It was originally for PC, and I tried it out with a boat load of upgrades in my G4, and my score said: 292. That's like a 3-point upgrade from the STOCK version of my Mac!
Depending on the mac and the particular upgrades you added, they very well have amounted to nothing. For instance, if you added more ram to a system that already had sufficient ram to run the benchmarking program then it would've shown no increase in performance. In fact, having additional ram sometimes slightly does slow down your system in reality because it now has more memory space that it needs to keep track of. All benchmark programs that I know of will show no benefit to adding more ram.
Another example is if you added a faster hard drive to a Mac that has a slow system bus. If the computer's system bus's bandwidth is saturated, it doesn't matter how fast the new hard drive is, it'll give you no performance benefit.
So which Mac did you do these upgrades to, and what were the upgraded components?
EDIT:
And also, if you're comparing your upgraded system to the score listed on that site for the stock machine's score, then your making a completely false comparison. Just look at the benchmark results site for xbench and you'll see that different people's same-model machines have very large differences in performance from one to another, because all machines are different. Even if you assume that all systems of the same model have exactly the same performance when initially taken out of the box, which is not true due to minor silent revisions to the hardware that's officially released as the same model, even then, as soon as you do anything at all to your computer it'll affect its performance. The systems they use for testing are new machines that have fresh installs. So if you're comparing that system's score to the score your computer reported, then you're way off base to use that as proof that their comparison results are inaccurate. Your old computer will have a fragmented disk, extra services turned on, extra drivers installed, background programs running that you're not aware of, etc. etc. So just because your system reported a lower score than their test system of the same model, that doesn't automatically mean their comparison ratings are off base at all.