Friday, September 10, 2010

Can Voice Messages Replace SMS in Africa?

This is the question posed by Professor Eric Brewer and his team at Berkeley. Eric is a Visiting Scientist at Google, and I had the privilege to chat with him last week while I was in New York. One of the subjects we touched upon during the conversation was a study his team had done last year, implementing a new system of asynchronous voice messages in rural Uganda to supplement voice calls and text messaging (Heimerl et al, 2009). SMS (i.e. text messaging) has become hugely popular in Africa for several reasons: low cost, asynchronicity, and the unreliable cell phone coverage. Let me unpack each of these components.

Cost. With voice calls often costing between $0.10 and $0.20 and text messaging costing only around $0.04, it can often be much more cost efficient to communicate simple messages through SMS.

Asynchronicity. Often a person simply wants to communicate a message without requiring immediate consumption or feedback. I can only speak as an American, but in the US, the immediacy of voice calling has in some situations come to be seen as obtrusive, whereas SMS allows recipients to digest and respond to messages at their leisure.

Unreliable Coverage. Connectivity across Africa is shoddy at best, and in rural environments, it can often be nearly inexistent. "Sometimes [users] have to wait hours, hike miles, or climb trees to have voice conversations," (p. 4). Furthermore, asynchronous messages provide the extra advantage that only one of the two parties needs to be within range of the network at a given time to communicate.

However, there are still issues that prevent SMS from being the sole method of communication across Africa. Among others, these are their impersonal nature, high rates of illiteracy among the population, and the difficulty of typing in local dialects. Users in Brewer's study expressed a clear preference for voice communication, citing that "hearing a particular person’s voice made the conversation more serious for business purposes, and more meaningful for personal uses," (p. 4). Furthermore, with high rates of illiteracy and local dialects that do not lend themselves to English character set phones, SMS is not always a feasible mode of communication either.

Brewer's group seeks to take the most valuable aspects from both voice calling and SMS to combine them for a novel solution: asynchronous voice messaging. AVMs (my own acronym) offer a potential low cost alternative to calling while also allowing users to convey the personal essence of one's voice. They circumvent the limitations of illiteracy and local dialects while also allowing recipients to consume messages at their own pace. Since AVMs do not need to be delivered immediately, providers could transit them by utilizing excess bandwidth at very little marginal cost, theoretically passing that savings on to the consumer. To quote the paper, "the resources of a synchronous communication system are idle most of the time. This is due to the resources required for synchronous communication, which must be provided exactly when needed. Asynchronous communications improve this situation by allowing for load balancing, moving messages to times when the system is under-utilized," (p. 2).

Brewer and his team conducted a pilot study with approximately 230 individuals in ten different villages, and their findings were hopeful, providing a solid proof of concept for this potentially major paradigm shift in African telephony. What do you think? Could voice messaging provide a viable alternative to SMS in Africa? Is this an area Google should be looking into?

Sign up for email updates here.


Robyn said...

You refer a lot to the fact that connectivity is shoddy in many areas of the continent, and it seems as though what you're talking about is a way of circumventing these issues to some degree. But is Google Africa also addressing the underlying issues behind what you call "shoddy" connections - the nature of poverty, politics, and class, who has access to technology and who doesn't? Maybe it's implicit in Google Africa's mission, but articles I've read on the subject seem to treat these issues as an afterthought, an obstacle in the way of spreading technology, as opposed to the source of the problem. An image of bottlenecking comes to mind - companies trying to push tech development in areas that are still unable to support it. Maybe you could write a post about it?

Benjamin Cole said...

That's a great point, Robyn. Google is actually pushing hard on this issue, working with telecom companies and universities to open access. Lots more about it here:

It's not totally within the purview of my job, but since the team is so small here, I'm sure I'll get to see a lot of what's going on, in which case, yes, it would definitely be a interesting topic for a post.

Eric said...

How would AVMs work? Would it be little more than the transmission of brief voicemails?

Benjamin Cole said...

There's actually an existing standard called MMS (Multimedia Message Service). It's the same way phones currently send combined picture/text messages.

More here: