use of direct mail marketing, like catalogs and personalized letters. In fact, an experiment by Time magazine in the 1940s found that simply addressing recipients by name (“Dear Mr. [Last Name]”) increased response rates by 600% compared to non-personalized letters.⁴ While this approach was effective, it also sparked early concerns about consumer privacy and the amount of data companies had access to—a concern that’s still relevant today!⁵

However, by the 1970s, the novelty had worn off. Response rates declined, and personalization, at least in its early, manual form, became expensive and harder to implement.⁵ To control costs, marketers became more selective about who they targeted. Catalog companies, for instance, analyzed buyer profitability and introduced practices that looked at data points like how many times (or if) a customer had made a purchase in the last year, how much they spent, and how recent their last purchase was. Sometimes this data came directly from consumers, but other times the data came from sources that the consumers hadn’t even interacted with, sparking a lengthy privacy debate.⁵ 

Once the 20th century ushered in an era of mass media with widespread access to newspapers, magazines, radio, and TV, the way businesses reached audiences completely transformed. Although most ads were relatively generic, companies still had to choose which shows, time slots, and broadcasters they would use for their messages.⁵ This was the beginning of a new wave of modern targeting and market segmentation. Today, social media and streaming services have capitalized on the ability of their platforms to target consumers even more directly, making the recommendations, advertisements, and marketing we’re exposed to even more personal.⁶ 

As the capabilities of our smart devices, AI, and other technology have progressed, so has our capacity to personalize the role that technology plays in our lives. With the invention of wearable devices like smartwatches that can monitor vital signs like heart rate, blood pressure, or glucose levels, those in healthcare have access to a plethora of real-time information that they didn’t have before.⁷ Immediate feedback from the various medical devices people use allows patients to track their own needs and respond immediately while they’re out in the world, but also gives providers a better picture of their individual patterns and needs, which can’t be defined by a one-size-fits-all model.⁸ 

Johannes Gutenberg

A German craftsman and inventor of the movable type printing press in the 1400s, which influenced early marketing strategies.²

John Dewey

An American philosopher and psychologist who theorized that connecting personal experiences to education enhances learning. This theory is known as “learning theory,” or “Dewey’s theory.” As Dewey puts it, the “educational process has two sides—one is psychological and the other is sociological.”⁹

Petter Törnberg

A Dutch social science researcher whose work lies at the intersection of AI, platforms, and politics. Törnberg’s research and writing has reflected on the role of echo chambers in modern society and the impact of over-personalization on our newsfeed.¹³

Richard Thaler & Daniel Kahneman

American behavioral economists known for their work on nudging, cognitive biases, and decision-making, which influence how we respond to personalization and help us conceptualize personalized nudges or decision-making strategies. 

The concept of personalization is a broad one, and its impacts are equally extensive, transforming everything from education to healthcare to consumer experiences. In education, personalized learning keeps students engaged by aligning lessons with their interests, improving motivation and outcomes. Similarly, wearable health devices provide real-time, customized insights for better diagnostics and preventive care, while AI-powered algorithms enhance consumer experiences by offering relevant recommendations, streamlining shopping, and increasing engagement.

Sparking Personal Interest in Education 

One of the most exciting applications for personalization is as an approach to improving education. If you’re a parent (or if you’ve ever been a child), you know that every child is unique and has individual needs. This seems to be even more true as an increasing number of children are diagnosed with learning differences like ADHD and autism.10 Personalized education seeks not only to support students who might be struggling in a traditional classroom environment, but also to increase interest in school subjects by using a student’s individual interests, values, and preferences.⁹ Although there are definite drawbacks to over-individualizing the education system, there are also some major benefits that should be noted. 

The idea of connecting a child’s experiences outside of school with learning materials dates back to the work of psychologist John Dewey. His theory suggested that while increasing classroom interest can be achieved through methods like teacher enthusiasm, the ideal scenario involves tailoring learning materials to each student's personal interests.⁹

A recent study on personalized education identified three main kinds of interventions, and in each category, the student’s interest is key. Capitalizing on students’ interests is essential; when they’re interested, they’re more engaged in the learning process, they stay persistent when things get tough, and they actually learn more.⁹ Teachers can harness a student’s interest by customizing the content of what they’re teaching to appeal to an individual student by simply including personal details or adding content that the students are excited about. 

