This is a good bit of perspective. There are 351 total D1 programs and 32 conferences. Each of those teams will conservatively play around 30 games by the time selection Sunday arrives. That means that in any give college season there are well over 10,000 games (a/k/a data points).
The committee has to take all that data and whittle it down to a field of the best 68 teams and seed them, and they have to do it in the most fair/objective way possible.
Based on that, I can all but assure you they don't dwell on any one loss or win or margin of victory the way that fans tend to. The biggest reason why is they physically cannot. The questions would become endless. Where do you draw the line on what you would consider a really bad loss based on margin? Is the line 15? 20? 30? What about a game like OU vs. KU where it was, actually a blowout until OU pulled their starters and KU cut it to 13 on the steam of some meaningless points in the last few minutes. Is the committee expected to analyze every margin of victory in this way?
This is why objective tools like NET, BPI, KenPom, SOS, RPI (less so now), etc are so important. The committee cannot possibly analyze every game NCAA game on the subjective basis being suggested here.