Multivariate correlation time series data

I have a timeseries billings and usage dataset. I am trying to find factors correlated with churning accounts. For this I am trying to run a correlation analysis. Can we use time series data to do a multivariate correlation analysis? I have read several articles on issues like spurious correlation that make me wonder if we will be identifying correlation where it does not infact exist.

You may want to look into the cross-correlation or cross-covariance functions.

Alternatively, since you have a time-series, you can practice some other techniques, like decomposing the time-series or forecasting. Here’s a good, free book on forecasting methods.