Does it really matter one way or the other?
The question that matters is whether or not FitBits led wearers to exercise more, not whether it led anyone to lose weight.
According to that same 2 year study, the answer to whether or not FitBit users moved more was no.
That's both disappointing and not particularly surprising. It's disappointing because moving more can markedly improve health and quality of life at any weight. It's not particularly surprising because no one has yet figured out how to affect a population based sustained increase in physical activity.
No doubt there are specific individuals who are motivated by their trackers and do in fact sustain long term changes to behaviour, but once again here we have an outcome that speaks against the ongoing singular focus of personal responsibility as the means to improve healthy living.
If we want people to move more we need to find ways to make doing so easier, more fun, and/or more valuable. Simply telling them to move more, or apparently showing them how little or how much they move, isn't going to be enough in and of itself.
[UPDATE: The study though did have a major limitation in that it used a now obsolete wearable. It's possible that the results with newer wearables would differ. Here's Aaron Coleman's coverage of same. Thanks to @Chimbo23 for pointing me to it.]