the other thing is you know i said we focus on the tradeline data. this credit card data. but if you mix in other data sets, you're going to get more information. maybe a data set that's not as predictive as the credit card payment data. but still has some kind of caloric value. >> and a frequent guest on this show is doing something with a firm. where he says i'm taking although he uses your fico data, too, doesn't he? >> yes. >> but kind of what you're talking about this revolutionizing finding other ways of predicting how somebody who maybe doesn't have a great score, but there may be evidence that they would pay back that loan? >> yes, very much so. in the old days you would look for a single data set that had the most predictive data. >> and it worked well. >> and we still use it. today it's much easier to mix and match additional data in. rental data, utility payment data. telephone payment data. >> laura i interrupted you, let's take you with the last question there. >> this is all personal data. how do you incorporate sort of external events which can disrupt every