Compared to generic materials, content and context personalization are highly effective in increasing student interest in the topics and can encourage them to create personalized connections with the material. The second method of personalization involves choice; for example, students can be given the option of multiple types of worksheets, all covering the same topic.⁹ Although all the worksheets might focus on multiplication skills, allowing students to choose the kind of worksheet they prefer can allow each learner to feel more empowered and engaged, even if all the students are working on the same skill.

Lastly, the study found that when students can form a connection between learning materials and their interests or future aspirations, they not only become more engaged in the short term but also increase their interest and performance in the long term. Particularly for those who feel disconnected from the topics at hand, an intervention like asking students to write an essay about the potential value of what they’re learning in their life or future career can actually keep them motivated.⁹ 

Personalized Algorithms 

So much of what we often hear about personalized advertising and the power of algorithms focuses on the negatives of data tracking: privacy concerns, ethical issues, and more. But in our modern world of targeted advertising, it can be easy to overlook some of the benefits that come with machine learning. Maybe you’re someone who does recognize or even appreciates when an AI algorithm provides you with tailored ads; maybe you’ve been scrolling through social media, only to serendipitously encounter a coupon code in your feed that is for the exact pair of shoes you’ve been looking for, in your size and the color of your dreams. Maybe you’ve been helped out by a pop-up letting you know when that flight to a tropical getaway is on a super sale. It might be eerily obvious that your devices have been tracking your conversations and search history about that dream vacation, but it sure is helpful in the moment.  

In fact, research shows that algorithm personalization, specifically with AI-powered tracking, can improve consumers' pleasure and level of engagement with the brand.¹¹ Happy customers make for happy companies, and because these AI algorithms can increase customer loyalty through personalized product recommendations and improved customer engagement opportunities, the companies in turn benefit from boosted revenue. Personalized marketing or shopping technology can even make the entire shopping process smoother; when the system can anticipate what you’re looking for, where you might need help, and what you want before you even realize you want it, then the entire shopping process becomes more seamless and enjoyable. 

Improving Healthcare through Wearable Devices

Wearable and AI-powered biofeedback devices have revolutionized healthcare and biomedical monitoring by enabling continuous, real-time measurement of critical biomarkers. As the global elderly population grows, the demand for personalized, point-of-care diagnostics has increased, which has shifted the medical industry toward more tailored health monitoring. Think of all the wearable biosensor accessories we have now: watches, rings, hats, chest straps… the list goes on! In addition to the biosensors that are integrated into our clothing, we also now have many implantable devices, driven by advancements in electronics, nanomaterials, and biocompatible materials. The rapidly advancing technology has allowed us to gather more and higher quality physiological data, improving diagnostics, treatment efficacy, and overall patient care.^7,8

The more opportunities we have to personalize these wearable biosensors, the more effective they’ll be, as many health factors vary based on genetics, lifestyle, and medical history. For example, a normal blood pressure or heart rate for a young marathon runner may be incredibly low for someone else, and could indicate a problem. Devices like your smartwatch and more elaborate wearable ECG monitors often have these integrated personalized features, which can give users insights about their well-being based on their own unique health data.^7,8 

However, the devices we use are only as helpful as they are calibrated, and we need to ensure that the devices we rely on for information, as important as our health, are accurate. The reliability of wearable devices is vital in detecting early symptoms and managing chronic conditions. As these technologies continue to evolve, wearable devices will likely become an integral part of preventive medicine, allowing individuals to monitor their health proactively and reducing the burden on traditional healthcare systems. Ultimately, this shift toward personalization will hopefully lead to better patient outcomes, lower healthcare costs, and a more data-driven approach to medical care.^7,8

Although personalization can have a positive impact in applications to marketing, education, or healthcare, there are limitations to every instance of its use. For example, when educators spend too much time and energy crafting individualized learning plans, other students may fall through the cracks. AI algorithms also have the power to be creepy or annoying, or even put us in danger of manipulation. 

