Delivering Ethiopian teff via Czech taxi?

Forgive me, gentle readers. I’m on the road again, in Budapest for the next few days – hence, the sparse blogging. In a desperate bid to get myself off the road for much of the next few months, I’m traveling like mad between now and mid-November. One of those trips is to Pop!Tech, which means you can expect to be blanketed with posts for at least a few days late this coming week.

The reason for trying to get off the road is that I’m discovering that my brain is full. It appears to be taking me longer and longer to process the meetings I attend, the new projects I see, the talks I hear at conferences. So two weeks after returning from Rome, I find myself processing a conversation from the Web2 for Dev conference. I participated in a very strange panel on eAgriculture, where the main topic of conversation seemed to be the fact that none of the panelists quite knew what eAgriculture was or should be.

In the room, but not on the panel, was my friend Mark Davies, who might have had the best answer to that question. He’s the instigator of TradeNet, a promising system for sharing pricing information on agricultural information in West Africa, and enabling online trades between farmers and merchants.

It’s the sort of idea that’s so clearly in the right direction that it’s hardly a surprise that other groups are working on it as well. Speaking at TED Global, Eleni Gabre-Madhin outlined a hugely complex problem and proposed solution – the difficulties with establishing a national market for commodities in Ethiopia, and her proposed solution, the Ethiopia Commodity Exchange. In this, Gabre-Madhin is inspired by the Chicago board of trade – in a recent Economist article, she points out that the gangsterism that plagues the Ethiopian commodities market parallels the situation in Chicago in 1848, when the first mercantile exchanges, standards boards and futures trading systems came into play.

Gabre-Madhin’s system is supposed to launch this December. In the meantime, friends at Feedelix, a company that’s built an SMS client that supports Amharic characters, are pioneering a system called BoonaNet, which will provide pricing information on commodities like teff, coffee and berbere in the Amharic, Tigrigna, and Afaan Oromo languages. There’s a very rich powerpoint online showing the success and challenges to the system, which range from technical challenges (like system uptime) to participation of partner farmers and collectives. In some ways, Ethiopia is one of the hardest countries to do this in, as SMS was blocked by the sole mobile phone operator for two years for fear of SMS being used for mobilizing opposition movements.

In our discussions in Rome, one of the subjects that came up was the problem of providing key information to farmers through appropriate technology. In the case of farmers who aren’t literate, even localized SMS may not be the right method – broadcasting information over AM or FM radio might be a better method. It strikes me as surprising that there hasn’t been more work done making interactive voice response systems usable for development purposes. (We funded a project at OSI around this issue, but without much success.) Paul Meyer’s company Voxiva (disclosure: I’m an investor) builds IVR systems for health and microfinance applications – it would be great to see systems like TradeNet or BoonaNet using IVR to deliver key data driven by voice or keypad commands.

An example of how this might work comes to mind because I’m in Central Europe, the land of functional cab systems. A year ago, calling a cab service in Prague, I was stunned to discover a highly functional IVR system that let me choose a language to interact in, let me speak my destination, put me on hold for sixty seconds until a human operator matched by location to the nearest cab, and then gave me the number of the cab enroute to pick me up. When then cab was outside, the system phoned me again and told me the cab was outside. I never spoke to a human during the process, and it was a vastly superior user experience to calling for a cab in Cambridge, which usually involves being cursed at.

There’s a reason to think about using taxi IVR systems for development purposes: part of the challenge of building an agricultural trading system in a developing nation is working out transport issues. If a Ghanaian farmer wants to sell groundnut to a Togolese trader via TradeNet, he needs to put his load on a lorry. Using some sort of IVR system to find a truck making the trip – or to give drivers a chance to bid for the load, having their phones ring when the job becomes available – would be very cool. I suspect that adapting Czech cab-dispatching software isn’t the way to go… then again, the toolkits some companies are offering are pretty damned broad-reaching. (Does anyone know of an open source package that covers this space?)

I’ve got high hopes that Mark Davies, Eleni Gabre-Madhin, my Feedelix friends and others are talking so that they’re developing their tools cooperatively. Perhaps they should invite some IVR folks to the conversation as well.