Below is a video by a fan of a Twitch streamer by the name of Asmongold. He made this video back in October, when Asmongold reacted live to another video by another streamer calling himself Preach. Both are World of Warcraft veterans, and both are experienced in high-end endgame play. What they're talking about here are decisions with regard to game design and how the current expansion (as of this post, Battle For Azeroth) has a lot of its flaws due to decisions made over a decade ago. This is not a short video, but worth it if you're into game design and game theory, be it at the macro level or (as they are) the micro level.
Let's break down the core of Preach's position: Wrath of the Lich King is where Player Retention trumped all other concerns regarding game design, but the full expression of this priority didn't manifest until two expansions later with Mists of Pandaria and it's at that point where the cancer metastasized and the freefall began in earnest.
Player Retention is the measure by which an online game holds on to the players it has, and decisions motivated by this concern inevitably involve making the game easier in some manner or another. During Wrath, the developers made design decisions with the explicit intention of giving high-end (and thus high-powered) players incentives to go into easier content and help the suck players actually finish that content. This carried over, notoriously, into Cataclysm after their initial removal when those same suck players turned out to not have improved at all; they let the good players baby them through Wrath instead of get good themselves.
It was this same concern that, back in The Burning Crusade, had the developers changing raiding into 10-Man and 25-Man tiers; the latter got the best loot at the time. This changed down the road, removing the existing incentive to deal with the far greater logistical and managerial issues of 25-Man raiding, and as such that scene died faster than vapid sorority girls in a slasher flick. Mists introduced RNG-based procs to upgrade items that dropped in raids, "Titanforging", at the time to deal with this.
It worked, big time. Why? Because it uses Gambler Psychology, exploiting the same flaws in human cognition that casinos use to rope in people and bilk them of their cash.
Having seen it work so well, it carried forth and expanded with Warlords of Draenor and did so since. The introduction of Mythic Plus (timed speed runs, with higher quality loot as the reward) in Legion iterated further on the matter, leading to the problems mentioned in this video: suck players being carried well beyond their actual skill, making the in-game measurement of Item Level worthless for assessing a player's skill. This lead to third-party tools coming into use and dominating high-end play, with predictable results of exclusion and complaints thereof.
In short, we have cascading failures of developers thinking through the full extent of their decisions, with each short-term fix causing the next problem. Instead they should have started at "How do we want players to play the game?" and work back from there to the mechanics required to achieve that end. There is your takeaway: Game Design deals with known quantities, so you start with the conclusion and work backwards from there to discern which rules and mechanics best achieve that conclusion.