Personalization in Education 

In a recent conversation I had with some friends who are school counselors, I was shocked to learn just how many K-12 students have Individualized Education Programs (IEPs). These personalized plans are designed for individual students with special education needs, whether it’s due to a speech or language difference like dyslexia, a disability or chronic health problem, or autism spectrum disorder.¹⁰ Regardless of the reason that the student or the parent feels that there is a need for an IEP, the rise in students with their own IEP has created its own challenge for students and educators. In the last 50 years, the number of students with an IEP has doubled, but the shortage of special education educators has only gotten worse.¹⁰ While my friends explained the impossibility of giving every student the kind of special attention their plan outlined while in a classroom of ten other students with special needs (and the rest of the classroom with students who might not have IEPs but certainly crave personalized care and engagement), I wondered how we could ever find the right balance of catering to the needs of each individual child to keep them engaged without overemphasizing the need for personal learning plans, which may even take away from the experience of primary school, where many students simply learn to coexist with those are different. 

Perhaps one of the most salient examples of an attempt to personalize education is the myth of learning styles. Usually, the concept of learning styles posits that people have one specific style in which they learn best: visual, auditory, kinesthetic, or through reading and writing. However, the idea that everyone is born with a pre-defined and static learning style has been debunked; instead, research shows that people actually learn best when they’re exposed to many teaching approaches together (such as words and pictures).¹²  Combining methods tends to improve learning across the board, a phenomenon known as the multimedia effect.¹² 

Unfortunately, the myth of learning styles has persisted: up to 90% of people surveyed in Western industrialized countries, including educators, think that learning styles not only exist but are genetic, unchanging perspectives. Not only does this view lack any substantial evidence to back it up, but it completely undermines the role of the environment and socialization. Research even shows that the learning style model can undermine education, because as educators spend time and money trying to personalize lessons to fit certain learning styles for each student, the students miss out on exposure to a variety of teaching methods, which is ultimately what helps them the most.¹² 

Echo Chamber

Social media platforms use personalization algorithms to tailor content to individual users, and while those do offer some benefits like providing more relevant information, personalized algorithms also pose risks related to privacy, autonomy, and limited diversity of information. Although the ramifications of unethical AI algorithms are wide, it may be helpful to focus on just one of the ways in which these personalized algorithms can negatively impact us on an individual and societal level. 

For context, algorithms begin by first filtering and prioritizing content for each user based on their demographics, online habits, preferences, activities of friends, and other factors that even many developers aren’t fully aware of. Unfortunately, this is one way people find themselves in an online echo chamber, where they only encounter information or opinions that reinforce and reflect their own. This phenomenon is tied up with confirmation bias, which is the tendency to favor information that reinforces one’s existing beliefs. Because personalized AI algorithms continue to serve users the same types of information and the same or similar sources, these echo chambers can cause people’s views to become incredibly distorted, with their worldview no longer resembling reality. Researchers like Petter Törnberg have shown that it can become harder for people to consider opposing viewpoints or discuss complicated topics, especially when they lack any exposure to outside opinions.¹³ 

Although the appeal of a personalized feed where users don’t have to encounter any posts that they disagree with is high, echo chambers can create a dangerous breeding ground for misinformation to spread. Particularly with the risk of political interference, if personal data gets into the wrong hands, then fake news can target those who may be most likely to believe it (such as Russia’s interference in the 2016 and 2024 elections).¹⁴

Personalization Pitfalls in Marketing 

Besides the dangers of polarization and privacy concerns, there are also some potentially annoying and creepy ramifications of personalization in marketing. In terms of annoyance, people may generally feel positive towards their personalized feeds but become annoyed when one of several things occurs. This could be due to irrelevant recommendations, such as after you’ve purchased a car, you might be inundated with car ads, which makes very little sense. Clearly, you’ve already bought the car, and now is probably the time you’re least likely to be in the market for another vehicle. The recommendations might also be insensitive or invasive. For example, I’ve noticed a male coworker’s ad space displaying ads for hair loss and erectile dysfunction. Regardless of whether these targeted ads are due to their demographic or search history, they may not want to see these unsolicited sites being recommended to them.

