The hyper personalization paradox

“I quit Netflix this month”. “I uninstalled the Facebook app from my phone (using the web app infrequently)”. “I’m almost on the verge of shifting to Safari from Chrome.” “I might never subscribe to YouTube Premium.”

While companies and marketers focus on cracking the science of personalizing experiences for their end users, they might be overlooking the basic and latent nature of human psychology to anticipate 'serendipity' in everything they engage in.

The problem with hyper personalization

Amidst the Covid-19 situation, most of our browsing history has been around "death toll", "latest news of Covid-19", "symptoms", "vaccination" or similar. Although contextual, this has led to a unique problem for me.

My Facebook video feed is now swamped with news about the impending disasters of Covid-19. This was impacting my wellbeing so much that I had to move to a state where I wasn't consuming any news around Covid-19, at least not anymore on the Facebook video feed.

And nearly everyone I know has been thrust into varying degrees negativity with similar news, content, and social media platforms.  

Now was that a great experience offered by these platforms that just tried to 'personalize' the content feed for me based on my (or my social network's) search history?

Technically the algorithm did its job, but can we say that the experience was human-centric? Probably not.

This experience got me thinking that the phenomena applies to so many other areas of our lives—the content we consume, e-commerce websites we visit, fitness and wellness apps we use, recommendation websites we depend on and many other digital experiences. Everything has been personalized at the cost of nothing being offered as a “new", or what I call, "serendipitous" experience.

Speaking of serendipity, the human mind is a seeker, always lurking to discover serendipity (a happy coincidence) in everything we do, the people we meet, and the experiences we engage in.

Quick detour—interestingly, the word 'serendipity' was coined by Horace Walpole, a member of the British House of Commons in the 18^th century. His unique claim and talent for finding what he needed just when he needed it prompted him to coin the term. 

Coming back to how that is relevant, here's a binary logical reasoning question that appreciates the problem of debunking the importance of serendipity around human experiences:

Fact 1: Mary likes to watch sitcoms on Netflix. True

Fact 2: Mary was looking for "Top 10 romcoms on IMDB" on Google. True

Now, does that mean, Mary doesn't like thrillers is TRUE as well as, Mary doesn't like Documentaries is TRUE.

This is the point where the intelligence oversees or “un-detects” this latent opportunity completely. How is that a problem, you ask?

The app ends up suggesting only sitcoms and romcoms to Mary until one day when she realizes she has overdone it and moves to another platform in search of "better content".

A lot of us may relate to this problem, especially now, with a never-before dependency on digital content consumption apps, most of us already feel that the majority of the digital content platforms are “out of good content".

This feeling could mean one of two things—either you have watched it all (which is highly unlikely), or the recommendation engine is not doing a good job with pushing results under 'Suggestions for You'.

Could “un-personalization” be the new trend?

I asked three simple questions to a few people in my network about their experience with personalization and it was quite interesting to hear what they said:

Q1: What is wrong with a hyper-personalized experience?

"I only see what I had once searched for. There is nothing new that the app suggests"

"The app feels like a black hole—The more I watch a certain kind of content, the more I get suggestions for similar content."

"My feed is so corrupted now, I want to cleanse and reset it"

"Stop personalizing everything for me. I fear I'll never chance upon anything new"

Q2: Is the concept of "breakpoint" for real? If so, what does it mean in the context of a customer using your product or consuming your services?

Yes, “breakpoint” is for real and has been traditionally referred to as the churn stage. In literal sense, it is the point at which a user is thinking of exploring other alternatives because they are no longer happy with the experience your platform offers.

Q3: How to determine these breakpoints?

That's easy, and most applications already have a system in place to detect when a user has declining consumption rate, there is a frequency drop in visits to the platform, or there are high abandonment rates.

The part that is challenging, yet valuable, is to identify what you suggest to the user when they are at their “breakpoints” or even before they get there. With the maturity of intelligent technologies today, there isn't one right answer to this problem. Some of the ideas or questions that could start solving for this are:

• Can we build and add a unique layer of serendipitous intelligence to the Data Management Platforms?
• Can these layers have the ability to identify insights around Network 360° data (and not just Customer 360°) from clusters of information around the users—people they engage with, the content their networks view, the kind of content that may be contextual around a certain time in their life (or globally, like the Covid-19 situation currently), the "next best " pattern recognition, etcetera?
• Can the system rapidly build a way to take contextual feedback from users? Surprisingly, none of the content consumption apps have a feedback system in place today.

Although leisure content consumption could be just one-use case, the problem exacerbates when it comes to disturbing content, such as posts dealing with self-harm, news, or fake news. The corroboration factor is so high in case of such hyper-personalized content that users end up slumping into isolated zones that can completely detach them from the logical ability of looking at both sides of things.

Can experience-moderation with a layer of "serendipity" be the solution?

Using a sophisticated Artificial Narrow Intelligence (ANI) program, it might be possible to train machine learning classifiers to embed a layer of moderation for that extra attention to identify if the user might be at the break point of their customer lifecycle.

In the short term, this might mean a setback to your user base and lead time with validation of your recommendations (considering, serendipity comes with a balanced risk) by the users but in the long term, these classifiers can invigorate the hyper personalization play, completely.

These serendipitous layers of intelligence could pre-empt the human behavioral economics and offer experiences that are not hyper-personalized but hyper-consumerized. This is where products will truly humanize experiences for end users. It may not be the complete recipe to design a solution that addresses the hyper personalization problem; however, this might be a good starting point. And there will be critical considerations around data security and privacy along with transparency of what and how much is being shared.

In summary

The distressing news all around has been taking a toll on our mental health, hinting at a global concern of looming mental health crisis. The problem is magnified since people around the world being forced to stay apart from their friends and loved ones.

During my research, by the time I got off the phone with my last respondent, I was worried more about this distress than how my weekend was going to look like. May be, that was the real hook.

What are your thoughts around hyper-personalized experiences on content, ecommerce, social, search platforms, and the likes? Have you experienced anything similar?

Do you also think it’s time companies revisit personalizing everything for their users?