Hi!
What a very late response from me, but hope this brings a new idea or two to your table!
I assume that you're used to using correlations and regressions to determine what variables in your survey data are most related to your KPIs, so I won't mention that here. But here are some other things I've done in StatsIQ recently:

• Utilized classifications and sentiment scores from open ended survey question coding in a regression analysis to "dig into" variables that we knew were important but we wanted to know more about. (Example: we know from our regression analysis of survey variables that wait time is a driver of how people perceive their doctor's experience, but we didn't ask any drill down questions about wait time in the survey. What part of wait time are people talking about in their open ends, if anything, and how is this impacting ratings of wait time? So, we used coding of the open ended questions in a regression with the wait time experience variable from the survey as the key (dependent) variable and ultimately determined that wait time experience is highly influenced by people talking negatively about the wait for the doctor to come into the exam room. I thought this was a neat use of the text/stats functionality.)


• A more basic example: we have a lot of demographic variables and want to know where the key differences in our KPIs are. I am able to quickly run multiple pivot tables, each with a demographic across the top and our KPIs down the side to determine which cuts of data display the biggest "deviations from normal" (so basically, finding our high performers and low performers). If I find interesting stuff from that - say that in my pivot table of gender, women have lower satisfaction scores and that in my pivot table of age, the 18-25 crowd has lower satisfaction - then I can start looking at the intersectionality of those two demographic variables to understand...are there ages for which the satisfaction scores of women are essentially equal to that of males? Are there age range/gender combinations that are even worse in satisfaction score than just looking at one demographic variable? This kind of exploration allows me to make recommendations to my stakeholders on where to put focus and explore further (example: we identified that mothers of young children who came into the ER were particularly unhappy, so gives us a sub-segment of the population to target for further research on what could make them more satisfied in their experience).