Thread · 5 tweets · 6 Jul 2020

I think that single ID are exactly the wrong solution to this problem. Assuming the privacy issues were solved (which seems impossible on first principles, but let’s assume!) IMHO it still fails to address the competition issue.

Data is non-rivalrous, multiple parties can observe it at the same time without detracting from each other. So perfect cookie syncing takes nothing away from Google, though it does add to others.

However, the value of data isn’t additive. Individual datum don’t have value, it’s the *links* between them that do. “3” is without value, but tied to me you can use it as my birth month. The value of tracking a user on sites A&B is more than the sum of behaviour on both sites.

This effectively translates to network effects in the value of data. The entity with the most data, the most links, can derive more value from the same data that another party also has, because it also has links to other data.

So a universal ID still structurally privileges the biggest party. It also furthers intermediary capture, which means less accountability. I think the trick isn’t to spread a non-rivalrous commodity around more; it’s to make it rivalrous. Prevent all sharing across contexts.