This is largely because people cook up all kind of bull-crap work just to keep busy and have it look like the equivalent of 'moon-landing' just happened.
Over years I have seen this phenomenon play out in every major company. The absolute worst I have seen is a whole team of 10-20 members dedicated to build that would be the equivalent of cron + ftp, implemented in super verbose java written and rewritten to make class hierarchies as complicated as possible to make it look big, and then have it projected as the biggest project ever. The team went to bag a lot of bonuses and promotions. Stuff like this is why goal based evaluation don't work.
Also look at: http://strikemag.org/bullshit-jobs/
The entirety of his argument has one data point: he has a friend who is a lawyer who thinks his job is useless for society.
The book company could then buy a standard product as a simple cost, while the security company attaches direct revenue to effective security.
I guess the IT/security company would need to run your infrastructure to be able to set up monitoring and such?
That can still be hard to accurately measure. You can measure, of course, how much revenue was attributed to sales of the product developed by the team, but how much of that is attributable to the team's work and how much is attributable to company-wide marketing and how much is due to company image developed by public experience with products developed by other teams is...not always readily obvious. (And, of course, then you have teams that aren't devoted to particular products or which provide company-wide services supporting product teams, but not directly generating revenue -- unless, of course, you organize your company so that accounting, HR, marketing, etc., are all broken up and product-team specific.)
Of course, if you measure by the entire company a lot of that uncertainty gets washed out.
I'm guessing that's true for a lot of areas in a large company. What about the department that takes care of buildings or HR for example?
Unfortunately, there's another camp that (with credible arguments) claims that task-estimates are almost always meaningless. If you buy into that camp, and if someone under-delivers, it doesn't tell you anything about whether or not they are performing poorly. They could simply be the victim of an overly optimistic task-estimate. As a manager, what are you going to do then?
This is pretty much the double-edged-sword in the NoEstimates movement. Either you judge people purely on the basis of results, and end up punishing people who are victims of poor task estimates. Or you don't judge people at all, and watch as people take advantage of the situation to slack off as much as they want. Or you judge them on the basis of hours-spent-at-work, which will then induce the same complaints we see in this thread.