4 Tips on creating an African micro-blogging platform

2009 July 9

Twitter has yet to hit mainstream in Africa, it may have jumped the chasm in the US spurred by Oprah, but the service is mostly being being used by geeks and bloggers here. Some entrepreneurs are stepping up to the plate in the hope of becoming the Twitter of Africa. In the past few months i have seen a number of micro-blogging services pop up.


We all know about some of the challenges related to bandwidth, high costs and not enough of it. This should change soon with a number of new cables, but it’s still a challenge currently being faced.

As an example: Twitter desktop clients really consume bandwidth. Try leaving a desktop client running for a week during offices hours (10 hours a day), you will eat through your bandwidth cap very quickly. It’s one of the reasons i mainly use the web. Elan Lohmann learnt this after he hit 6gb in one month.



Desktop clients, constantly poll and download the entire XML/JSON file using the twitter API, which is fine and dandy if you are using bandwidth you do not have to pay for (i.e at work) or you are somewhere where bandwidth is cheap. There is no ‘push’ of just the new messages. Imagine for a second you only follow 10-20 people who update infrequently, your desktop client would be constantly updating with the same messages over and over again which is not the optimal use of bandwidth.

So here are some tips, for building micro-bloging platforms in Africa:

1: Reduce the overhead (XML/JSON etc.)

Mxit modified the standard Jabber protocol to reduce the bandwidth used by their mobile instant messaging client. From Defza:

Mxit does not use the standard client2server (c2s) protocol, because it would get too expensive if the standard jabber protocol in XML was used:

<message to='romeo@mxit.co.za'
<body>Wherefore art thou, Romeo?</body>

Instead, mxit had to come up with their own c2s protocol in order to ‘compress’ this into something… cheaper (less bytes of data)!
So, we have this instead:

ln=69.id=juliet.cm=10.ms=juliet@mxit.co.za.Wherefore art thou, Romeo?..1

So, instead of 146 bytes being used for the message, only 72 bytes or so is used. (plus 9 bytes for the confirmation message)

2:  Rather then constant polling, push updates

Constant polling by desktop clients can be expensive with regards to bandwidth a more elegant solution would be to push updates alone, this will reduce the bandwidth costs of running a desktop client. This is where an intermediary service like MyQron becomes useful, which does the polling and then only pushes updates.

3:  SMS Integration

SMS integration is key, but the integration should go beyond just getting updates via SMS. A user should be able to use the service without ever needing to go to a PC to: sign up, follow users, update their profile etc. The issue with this though, it may become costly very quickly and most of these start-ups are being bootstrapped.

4: Mobile integration

It goes without saying, the site needs a light weight and simple mobile website. The site should be able to be used on older wap browsers and low end phones. In addition to this there are other ways these local micro-blogging solutions could work and integrate better with the mobile networks. Most operators have some services which make use of USSD embedded on the sim card. The cost of requesting and sending information via USSD is usually 0, and the micro-blogging platforms could be integrated with this.

Can you think of any other ways micro-blogging services in Africa can differentiate themselves?

Some of the micro-blogging services in Africa

No comments yet

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS