coates’s Twitter Archive—№ 19,182

Twist: analytics processor doesn't do any actual measurement and simply uniformly confirms that mail list popovers are always bad for users.