Product growth: Launch premium service with rev share biz model [004.3]

Premium ServiceAs I mentioned in my earlier blog titled “Product growth: Scaling the SaaS platform [004.1]“, we had realized within the first month of launch of visual voicemail that cost per subscriber per month business model was not sustainable and we had to somehow figure out how to deliver voicemail to text premium service using automated transcriptions… YES using Automated Transcriptions. This was year 2008 and only human or hybrid transcription based services were available.

The biggest challenge was the cost of the service due to which it was not possible to sell to thousands or millions of users. To bring down the price point, we had to somehow use automated transcriptions only and accuracy had to be be good enough to launch commercial voicemail to text service.

We had the voicemail user base but did not have the voicemail to text transcription technology. So we started engaging with speech transcription technology providers to partner with in the marketplace. After several months, we were down to two providers named Yap (acquired by Amazon possibly powering Amazon Echo product) and Ditech (acquired by Nuance). We worked with them to further improve their voicemail to text engines and fix a few deployment issues to setup a trial with our Carrier customer around April 2009. The trial was deemed a failure by many folks from Carrier customer team but for us and the Carrier product management team it was positive sign of hope. We reached a joint conclusion that it was practical to launch this service IF & ONLY IF we could get the accuracy increased to a level high enough to get between 5-10% of adoption.

But how do we measure accuracy? Every speech transcription provider had their own method of measuring accuracy. Finally, after extensive research we found a standard program from NIST that measures speech transcription accuracy. After multiple iterations, we reached the conclusion that if we could consistently get 82%+ transcription accuracy from the voicemail to text engine then our business case of 5-10% service adoption will work. We worked closely with Yap and Ditech for months and finally we had a breakthrough in Q1 of 2010. Yap hit 82% accuracy consistently against new voicemails day after day. We conducted another trial with our Carrier customer and this time around the feedback was that if launched right this service could generate significant revenues.

There were several challenges with launching the voicemail to text service:

  • Accuracy was high for speakers with American accent only, otherwise poor.
  • Carrier customer directive was to bundle the service in a manner that users should never call customer support for help at any time.
  • The service adoption and provisioning process had to be 100% client application driven for trial, provisioning, and cancelling of the service without any call to customer care.
  • Users cannot be auto opted into the voicemail to text 30 days free trial, as per legal team of Carrier customer.
  • Price point should be below $2 a month even though others were charging $4.99 and $6.99 per month.

We addressed all of the above and designed the user experience for the launch as follows:

  • Users had to manually opt-in to the voicemail to text service trial for 30 days. (we decided to offer 2 weeks more than anyone else to get higher adoption)
  • Trial was auto opt-out based such that the voicemail to text service expired automatically after 30 days. (unlike most services which continue and start charging from 2nd month)
  • Users had to explicitly opt-in to the paid premium voicemail to text service and it was priced @ $1.99 per month only.
  • In addition, users had the option to stop the trial at anytime. (since transcription accuracy was low for non-American accent individuals we provided option for them to stop asap)
  • To ensure adoption of trial and subsequently paid service, we added limited notifications for users to nudge them from inside the voicemail application.
  • The service provisioning process was completely user driven from the client application integrated all the way with the Carrier billing system in the backend and voicemail to text engine of the provider.
  • In addition, we provided a bonus auto forward to email, which became a favorite of most users since they did not have to open the voicemail app to read the voicemail message.

After we received approval from our management team and Carrier customer by end of June 2010, we managed to launch the service in 4 months flat. This was first of a kind service launched by any Carrier worldwide and we put it together working with multiple Carrier teams … product, marketing, device, network, uxd, legal, supply chain, and others. Also, we worked very closely with our voicemail to text engine provider Yap to help them scale the service to support the project. Overall, it was a great project working very closely with Partners in launching a successful service.

Since we were very confident about our assessment and adoption numbers, we proposed a revenue share business model to the Carrier with higher percentage for us and they agreed. Carrier wanted to get rid of the high cost of visual voicemail service per subscriber per month and we took a calculated risk and got a Win-Win with the customer.

In less than 2 years, we had 500K paid subscribers. It was good but not fantastic as we were limited by the opt-in process.

In the meantime, Yap was acquired by Amazon and overnight Amazon wanted to shutdown their voicemail to text service. But with intervention of the Tier 1 Carrier, we got 1 month extension during which we switched from Yap to Ditech successfully. This was another major milestone since we had to take care of the traffic as well as the accuracy of the service without any downtime whatsoever!

Next up… data analytics, client upgrades & server based client configuration.

Previous blog in product growth series: Product growth: Impact of acquisition [004.2]

Leave a Comment