Identity is missing from the internet
Identity is missing from the internet.
Today, as you move through the internet, there isn’t currently an understanding of your knowledge state or your taste.
This was by design (or lack of it):
Shopify Tobi likes to say that the web was really great, but it had two critical design flaws. The first design flaw was that payments weren’t built in, even though they almost were. And that is kind of what allowed Amazon to grow up and win this whole slice of commerce. And the second flaw was that identity was not baked into the web. It was up to websites to decide how they managed identity and how they attract people’s identity. And that ultimately led to things like… it led to AOL having this initially last presence around owning people’s identity.
But then down the road, this is how you get to places like Facebook who provide this incredible service for people, by giving them an identity that they can use to talk to other identities. But ultimately (that) turned a lot of the internet and subsequently a lot of the world of commerce into something that happened inside walled gardens” ~ Alex Danco, speaking on Jim O'Shaughnessy’s Infinite Loops podcast.
Mechanisms we have for overcoming this today fall broadly into two camps:
1) Intent-driven Google search (a composite action made up of an a priori knowledge / previous time spent on reading something)
2) Ad platforms collecting expressions of your tastes through likes (not what you have bought) and renting this information out to companies to then find you and people that have made similar expressions via lookalike audiences.
The first relies on the user being able to answer the right questions.
The second relies on the user’s consistent engagement with aspirational content, plus significant ad spend by brands.
Neither are satisfactory.
What’s next?
Currently, we consume a firehose of products / content, and the filtering sits with the user. This needs to be flipped such that identity sits with the user and then filtering happens on the internet.
First, we need the ability to match the attributes of the user to phenomena on the internet (existing content, existing products) in a much richer way. We need to take advantage of the fact that human recall uses more sophisticated attributes than brand names, products or file names.
Then, we need to be able to account for the fact that our knowledge state / taste changes. Currently, recommendations are laggy. They give me what I did like, rather than move with what I am now liking. They don’t account for growth / shifts in perspective, but instead create personal echo chambers.
If you’re working on a product in either of these areas, I’d love to chat!