@parrot.pal: Every case file so far has tested a published study against our data. This one tests our data against itself. We run two apps. They share no user data: different products, different audiences, different databases, and in one case a four-year head start. Engaged women in Turtle Method give weekends 23.0% of their logging days. In ParrotPal: 23.5%. Two thirds of light users in both never log a weekend at all. Own-label supermarkets take the top five brand spots in both. The biggest advertised brand reaches 44% and 40% of Tesco's members. Four separate measurements, two separate datasets, no shared users. One app's data is an anecdote about that app. Two apps agreeing is a fact about people. That is the whole reason we publish both.