> Why do websites slow themselves down with trackers, etc.? It pays to do so.
Unpopular opinion coming from a Director in a media agency who has these “come to Jesus” conversations daily with stakeholders trying to gum up the works with tag managers full of mystery meat, purposely render-blocking monolithic trackers, and A/B testing scripts:
They simply don’t have technologists in the room explaining the trade offs between gathering mass amounts of data for analysis and actual site performance. (Consider also much of the data is for data’s sake, and is often inconsistent, contradictory, and/or redundant due to being JavaScript-based and complications such as ITP, Tracking Protection, ad-blocking, etc.) You have mounds of studies from Internet pillars such as Amazon, Akamai, and Nielsen Norman Group laying out how crucially performance equates to revenue by meeting visitor expectations of _less_ friction, but marketers get sold a completely different narrative by vendors and influencers in their own segment of industry in order to deliver “numbers”.
Classic mismatch of concerns and nobody to bridge them.
So they package up whatever giant bundle of crap they've been sold as "necessary" into a tag manager, and instantly (and negatively) impact the site with it. Whether that's on the initial render being delayed, or the phalanx of async scripts all firing when the poor visitor's just trying to scroll on their phone, it all flies in the face of conventional wisdom.
Coincidentally, this is also related to how marketers are sold into AMP. The sales pitch from GOOG is usually less concerned with addressing the actual problem, instead suggesting the marketer bypass it by using AMP as a solution.
A year ago I started synthetic and (responsible) real-user performance monitoring on our sites, which breaks down third-party impacts, to bring data to the table. That has helped substantially with these conversations.