NOTE: This post originally appeared in Techcrunch. (Photo credits)
Ironically, Rao and her pundits are missing the zooming race car that’s speeding by them as they wait for the personalization bus to arrive. That racecar is Pinterest and the new breed of startups marking the beginning of what I call the “Curated Web.”
The promise of personalized eCommerce began over 10 years ago with technology pioneered at Amazon. It was then that the mental dye was cast for what eCommerce personalization would look like, an algorithmic solution for matching customer to products. Web watchers came to expect that someday all online retailers would have such algorithms on their sites and the dream of personalized commerce would finally be realized.
For over a decade, startups took their best shot at making this apparition a reality. Companies like Hunch tackled the data collection piece of the equation, asking users endless survey questions to determine their tastes and preferences. Google’s Boutiques.com tried to crack the challenges of structuring the data associated with personalized shopping recommendations. Ultimately, these attempts failed.
In September, Google shut down Boutiques.com and the founders of Hunch sold their company to eBay, an outcome far short of their IPO hopes. Previous attempts at personalized eCommerce were unsuccessful because those involved failed to realize they were missing one key element, the interface. Curated Web companies are defined by their ability to use new interfaces to collect and structure data better than previous algorithmic solutions.
Users Want It All
While the tech world waits patiently, expecting the solution to personalized eCommerce to look like Amazon ported to other online retailers, “Curated Web” companies like Pinterest are changing the game by changing the interface. Pinterest will be the first company to nail eCommerce personalization because they understand the importance of having an interface, which matches what the user is there to do, discover stuff they like from across the web. Pinterest has cracked personalization right under everyone’s nose by doing the two things Rao says have yet-to-be invented, data collection and data structuring.
Collecting Data, One Pin At The Time
Pinterest is becoming the web’s personalized mail-order catalog. Each user is presented with a one-of-a-kind visual interface based on their tastes. They are presented with any product, from any retailer, anywhere in the world. The items they see are curated through people and topics they’ve identified as interesting and what is shown to them improves the more they interact with it. Every time they pin, re-pin, like, or comment on an object, the relevancy of the products displayed on their magic catalog improves. This is what personalization looks like, effortless, simple, social and fun. It’s the interface, stupid.
“Big deal?” you say. “Facebook can do this!” No, they can’t. Social media is for selectively sharing information about you and Facebook is built for presenting yourself in the way you want to be seen by others. Facebook is, by design, about creating a network of relationships and sharing with them selectively. This brings up all kinds of privacy concerns, which reduce the flow of content creation and sharing to limited circles (Yes, I said “circles.” Don’t even try it, Google).
Pinterest has no such restrictions. Pins are inherently open and there is no expectation that anything shared is private. It’s a community built on individuals acting in their own self-interest to capture and collect things that interest them. Facebook and Google+ are just not the right interfaces for capturing and collecting products, and attempting to discover new products amid the newsfeed clutter is hopeless.
Making Sense Of The Data
To some, the rise of the Facebook “like graph” foretold a personalized future, where retailers could utilize data collected from what users “like” to serve targeted recommendations. Here, Rao is spot on about why Facebook fails to provide useful data on users’ tastes; it lacks structure.
What does it mean if I “like” something on Facebook? Not much. “Liking” a brand or even a specific product, doesn’t provide useful information. If I “like” Babies’R’Us, does that mean I like the brand, a specific outlet, a product I found there, or am I just hoping to get a coupon? From a personalization perspective, it’s pretty low-value stuff. There is no structure to correlate my actual likes with my Facebook “likes”. However, Pinterest solves the data structure problem brilliantly. You’d think they’d need some fancy photo recognition technology to tag a handbag by color, make, and model, but I doubt they have any such technology. Pinterest doesn’t even try to solve the data structure problem because they don’t have to.
Again, it’s the interface stupid. By presenting users with a dynamic catalog, and tasking users with the job of deciding whose tastes they’d like to follow, structuring image data becomes irrelevant. Pinterest simply has to make sure the magic catalog appears, tailored to each user’s stated preferences. It’s doesn’t matter if Pinterest knows a thing about what’s in each image on a pin board; what matters is that it’s curated by the user and the user likes what they see. If they like the products, they’ll buy them, and Pinterest laughs all the way to the bank.
So while the rest of the web is waiting for the personalization bus to arrive, Pinterest, and perhaps a few other fast-moving Curated Web companies, will be far ahead. It’s clear, given Pinterest’s astounding growth that personalized eCommerce is here to stay, even if it looks nothing like what you imagined.
Note: I have no affiliation or investment in any company mentioned in this post.
Related Articles
- Schedule Maker: a Google Sheet to Plan Your Week
- Cancel the New York Times? Good Luck Battling “Dark Patterns”
- How to Start a Career in Behavioral Design
- A Free Course on User Behavior
- User Investment: Make Your Users Do the Work
- Variable Rewards: Want To Hook Users? Drive Them Crazy
- The Hooked Model: How to Manufacture Desire in 4 Steps
