Jott announced that they are discontinuing their free service effective February 2, citing a terrible ad market that is insufficient to support the free paid product. Customers will need to upgrade to paid plans at $3.95 per month or $12.95 per month to keep using the free service.
It's a surprising move, because the company says 30% of free users ultimately upgrade to paid versions. It's a high upgrade percentage, and seems at first blush like a great example of the freemium model at work.
It's an unsurprising move, because the cost of supporting these free users with a service that uses human transcription has to be very expensive, even if some speech to text automation is used for preprocessing. If the cost to support a user is even a dollars a month, it's easy to see to company going further into the red as the model (and free base) scales. The profits off the paid base are outweighed by the costs of supporting the free base.
It seems less likely that the struggling ad market affected the decision, because the most that advertising could contribute per user per month is a fraction of a dollar. Hit a user once per day at a $10 CPM and you're making only $0.30 per user per month. The ability to offset the human transcription costs with advertising seems challenging in good and bad ad markets alike.
Let's suppose that the entire ad market evaporated and Jott lost all advertising revenue. Since roughly 30% of the base is paid, Jott could offset that loss by increasing the profit on an average paid subscriber by $1 per month. They could do this by introducing new features, introducing a new subscription level, or optimizing price for profit - and maintain a free service and a freemium model that seems to be working well.
Since they didn't choose this path, and chose instead to sacrifice the freemium model, I suspect the costs of the free service are more significant than the reference to the ad model suggests and the profits on the paid service aren't near what's needed to offset the marketing expense. [Note: I've got zero insight into the actual economics of Jott's model - if my calculus is bad, I'd be grateful for a correction.]