Remember its a computer, so GIGO (Garbage in garbage out).
You should have a pretty good idea of of your speedo error (usually under reads unless you are running non-standard 22 inch mags). The error on your consumption will be related to that error. The car overestimates the distance you are covering. So your actual consumption is usually better than the computer.
If you have replaced the fuel level senders they may be in wrong and not reading correctly. There may be some calibration involved.
There is a margin of error in the computer distance usually. You should get 50km further than the DTE figure.
It doesn't hurt to run it dry once with a can of petrol in the boot. Take note of the ODO or better yet GPS distances that the DTE hits 0km, the petrol light comes on and when it dies. When I started out I had a E30 325i that used to run on fumes frequently and I got to know the car very well!! I once did a trip from Welkom to Grahamstown that I precicely timed to get the light to come on when I got into Grahamstown (picture James May on an economy run) It came on a few seconds after I stopped at the destination.

imp:
So my experience with computers is quite good. The Jeep's is also quite accurate as with bigger wheels she is now spot on speed with the GPS.
If you have a bad input it will throw the computer. Input injector pulse width, speedo/ odo for distance and for calculating DTE it is probably some average consumption of the last 100km or something). Ofcourse if you have been driving Miss Daisy for a week and suddenly have to drive your pregnat wife to the hospital I wouldn't rely on the DTE.
I am not too sure but I think that DTE uses the fuel level sender and average consumption so changing road conditions, inclines and aggressive driving will make the fuel move around and it will overestimate the fuel in the tank.