[roasm]

I generally agree with these points. We ended up with an interstitial the same way I hope most everyone else did: we A/B tested it, and the results were clearly positive on engagement and purchase.

We could choose to drop the interstitial as a matter of principal or on the assumption that the difficult-to-quantify long-term benefit will pay off later, but we'd have to make that decision not in the absence of data, but actually in opposition to the data.

Or let Google make the decision for us...