One's own accomplishments tend to look inconspicuous. Two aspects of my recent work are archive and backup--neither interesting at all until something goes badly wrong. (We're twiddling our thumbs on the archive plan: not sure if the delay is a NSF vs DOE thing or lower level negotiators.)
Go to collaboration meetings for a few years in a row and the patterns emerge. The great product that will ease everybody's life actually is, but it takes quite a bit longer to get into service than they estimated. The promising approach doesn't work so well as hoped, and the noise simulation still doesn't quite match the data. But it is closer.
Talk about your inconspicuous service! Estimating the noise rates accurately is critical for a number of different low energy analyses (e.g. supernova detection), but it doesn't have the same ring on the resume as "cosmological neutrinos". It probably needs a couple more man-years of work, too.