SimulScribe, the scrappy voicemail transcription company, didn’t get acquired exactly, but it just signed an exclusive partnership agreement with Ditech Networks that could be worth as much as $17 million. The deal is $7 million in cash up front with a $10 million earnout, and gives Ditech the exclusive rights to resell SimulScribe’s speech-to-text transcription services on a wholesale basis to telephone companies and developers. SimulScribe CEO James Siminoff will become the chief strategy officer of Ditech, and his co-founder Mark Dillon will also work there.
It is a decent outcome for a startup that raised only $5.7 million and is already profitable on sales of about $4 million, according to SimulScribe CEO James Siminoff. But competition is intense, with Spinvox on the one hand, which has raised an insane $200 million, and Nuance on the other, the speech-recognition behemoth which is nearing $1 billion in sales. SimulScribe offers its own voicemail-to-text service called PhoneTag, which has about 20,000 paying subscribers, and reaches about 80,000 more subscribers through wholesale partnershps with Vonage and British Telecom.
The big play is getting those deals with carriers, but with only 7 employees, SimulScribe didn’t have the resources or manpower to go after those deals. Ditech already has equipment sitting in nearly every major carrier’s network. It sells voice processing software that minimizes background noise on calls to the telephone companies. Adding a new service is a software upgrade. Ditech will resell SimulScribe’s service to its existing telephone carrier customers.
The companies will also combine their two technologies for more accurate transcriptions. Background noise is one of the biggest contributors to auto-transcription mistakes Simulscribe still supplements its transcriptions with humans, but the more accurate the auto-generated text is, the easier it is to fix.
If major telephone companies start adopting voicemail transcription services, it could quickly grow from what looks like it might be a $100 million market this year to a $1 billion million market. A lot of this adoption is being driven by a fear of Google, or more specifically Google Voice. Converting voicemails to emails is one of the great features of Google Voice, but its auto-transcription still needs a lot of work. Google Voice is using a homegrown speech recognition system developed for the GOOG-411 free directory information service (which is why it is not so great for conversational speech). But at least Google Voice is waking up carriers to the fact that the only thing voicemail is good for these days is converting it into emails so you don’t have to listen to them.