zlacker

I think your diagnosis is on target: "the cost was real and up-front but any benefits were largely hypothetical based on enough other people deciding to pay at all and some of them building apps."

But freemium is hard: Either you make a market for ads, and have intrusive data collection and analytics, and you have to compete with the most intrusive ad-supported products that already dominate the ad market and you become as bad as them, OR you have pure freemium and you have the question of how to make the up-sell compelling enough while keeping the free level of service interesting enough. The only good example of success at freemium is LinkedIn where they segmented their user population and only made the recruiters pay a very high premium price.