Neither one is a great way to measure market share. Using tracking pixels is notoriously bad though, because it tends to lead to an echo-chamber like effect. No

Neither one is a great way to measure market share. Using tracking pixels is notoriously bad though, because it tends to lead to an echo-chamber like effect. Not everyone with a device will visit a site with the specific browser tracking code, and many people will visit those sites many times from different IP addresses, and even different browsers installed on their device. There are many ways tracking code can lead to false statistics. Units shipped doesn't tell the whole story either, but it is a specific metric that doesn't suffer from bots gaming things or any of the other problems that using stats based on tracking code can lead to.