Data can help us discriminate between creative options. A retailer might test one creative option focusing on quality against another offering a discount for example. But if the resulting data suggests that the discount was more successful, that’s not to say the right move is to start discounting everything.
In scientific studies, the bigger the data set the better. A clinical trial with 100 people is better than 50 people. 1,000 people is better still. No doubt about that. It's the same story in marketing, right?
http://www.thedrum.com/opinion/2018/05/10/what-s-the-deeper-worry-around-big-data-and-creativity
