> You're also somewhat selecting for users who have disposable income, which can't be great for ad value.
Many years ago I worked for a company that had tens of millions of US subscribers, my job involved modeling their behavior in order to allocate resources at least a week in advance. The law of large numbers is pretty amazing to see play out in front of you like that, where you can clearly see the bright lines between your market segments - fundamentally different kinds of people. I have a feeling that there is only one kind of person who would pay for twitter, which will very likely end up as a flag in a marketing dataset that certain companies would find well worth whatever twitter charges them (or their data-broker). Not unlike Volkswagen, on the eve of a big sales push for beetles, wanting a list of everyone who regularly buys peanut butter and cat litter.
Many years ago I worked for a company that had tens of millions of US subscribers, my job involved modeling their behavior in order to allocate resources at least a week in advance. The law of large numbers is pretty amazing to see play out in front of you like that, where you can clearly see the bright lines between your market segments - fundamentally different kinds of people. I have a feeling that there is only one kind of person who would pay for twitter, which will very likely end up as a flag in a marketing dataset that certain companies would find well worth whatever twitter charges them (or their data-broker). Not unlike Volkswagen, on the eve of a big sales push for beetles, wanting a list of everyone who regularly buys peanut butter and cat litter.