By this measure, John Harper's design of Agon 2nd ed is bad design, or Luke and Thor's design of Torchbearer 2e, given that all three would be well aware that there games would be more appealing if they were 5e D&D variants.
I regard this as a reductio on your inference to bad design.
This assumes the D&D and Agon or Torchbearer are all targeting the same marketing segments.
The reality is that virtually all RPG design is in reaction to D&D, in as much as they target parts of the RPG market that want something
other than D&D. And within those smaller segments of the market, John Harper and Luke & Thor are certainly aiming to make their games as appealing as possible. I mean, that's ostensibly
why there's a Torchbearer 2e. Having gotten feedback through the release of its first edition, the designers endeavored to keep what everyone liked about the first edition, and improve on it for greater appeal, within that marketing segment.
As far as I can see, "do we design for good design, or do we design for mass appeal" is a false dichotomy. Good design is by definition design that appeals to its target audience. The only question is, how that target audience is defined.
Which is of course the crux around 4e. For certain segments of the market, it was tremendously well-designed. For other segments, it was a complete failure of design. The interesting question, to me, is what makes up those respective segments? And for 5e, what are the segments that it is appealing to that have made it such a success, and what are they looking for in design that makes 5e appealing to them? For good or ill, that is not the conversation that we have in the usual online spaces, because people prefer to define "good design" by what appeals to
them.
As far as 5e's playtest surveys go, the sample is inherently biased, but that is not really an issue for WotC's purposes. One of the lessons WotC took from 4e was the
Gnome Effect. Which is, you might find that only a minority of your players like playing gnomes, so it seems like a good idea to remove gnomes from the PHB and replace them with something new and fresh. But while those gnome-players as a category may only be a minority, they may be in a majority (or at least half) of the
groups that play. So the effect of removing them is far greater than anticipated.
All WotC wants to do is avoid landmines. That's what the surveys are for. The surveys don't drive design, they only tell WotC what isn't working. It still falls to the designers and developers to come up with new iterations that satisfy a super majority of respondents. And the segment of the population that checks out the playtest and responds to the survey are precisely the segment that will lead any backlash to design decisions they make.
