It's hard to predict how this will develop in the future. As is, I have yet to hear any positive success story from my peers. All I've heard of ResearchGate are complaints about their invitation spam.
If you sign up (or are already signed up), make sure to carefully check all settings. There are some "hidden" settings that will repeatedly send out "Invitations" and even "Invitation reminders" to your peers (make sure to disable this right at signup time, before "claiming" a single paper!). You can imagine that if your peer gets a dozen of them each week, he will get annoyed. And ReseachGate uses your name for the invitation, so he will get annoyed at you.
Before "requesting" a paper on Researchgate, make sure they have signed up. Otherwise, use regular email (even better, just use regular email right away!)
When you request a paper from a non-user, they will be sent an invitation letter, but they cannot answer without joining: researchgate does not include contact information in your request!?!
Other than that, I have the impression that ResearchGate is mostly used by students, not by established researchers. Therefore, do not expect much high-quality interactions to happen.
Most of my peers that were on RG once, seem to have left. This is not a good sign.
If they don't manage to get high-quality interactions happen soon, their name may end up remaining associated with spam. They may have been too aggressive at boosting their user numbers quickly. The overall idea is good, but they need to find a way to get quality contributions, not only quantity. As is, I see them only strong on the quantitative side so far...
Honestly, in my opinion there is one thing really important for the visibility of your research:
- Make it available. Publish with open access, or a publisher where you can put the preprint on your webpage. Make sure that Google Scholar indexes it, because that is what everybody I know uses for searching literature. You must get Google Scholar to return a PDF link for your article.
- Make it available. Also put data sets, detailed description, source code, etc. on your web page; offer collaboration.
If you make it easy for others to build upon your work, they are more likely to do so. In particular, you must make it easier to build upon your work, than on alternatives. Making a RG profile does not really help; effectively it just means your data is spread to one more site. Instead, put a lot of information on your homepage.
Here is an example. I'm convinced that one of the reasons why he has been very successful is that you can download data and software and tutorials right on his homepage. Compare to his ResearchGate profile (I was even surprised he has one) - it's a dead duck, absolutely useless. A Google Scholar Profile may be much more useful, as it will be linked from any article, making it easy for people to reach your other publications. Again, the same example: Google Scholar does a much better job at providing an automatic publication list for you.
IMHO, Google Scholar is a must, because the profile will make it easier to get updated on your other publications. I visit Scholar profiles quite often; in particular to see the latest and the most cited work of an author of which I already have discovered a good article.