The retail example is an oversimplification. When the consumers of a company's product are also its primary source of revenue, nobody is under any illusions about who that company is responsible to. You are the person most impacted by Gamestop's customer experience, and you are also the main way Gamestop makes money. Your satisfaction is paramount, so it makes total sense for a company to fire people who don't deliver. Similar example: there's nothing irresponsible about insisting your public relations firm writes you a flattering press release, because everyone understands that a press release is paid for by the company it concerns.

A think tank is a different sort of animal. The donor is not the main entity impacted by the policy proposals the think tank produces. Just like a responsible media outlet must have a notion of journalistic integrity, or a legislator's office should feel beholden to its constituents, a think tank has a responsibility to provide sound policy advice to its readers.

You can take the position that "think tanks do not have any level of responsibility to the people who consume their policy proposals" but I certainly would not follow you there. Policy institutes present themselves as rigorously data-driven nonprofits and write with an impartial air. If criticism / negative results are known to lead directly to a loss of funding, then it is disingenuous for the think tank to present itself as impartial, and I would consider it compromised.

We tend to solve this problem with regulatory agencies that represent the impacted consumer (and not the source of revenue), though I would note that this is the very solution the author was fired for praising.