One of the things that strikes experienced drug-company people is the cyclic nature of the industry. There's the boom-and-bust productivity cycle that's long characterized it, which I've had several chances to remark on. (Doubtless I'll have plenty more.)

But there are others. For example, how long should a drug discovery project take until it recommends something to the clinical development folks? A couple of years? Five years? As long as it darn well takes? Or if you think that's headed in the wrong direction, how about eighteen months? A year? Anyone for nine months? You can find companies that are trying all of these.

What happens, I think, is that a crop of managers get their experience under a given system, and while doing so they note the grievous failures all around them. Of course, there are some occasional winners, but there are always plenty more that never make it: fine projects, worthy ideas that never worked out because they just didn't have enough time (or just dragged on too long and got sidetracked, on the other hand.)

So, when they get the chance to run things, they resolve that they won't make the same mistakes. Nope, we're not having things drag on forever around here any more: you folks have a year to put it in the net, or you do something else. Or, conversely, we're not going to kill off perfectly good projects when they don't hit some arbitrary deadline: you folks keep at it until you get it right this time, OK?

This illustrates a general principle: just because you can screw things up in one direction doesn't mean you can't screw them up in another. It's a Manichean with-us-or-against-us fallacy that trips people up all the time. There's a case to be made both ways, of course, which is why this mistake has been around so long.

You really can run projects too quickly, sending half-baked slapped-together clinical candidates on to Development. And you really can let things drag on too long, pouring money and effort into a sinkhole because, hey, you've come this far already, and if you quit now you're just admitting failure, right? So the reverse approach, whichever one it is, breeds its own crop of should've-worked projects, and the managers in training resolve that they're not going to make those mistakes again. . .

Ideally, you'd split the difference, but that's the hardest way to do it. Alternately cracking the whip and witholding it can breed resentment at the perceived unequal treatment: How come those guys get to keep going when we're under pressure? How come that group got to present that junk as a candidate when they're still making us fix ours? Instead of just hearing one class of complaint, you get the whole spectrum.

Note, of course, that all the choices involve researchers complaining. It wouldn't be research, then, would it?