Monday, 10 December 2012

Failed ICT development projects: Sweeping it under the carpet and moving on?

by Inka Barnett

The use of information and communications technologies (ICTs) has become increasingly widespread. Even in remote villages in developing countries there are more and more people who have access to a mobile phone. ICTs have the potential to make development projects more efficient, lower costs and improve the quality of service delivery. Therefore, it does not come as a surprise that the development community and national governments have enthusiastically embarked on ever more ICT projects in health, agriculture, e-governance, education and many more. There are certainly a number of very successful projects (see for example m-pesa). However, when I speak to development practitioners, I get the impression that there is an equal (or perhaps considerable higher?) number of failed projects. Unfortunately, only very few share their experiences of a failed ICT project publicly.

Last week I attended a talk by Ben Taylor from Daraja, a NGO in Tanzania, in which he bravely presented the lessons learned from his failed ICT project. Ben and his colleagues founded the Maji Matone (water drops) programme in rural Tanzania in 2010. The aim of their programme was to encourage citizens to put pressure on their local authorities to maintain and repair broken-down water pumps by using mobile phones. Using a simple SMS-message local communities were asked to report on the state of their water supply to the authorities. Local radio stations were simultaneously informed and followed-up the action the local water authorities would take in response to the text message.
The programme received a lot of attention nationally as well as internationally before it had even started. Unfortunately, the anticipated success did not come after the initial pilot phase of the project. The team had anticipated more than 3,000 text messages but received only 53! After Ben and his colleagues overcame their disappointment, they decided to actively investigate what went wrong.

They found the following reasons for failure:
  1. Political reasons: The relationship between local communities and authorities is sensitively balanced in Tanzania and citizens are reluctant to report on their government.
  2. Gender-specific reasons: Water collection is generally the responsibility of women and children who often do not have access to a mobile phone
  3. Lack of electricity and limited mobile network coverage
Ben and his team decided to openly share the reasons for the failure of their project in talks, on the web (including social media such as Facebook and Twitter) and in leaflets.
Openly admitting failure is a relative new but very important development in international development. The annual failure report by the Canadian NGO engineers without borders is another admirable example.

The application of ICTs in development projects is still novel and there are a large number of new and additional variables that need to be considered in comparison to traditional development projects (e.g. challenges of private/public partnerships, development of sustainable business models, negotiations of complex intersections between technology and development). Given the current often slightly uncritical excitement about the potential of ICTs for development, expectations of donors and limited funding, reporting failure is a challenging subject. However, without through investigation of why and how an ICT development project failed and without honestly sharing these experiences, we will continue to repeat the same mistakes over and over again.


  1. Yes, I think this is a really important point. There is always a buzz about the potential of new technologies, and it is really important that the opportunities these present are exploited. However, as with all development it is important to be clear about what you are trying to achieve. New technology should be the servant and not the master of development goals. I attended a Communication for Development ( network event on mobile technology hosted by GSMA - the organisation that represents the interests of mobile operators worldwide. One of the panel speakers, Becky Faith, from Tatical Technology highlighted the importance of considering security of users, particularly when using mobile technology for advocacy purposes. Governments can now monitor mobile traffic, for example during protests, and get personal information about users from the mobile companies. Five governments had forced companies to hand over blackberry user information and 35 African governments now require people to register to get a SIM card.

    There were stories about really effective use of mobile technology, alongside cautionary tales where users (beneficiaries) were not sufficiently taken into account in the execution of the projects, which inevitably failed.

  2. Inka, I think you are right on target with your impression that "there is an equal (or perhaps considerable higher?) number of failed projects" in ICT4D (ICT for development). The fact that the successful example you cite, M-Pesa, is not in fact an ICT4D project at all, but rather the commercial product of a huge corporation, only drives home the point.

    In fact, ICT4D has been trying and mostly failing to create sustained, scaled technologies for decades, mostly by ignoring the most obvious and important factors to create sustainability and scale: utility and cost. While the mobile phone itself has managed to harness those two factors to become the most important ICT since the printing press.

    In terms of utility, funders typically prefer very specific projects; always preferring, for example, a system that lets farmers get corn prices by SMS to a system that lets anyone get any market price by SMS. You end up with a situation in which decision makers in DC and London are trying to figure out the most important needs of an Eritrean farmer -- and most of the time, not surprisingly, they fail to do very well.

    On the other hand, the mobile phone is designed to let the USER (not the mobile carrier or the phone manufacturer) choose what to use it for. It can be used to communicate crop prices, or news about the family, or sports scores. And it is the person on the ground in the country that decides.

    In terms of cost, there are now mobile phones that cost only $10-20 new, and which (of course) don't require any tech consultants to set up or use. Contrast that with the typical ICT4D project, often costing hundreds of thousands of dollars per implementation, and which uses "free" software that requires several full-time programmers and tech consultants every time it is modified or adjusted.

    Can you imagine having to call a programmer in every time you wanted to make a phone call? Of course that would not be economically viable: the mobile phone is a lesson in how to make technology that is actually affordable even to the poor. And while Safaricom, for example, employs hundreds of programmers and technicians, that is for the whole region of East Africa. The typical ICT4D project requires at least a few programmers for a project with just 500 beneficiaries!

    ICT4D is not without a few successes. Consider FrontlineSMS (on whose board I sit): most implementations require a laptop and a mobile phone and someone with pretty basic computer skills, and the software can be used to send messages about anything the user wants (not just what someone in Geneva or Tokyo wants). Or our own EpiSurveyor mobile data software (to be rechristened Magpi next month): 99% of implementations require no programmers or tech consultants and cost just about $0, and can be used to collect whatever data the people on the ground think is important. Not coincidentally, EpiSurveyor is now used by more than 11,000 people in more than 170 countries -- the most widely scaled ICT4D project ever, and one of the very few that is entirely supported by revenue from the users themselves.

    I think it's going to be a great day when every funder, on being presented with a new project proposal, says "First you need to show me how, when this startup money has been spent, we'll be left with a technology that -- like a cell phone or like Gmail -- can be repurposed and reused by the beneficiaries themselves, with no programmers, and at a cost that is affordable to those beneficiaries." Until that day, I'm afraid we'll be left trying to think of successful examples of ICT4D and mostly drawing a blank.

  3. Inka you are right that there are at least as many 'failed' ICT4D projects as successful ones and of course we tend to hear only about the positive stories. I think however the same could be said about development projects as a whole. The World Bank is candid enough to admit that the majority of its projects fail to meet there original aims. In the commercial world 9 out of 10 small businesses fail inside the first two years so we should not perhaps be hypercritical. The important thing is to learn from our mistakes. FailFaires are one way that the ICT4D sector reflects on failure in order to inform future practice.

  4. Inka it is probably true that more ICT4D projects fail than succeed when measured against their original aims. This however is true for both development in general and for commercial innovations. The World Bank are honest enough to admit that the majority of their projects do not meet the originally specified objectives and 9 out of 10 small businesses fail in the first two years. Perhaps we should not judge innovation to harshly especially given the crisis in the old way of doing development. The important thing is to learn from our mistake and FailFaire is one way that the ICT4D sector attempts to make public its failures in order to inform future practice.


Thank you for your comment. Our policy is to review the comments before publishing in case of spam or abuse.