Astronomers were astonished when their initial observations using the Doppler technique immediately revealed Jupiter-sized (and much larger) planets so close to ther suns that they had orbital periods of only a few days or weeks; no one had expected giant planets anywhere near that close.
As yet, we simply haven't been making observations long enough to be able to detect any gas giants out at the distances from their suns where gas giants rightfully should be -- and so our sample of observed extrasolar planets is completely biased in favor of such "hot Jupiters".
As I've said, about 5% of stars have turned out to have them -- but the total percentage of stars that have planets of some kind is thought to be much higher.
As yet, though, we don't know how much higher.
Second, this technique can only set a lower limit on the planet's mass -- that is, a measure of what its mass is if the planet's orbital plane just happens to be perfectly lined up with Earth's line of sight to the star, so that the planet is dragging the star directly toward and away from our Sun. But, naturally, in reality such planets must have orbital planes tilted at all sorts of angles relative to our line of sight.
Even a huge planet cannot be detected at all by the Doppler technique if its orbital plane is at right angles to our line of sight, so that it only drags its star from side to side as seen by us, and not toward or away from Earth at all -- and since it's a certainty that all the planets we've detected so far have orbits tilted at some angle to our line of sight, the Doppler technique can sense only part of the speed with which they are pulling their suns around. Thus all we can do is calculate the planet's minimum mass -- and an estimte of its "average probable" mass, based on the equal odds that its orbit could be tilted toward us at any angle from 0 to 90 degrees.
So in reality, it's a safe bet that our estimates of some of these planets' masses must be in serious error.
So we're looking for other techniques.