Another pitfall is when consumers get the eerie feeling that accompanies the realization of just how personal your feed is. This response can come in the form of reactance when consumers get angry that they’ve seen the same ad too many times or are excessively retargeted. The creepiness ditch may also show up with stereotyping. For example, one study found that when larger-bodied consumers received ads for a weight loss program, they felt “unfairly judged” by the matched message.¹⁵

Overly personalized advertising can make you extra wary of the potential for privacy violations. One time, when I was checking into my room at a popular hotel chain where I stayed frequently for work, I noticed the TV screen displayed an image of the outline of a young woman who shared an uncanny resemblance with me, running along the city’s bridge at sunrise where I did my own run every morning before work. Though it may have been a strange coincidence, it’s hard to settle into your hotel room with the eerie feeling of being watched. When messaging becomes too personalized, the marketing strategy can actually backfire and make consumers feel violated. 

Duolingo’s Personalized Language Learning Model 

Have you tried to learn or practice a new language recently? Even if you haven’t, you’re probably familiar with the friendly green owl from Duolingo, the immensely popular language-learning app. Their platform uses AI to create personalized education experiences for people based not only on the language they’re practicing, but also all of their strengths and weaknesses within the language.¹⁶ Like many AI algorithms, the app analyzes how the users interact with the app and adjusts the lessons in real time to make the experience more engaging (and hopefully make learning even more fun and efficient). 

The AI algorithms help the Duolingo team analyze user responses in order to identify common grammar mistakes and make corrections and suggestions. Although those at Duolingo review the data as a whole, to identify trends and points for improvement across the entire user population, the integration of AI allows for these adjustments to be made on a personal level too. In fact, the app can even analyze the sounds and patterns of your speech to provide targeted feedback on your pronunciation.¹⁶ 

Research has shown that language learning is one of the areas that has benefited the most from personalized learning systems, fueled by advancements in AI and big data. The increasing amount of data available and the power of the algorithms we’ve developed have led to even more advances in the field, providing more interactive and effective educational processes, as these predictive and personalized learning systems can adjust based on current performance and proactively adapt to future needs. Research has continued to show that this kind of personalization supports long-term student development and helps them maintain their motivation and progress. As these systems continue to improve, they’ll also continue to democratize education by making learning more personalized and accessible.

Spotify Wrapped’s Success

To understand the behavioral science behind Spotify Wrapped's success, we must first appreciate how well the company has capitalized on its ability to personalize the end-of-year summary it gives to its listeners. Sometime in December, Spotify sends each user an aesthetically pleasing and engaging summary detailing all of their biggest “hits” of the year: stats like how many songs they listened to, their top artists, even what their “vibe” was in each season. This feature became wildly popular, likely because each person’s annual wrap-up was unique and truly personal. Our music tastes are an expression of our individuality, so having an easy way to share our musical adventures feels like a great representation of who we are (or at least part of who we are!). 

In fact, Spotify’s Wrapped experience has become so popular that it inspires more and more copycats every year, not just with other music streaming services, but for everything from fitness apps to food-delivery services (not that I want to see how many times in the past year I ordered pad thai). While some of these personalized summaries may be more successful than others, the concept of personalization will likely remain successful across domains, because who doesn’t love to see something created for them, about them? 

This is Personal: The Do's and Don'ts of Personalization in Tech 

Because personalization is one of the biggest trends in tech, it’s important to understand where its major strengths and weaknesses lie. This article unpacks some of the things that personalization gets right when done well, and some of the areas where it’s more likely to flop. 

