Interesting technique, that DID. But it assumes the non treatment factors would affect both the treatment group and control group equally, that the effect would scale linearly. If the treatment group was more exposed to the non-treatment factors, then an increase could account for a larger difference than the one seem at time 1. Idk which other industry they used as the controll group but interest rates could have a superlinear effect on tech as compared to on that, so the difference of difference would be explained by the non-treatment factor too