The backend (scripting) needs improvement to be more like a real ETL BI tool. The script editor is, in my opinion, no more than a basic scripting editor with too-limited debugging features.
I have been using it for five years (still today).
I have not had any issues with deployment, stability or scalability.
Technical support is very good.
The initial setup was straightforward.
We usually implement in-house, but we also rely on external resources.
We are currently using two different BI solutions for our customers: Business Objects and QlikView.
QlikView lets our customers develop data analysis projects independently with a little initial training.
Only a recommendation in order to assure the growth of a project with minimum impact on the previous deployed portion: As much as possible, try to solve data complexity on the backend side (where the logic is easily documentable), rather than in the front-end using only analytic functions.
