Just an addition (or an extended comment) to other answers.
I wouldn't say that people being conservative could not change time units quickly. They use time measurements everyday, it is more often than eg. measuring distances, however, people use money everyday also very often. Most money reforms (the largest are converting to euro in EU countries) takes at most two-five years. So it is not a big endeavor. And the time switch would also not be (if related to brains only, the financial effort would be large), but it should allow to use two units of time in parallel (eg. "old hours" and "new hours", or maybe "hour" and "dour" - "decimal hour"). The calendar switches were performed and nothing bad happened.
What is most important here, is that in fact no-one can force anybody to abandon the old unit. A government can make you use new money by removing the old ones from the market. Of course, it can punish for using "hours" instead of "dours", but it wouldn't be accepted by people in democracies.
The time reform would also be difficult, as it should be performed worldwide, probably in the very same moment. This is difficult and requires lots of effort. There are many things that are time-dependent, in today's era software, which 1. has to be modified, 2. has to be verified and tested, 3. has to be introduced at once, 4. should understand older time, 5. should be rapairable (support for all devices would require lots of people, and this is sure that software would have bugs). Also, there are devices that use time not to count your worktime, but which guard and control power plants, trains, planes, nuclear weapon etc. Do you remember the "Year 2K problem"? Even if all the measurements taken were not necessary, you can remember the effort that was taken.
The option why people are conservative is in fact the life experience "if it works relatively good, why should I change it?". Everyone knows that any small modification of a system is a risk of big collapse. The only problem with 24 hours a day (not 10 or 100) is when you need to calculate some time difference - how long did something last? But it's quite easy to convert parts of an hour to minutes, as someone mentioned, 60 is very good number for division, it has many dividers (2, 3, 4, 5, 6, 10, 12, 15, 20, 30), having 2, 5, 10, 20 - all the necessary that one use most often (take a look to your wallet and see, what bills / bank notes you have). Would it be easy to divide an hour on three parts if it had 100 minutes?
The main problem with time is then only when calculating time differences - how often do you perform it daily? How many times have you a problem to calculate "it started 4:45, ended 7:08, how long did it last?" Damn, that's easy: there are eight minutes between 7:00 and 7:08, fifteen minutes between 4:45 and 5:00, so together 8+15=23, there are three hours between 4 and 7, so 3 hours minus 60-37=23 minutes is 2 hours and 37 minutes. Is this really harder than subtracting from 100? How often do you need to solve "how much time are 157 minutes?" And - how often you need to know exact difference? Most people would say "about two and half hours".
This revolution could have been made before, for example 200 years ago, or even in Renaissance - not now. Would it be possible for the United Kingdom to switch from the left-side-driving? There are so many cars to be changed, so many traffic signs... (IIRC Sweden did it in the beginning of 1960s. as the last country). But is it necessary for the UK to do this swap? Who would benefit? How would this change pay back? When?
So the reason is that it in fact is not necessary. People do not have problems with this. Only those who count time, and software engineers who code software for them. I suppose this is a minority in each population.