Sramana: How did you determine which clients to build custom offerings and courses for?
Shiv Rajendran: We selected them based on the market we wanted to target. We knew we wanted to target business people, so it made sense to customize courses for individuals who were business people themselves. Students were another easy category for us because it was easy to tell why wanted to learn English.
We ended up developing a system whereby the content was produced fresh every week. If a group of students joined who were interested in kangaroo farming, then within two weeks we would have lessons developed based on kangaroo farming. Our system was adaptive because we had real human teachers producing the content each week. The teachers took inputs from the students and developed content specific to those interest. Our system made sure that the developed content was delivered at the right time. It was a sophisticated system in terms of content development and scheduling to ensure that students got maximum value.
Sramana: I am curious how you managed all the permutations of accents, time zones, and contextual scenarios. It sounds like a complicated and cumbersome set of variables.
Shiv Rajendran: It was a very cumbersome set of variables. Before we could deal with them we had an even bigger variable, and that was that in 2005, hardly anybody knew what a virtual world was. When we brought people in to learn in our language courses, but the first thing they had to do was learn how to use the virtual world. We solved that problem by hiring customer service representatives and making them available 24 hours a day. As soon as a new student appeared, the representatives would guide him or her through the initial setup and direct that person to the class. That was very expensive.
Sramana: How many people did you need to do that? I may be simplistic, but that sounds complicated.
Shiv Rajendran: We needed one representative online at all times, and it would take five or six people to cover a 24-hour period. There was a benefit to having a human fulfill that customer service role rather than a software tutorial. When a new student joined, that representative could engage with that student to determine why he or she wanted to learn English. That engagement made it possible for us to categorize and group our students. Business people were an early target segment for us. Business people from Europe wanted to communicate with Asians and Asians wanted to communicate with business people from Europe. There was a lot of overlap. We could put those groups in classes together that allowed them to communicate with each other.
Sramana: How did you acquire customers?
Shiv Rajendran: Initially we ran through 500 students free. We put advertising on a range of sites and allowed them to join at no cost. From there we started acquiring customers through PPCs. This was in 2007.
Sramana: Language learning is a fairly complicated space. What was the PPC landscape like?
Shiv Rajendran: It was very difficult. The web is saturated for English language training, and it was saturated back then as well. There were some massive companies with huge budgets. We had no hope of competing with them. We actually found a way of bidding for the lowest position on the Google search page, and we did that by complete accident. That got our advertisements out there and brought us traffic. After that we did not focus on anything other than getting our ad in that very last spot. We also targeted countries in the Middle East and China because larger companies were ignoring those areas.
We also focused on our messaging. Our message was to learn English from a qualified teacher in the US or the UK. We did not mention virtual worlds or our technology. That got people in the door. From there we had to ensure we could get them up to speed with the technology so that they were able to interact with our teachers.