Comment by texasbigdata

Building and maintaining the scraper is the not cost they would use to measure it internally. It’s the cost to build the API, and support it and perhaps any perverse incentive it creates where even more data flows out to competitors.