Why not SMS in Africa?
I read an interesting post from Steve Song earlier entitled “Why SMS in Africa” which outlines why SMS will remain relevant in the African context for some time to come. I agree with Steve that both IP and non-IP technologies should be considered but want to elaborate on the challenges SMS-based service face.
Naturally, through our work on Mobile Researcher and Mobenzi, we’ve had a lot of technology decisions to make along the way. It goes without saying (although I’ll say it anyway) that I do not believe in a “one technology fits all approach”. SMS is obviously great for ad hoc, instant communication and a range of other communication-oriented tasks. But it falls down in a lot of other, more complex (and sometimes more valuable) scenarios.
Here are 5 reasons why:
1. Usability
Although people are familiar with how SMS works, they aren’t normally familiar with what your service is or how it works. I’ve seen loads of solutions which force users to format responses (commands, requests, etc) in cryptic ways so that a system can understand them and render the correct service (e.g. looking up information). This is one of the main reasons we’ve built Mobenzi in an attempt to improve the usability of SMS by adding real people to the mix.
2. Cost
Although SMS costs are probably coming down (I hope?), when one considers the transactional cost of an SMS, it’s a hellishly expensive way to conduct multiple interactions with people. Cost can even be a factor with reverse-billed services because SMS has a perceived cost which can act as a barrier to user participation.
3. Service portability
If you’re planning to use a short code, you can almost forget about using the same one worldwide. It’s often difficult, expensive, and/or impossible to register the same code in different countries. This again reduces the usability of the service (or at least increases the training you need to do in each country).
4. Connectivity
While SMS coverage is generally much better than data coverage, it’s still patchy in areas (even in very mature mobile markets like SA). Using a system which can’t work offline (rather than a smart client application) puts you at the mercy of coverage dynamics.
5. Building for the future
I don’t have anything more than anecdotes and my own experience to support this but as if you look at even the low end models being brought out these days, its a relatively small percentage that don’t have internet connectivity. If you’re going to base your service on SMS now, it seems somewhat sensible to have a medium term plan to migrate it to IP. And be prepared to expend significant resources in that process. Swapping an engine whilst in flight might not be impossible – but rather expensive I would postulate.
Trying to take sides or apply a “technology x is better than technology y” approach is futile and in my opinion a waste of time which could be better spent leveraging an appropriate technology to solve a problem.
I believe the biggest barrier to wide-spread adoption of IP-based mobile services is that operators and manufacturers are generally pretty poor at shipping “internet-ready” devices. Don’t even get me started on signing applications which are able to use protected handset APIs. That’s a rant for another, longer blog post.
Valid points Andi. Illustrates that mobile environment is complex and that there are no simple axioms. In particular, I agree about the cost of SMSes which are frankly ridiculous. I think that operators are missing an opportunity to actually make more money by dropping the cost of SMS.
Have you considered capturing the thinking and experience in pattern format? Rather than discuss pros/cons, you can capture the problem to solve, the context and the good and bad aspects of a particular solution. Context is key.