In chemical engineering, a "pilot project" is an attempt to move a process off the lab bench and into a small set-up that will allow the researchers to find out if their new process will actually work in a commercial setup, without the risk of a full-scale industrial plant. They regularly demonstrate that the original process, which worked well in the lab, doesn't scale up. That is, they fail.
In the library setting, people regularly talk about "pilot projects". In general, this seems to be code for "We're going to do this, but if we just tell staff that, they'll get upset." At least that's the way some staff seem to interpret such. This interpretation is reinforced by staff who react to any unsuccessful attempt to provide a new service as a total failure that throws the entire library systems into disrespect in the eyes of the user community.
It seems that the way around this for new web services is to label them as "beta", since "users are familiar with 'betas' from Google, so we can change it any way we want without going back to the committee for approval." Whatever we call it, we need to encourage trials, and we need to be honest about what happens when something doesn't work: we fire the... I mean, we learn from the experience and try to improve the services we continue based on what we have learned.