>Startups generally have a bias towards doing things that they can measure accurately.
As far as biases go, preferring to pay for things that have a demonstrable effect is a pretty good one, especially for startups that generally don't have the resources to waste on a large number of failed initiatives. Sure if you have some incite that gives you a better than normal chance of getting a very high ROI on a creative initiative, then go for it, that's what startups are for. But if you don't have any special incite, stick with what has been shown to be effective, and avoid things like Facebook advertising which appear to give low or negative ROI in the typical case.
As far as biases go, preferring to pay for things that have a demonstrable effect is a pretty good one, especially for startups that generally don't have the resources to waste on a large number of failed initiatives. Sure if you have some incite that gives you a better than normal chance of getting a very high ROI on a creative initiative, then go for it, that's what startups are for. But if you don't have any special incite, stick with what has been shown to be effective, and avoid things like Facebook advertising which appear to give low or negative ROI in the typical case.