Behavioral Targeting 

Behavioral targeting is a digital marketing strategy that analyzes users’ online activities—such as website visits, clicks, and search history—to deliver personalized ads and content tailored to their interests. This article explains how, by leveraging this personal data, behavioral targeting can enhance user engagement and increase the likelihood of conversions by showing individuals content that aligns with their preferences and behaviors.

1. Rife, N. and Yan, R. (2020) ‘The Evolution of Personalization in Marketing’, The Jabian Journal. Available at: https://journal.jabian.com/wp-content/uploads/2020/09/JJ-Fall-2020-14-evolution-of-personalization-in-marketing.pdf 
2. Ross, N. (1992) ‘A History of Direct Marketing’, Direct Marketing Association.
3. Tedlow, R.S. (1996) New and improved: The story of mass marketing in america. Boston, MA: Harvard Business School Press. 
4. Reed, O. (1949) ‘Some random thoughts. on personalizing’, The Reporter of Direct Mail Advertising.
5. Petrison, L.A., Blattberg, R.C. and Wang, P. (1997) ‘Database marketing: Past, Present, and Future’, Journal of Direct Marketing, 11(4), p. 2. doi:10.1002/dir.4000070306.
6. Paz, E. (2023, November 15). The history and evolution of personalised advertising. Medium. https://medium.com/@emilianopazf/the-history-and-evolution-of-personalised-advertising-790e3ec625fe
7. Guk, K., Han, G., Lim, J., Jeong, K., Kang, T., Lim, E. K., & Jung, J. (2019). Evolution of Wearable Devices with Real-Time Disease Monitoring for Personalized Healthcare. Nanomaterials (Basel, Switzerland), 9(6), 813. https://doi.org/10.3390/nano9060813 
8. Johnson, K. B., Wei, W. Q., Weeraratne, D., Frisse, M. E., Misulis, K., Rhee, K., Zhao, J., & Snowdon, J. L. (2021). Precision Medicine, AI, and the Future of Personalized Health Care. Clinical and translational science, 14(1), 86–93. https://doi.org/10.1111/cts.12884 
9. Gunawardena, M., Bishop, P., & Aviruppola, K. (2024). Personalized learning: The simple, the complicated, the complex and the chaotic. Teaching and Teacher Education, 139, 104429. https://doi.org/10.1016/j.tate.2023.104429 
10. Blad, E. (2023, July 31). The number of students in special education has doubled in the past 45 years. Education Week. https://www.edweek.org/teaching-learning/the-number-of-students-in-special-education-has-doubled-in-the-past-45-years/2023/07
11.  Misra, Ruhi & Kapoor, Shikha & M a, Sanjeev. (2024). The Impact of Personalisation Algorithms on Consumer Engagement and Purchase Behaviour in AI-Enhanced Virtual Shopping Assistants. 10.21203/rs.3.rs-3970797/v1.  
12. Nancekivell, S. (2019, May 13). Learning styles myth undermines education, report shows. American Psychological Association. https://www.apa.org/news/press/releases/2019/05/learning-styles-myth
13. Törnberg P. (2018). Echo chambers and viral misinformation: Modeling fake news as complex contagion. PloS one, 13(9), e0203958. https://doi.org/10.1371/journal.pone.0203958 
14. Shuster, S. (2024, November 2). Russia ramps up 2024 election disinfo ops. TIME. https://time.com/7171326/russia-2024-election-disinfo-ops/ 
15. Kotamarthi, P. (2021, June 14). This is personal: The do's and don'ts of personalization in tech. The Decision Lab. https://thedecisionlab.com/insights/technology/this-is-personal-the-dos-and-donts-of-personalization-in-tech
16. Wodzak, S. (2023, March 30). How AI is improving education on Duolingo. Duolingo Blog. https://blog.duolingo.com/ai-improves-education/ 
17. Xia, Y., Shin, S.-Y., & Shin, K.-S. (2024). Designing Personalized Learning Paths for Foreign Language Acquisition Using Big Data: Theoretical and Empirical Analysis. Applied Sciences, 14(20), 9506. https://doi.org/10.3390/app14209506