Think Like a Futurist: How to personalize your product
There has never been a better time to be a consumer. According to futurist Kevin Kelley, “Every 12 months we produce 8 million new songs, 2 million new books, 16,000 new films, 30 billion blog posts, 182 billion tweets, 400,000 new products.” This results in an infinite sea of things to learn, ways to be entertained, and products to try — all clamoring for your attention.
As consumers, how will we adapt to this deluge of choices? One thing’s for sure: we will inevitably turn to products that help us filter and personalize the content we consume. Products that help us triage our options, and ultimately choose the things it knows we’ll love the most.
Soon enough, every song that pops up on your Discover Weekly, every new gadget that’s suggested for you on Amazon, and every film you see on your Netflix homepage will be the crème de la crème — consisting of new discoveries related to the things you currently enjoy and a sprinkle of a few surprising “wildcard” choices to expand your horizon of preferences.
The three components of a perfect filter
The “more like this” filter
Many consumer products already employ one aspect of personalization. Think of Amazon, Netflix, or LinkedIn serving up more of what you already like. This is called a recommendation engine. I personally find these recommendations incredibly reliable, sometimes even better than recommendations from experts or friends. This may be because they’re driven by algorithms and based on cold, hard data.
Personalized recommendations are so valuable that Netflix has 300 people working on its recommendation system with a budget of $150 million, and “more like this” offers are responsible for a third of Amazon sales — a difference amounting to about $30 billion in 2014 (Sources: Gigaom, McKinsey).
The “people who liked this, also enjoyed this” filter
In addition to filtering based on our own current preferences, we also turn to our peers to recommend new things. Think of Facebook and Twitter, where your feed is composed of direct posts from pages and people you follow, but you can also see other, third-party posts that your friends have engaged with. This allows you to learn about things your friends like, that you may have not known about.
Similarly, this is how Spotify makes music magic with Discover Weekly, their personalized weekly playlist. In addition to utilizing the information from a listener’s unique “taste profile,” they also leverage other people’s playlists to find new music selections. For example, if Spotify notices that two of your favorite songs tend to appear on playlists along with a third song you haven’t heard before, it will add that new song to your Discover Weekly. This is known as collaborative filtering.
The wildcard filter
That being said, a world where you only consume what you and your peers enjoy would be a pretty boring, homogenous world. And a scary one at that — we’d be discouraging open mindedness and eliminating new perspectives from forming.
The danger of being rewarded with only what you already like is called “overfitting,” which is seen often in the political realm: Members of one political party who depend on a simple filter of “more like this” will rarely consume content of another political view, which hardens their existing beliefs.
As technologists, it’s our duty to prevent overfitting by introducing people to new things. How might products suggest stuff that I don’t like, but would like to like? Or help you discover more of those things I used to hate, but have learned to love?
This is where we’re at with technology. While we’re experimenting with deep learning and neural nets, we’re still working on creating this third pillar of the perfect filter — the one that introduces those divergent recommendations that can expand our horizons as consumers.
How to personalize your own product
So, how can you employ these components of personalization into your own product? Here are some tips and questions to start thinking about. And, I’m sure this comes as no surprise, but it all boils down to data:
Start collecting more nuanced data about your customers.
In addition to the information provided by Google Analytics, start creating dashboards that tell you more about your customer’s journey, habits, and demographics. Solicit information through fun onboarding quizzes, or through more in-depth user surveys. Having a bank of information about your users makes it much easier to begin using that data to create a personalized user experience.
Gather data about similar customers.
In order to execute the second level of personalization (“I want to know what my friends like”), you’ll need to figure out the best way to group your users. If you’re at an interior design startup, it could be by home decor style or project budget. If you’re at a B2B software company, perhaps it’s job title or decision-making power. From there, you can gather data about other people in that cohort, and use it tailor that persona’s experience.
Think beyond collaborative filtering.
How can we introduce novel or even drastically different ideas and experiences to our users? While we’re waiting for the development of large collaborative databases, we can employ creative, scrappy solutions in the interim. Two ideas I’ve been thinking about include:
- One-on-one live chats: In a world of bots, human connection is scarce and valued more than ever. Opening up a chat service that allows users to connect with an actual human allows you to collect nuanced data, offer more unique recommendations, and gather information about how you should continue personalizing your product.
- Utilizing adjacent clues: Is there other, lesser-known information that can help personalize your product? Weather and location based information could influence a ride-sharing experience, or an eCommerce recommendation algorithm. Gaining access to a consumer inbox could help a custom-suit startup personalize their communications or influence their product roadmap based on event invitations. There may be valuable information hiding in corners you haven’t explored yet!
Overall, we’re headed in an exciting direction for consumerism—one that soon will become expertly filtered and personalized. But how can we ensure this personalization is diverse, suggesting novel ideas in addition to existing preferences? I’d love to hear your thoughts in the comments below!
This was first published on Katerina’s bi-weekly mailing list, where she shares what she’s learning about product, technology, and productivity.