Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
Ever since Michael Porter published The Competitive Advantage of Nations in 1990 Ministries of National Planning throughout the developing world have focused their national development strategies, at least partially, on creating and fostering sectoral “clusters.” Clusters are groups of related firms -‐ suppliers, buyers, competitors, and collaborators -‐ all in the same industry, and all in one place. The two most famous clusters in the world are likely the entertainment cluster based in Hollywood, California, and the technology cluster in Silicon Valley. Everyone generally agrees that clusters in and of themselves are a good thing, and generally for the reasons that Mr. Porter cites in his book. His central claim is that economic growth in a modern world, where natural comparative advantages in resources and labor can be rendered meaningless by globalization, is all about productivity; and the productivity of firms is dramatically improved by operating in a cluster. In his words, “clusters allow companies to operate more productively in sourcing inputs, accessing information, technology and needed institutions; coordinating with related companies; and measuring and motivating improvement.” Translated, this means that companies spend fewer resources looking for things they need: employees, supplies, and know-‐how, because they are all available to them within the same geographic area. Consequently, more of those resources go into production. Further-‐ more people in close proximity working to solve the same problems (i.e., how to streamline production or build a better product) means better solutions are continually developed and disseminated more quickly, and the outputs of everyone in the cluster improve. In economics this is called a “positive production externality,” and is one of the textbook justifications for government intervention in a market. So it makes sense that so many governments have decided they are going to intervene in their national markets to help create clusters. Further, they tend to follow Porter’s advice when picking clusters to support. He argues, “governments, working with the private sector, should reinforce and build on existing clusters, rather than attempt to create entirely new ones.” So when developing countries take stock of their existing economic assets they tend to emerge with strategies that focus on a handful of similar clusters: notably agriculture, tourism, and ICT. Why ICT? In most countries, even very poor ones, there is generally a small community of well-‐educated and computer-‐savvy young people. They may already be programming. So the “firms,” in this case individuals, already exist. A focus on ICT is a logical choice. But despite the reasonableness of the decision, and all the resources being poured into the creation of ICT clusters, very few, if any have taken off. Why? The problem comes when we translate strategy into tactics. Cluster theory focuses on contact as the mechanism through which productivity is improved. Porter writes: “tapping into
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
the competitively valuable assets within a cluster requires face-‐to-‐face contact, a sense of common interest, and ‘insider’ status. The mere collection of companies, suppliers and institutions creates the potential for economic value; it does not necessarily ensure its realization. To maximize the benefits of cluster involvement companies must participate actively.” So to encourage the formation of an ICT cluster a government should logically focus on interventions that bring entrepreneurs (the firms of the ICT world, especially the programming world) together. If a government would, for example, provide a space for ICT entrepreneurs and potential entrepreneurs to gather, collaborate, share ideas, and learn from each other, and if they designed it in a way that reclusive computer programmers would actually use, then they’d go a long way towards building a foundation for an ICT cluster. Once the entrepreneurs were all in one place collaborating, then technology companies that supply the things these entrepreneurs need, from hardware to training programs to legal services, would have a large and static market to cater to, and they could operate more efficiently. The presence of all these entities together would further increase everyone’s productivity, resulting in the virtuous circle described in the literature on clusters. This path leads directly to economic growth and job creation. This type of intervention is simply a logical extension of Mr. Porter’s theory, which governments have already bought into on a more strategic level. But it is not at all representative of the type of intervention governments and donors pursue when they attempt to create ICT clusters. In fact, governments (and donors) tend to pursue policies that actively encourage fragmentation of the market, rather than cooperation between firms. Take for example, the very popular intervention of setting up an ICT business incubator. This institution provides a very deep level of support to a very narrow group of individuals or firms. The idea is that due to limited resources the government/donor will provide training and resources to a small number of groups, and the knowledge and productivity they gain will spill over onto other entrepreneurs. The idea sounds good, but is in fact predicated upon a faulty assumption: that there exists a vibrant social and professional network of entrepreneurs and firms within the industry that will ensure these productivity spillovers take place. Such a network almost never exists. Spillover effects need transmission mechanisms, and to assume these mechanisms away will mean the government’s intended benefits won’t happen. Instead, they will subsidize training for a very small group of companies whom, when they’ve finished the training, don’t have a peer group. The training has given them tools, but left them isolated (fragmenting the domestic market). Firms are left out on their own to either try to innovate in a vacuum or to compete, by themselves, in a global marketplace that consists of firms operating out of other geographic clusters. These newly minted
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
firms are therefore born with a handicap, and cannot possibly lead a massive sectoral development push that governments and donors so want to see happen. Another example of a popular intervention is the “innovation center.” Due to the private sector’s eagerness to fund these centers (for reasons of capturing a market), and the government’s/donor’s eagerness to be seen to be playing nicely with the private sector, these innovation hubs are invariably owned and branded by a private technology company. If the goal, however, is to encourage open and strong networks of innovation and collaboration this privatization of innovation hubs cannot happen. If, for example, Microsoft operates one of these hubs it is highly unlikely that Nokia, Google, or any other firm is going to contribute resources to it. That will limit a number of things. First, it limits the output of the lab. Only people that use MS platforms can use the products produced within the lab. Second, it limits who will use the lab (it excludes programmers who are proficient in non-‐MS platforms) and creates segmentation within the developer industry. Third, because the group of entrepreneurs is limited, the opportunities for other firms to associate with that group are limited. In short – the center is too small to become a pole around which a cluster can emerge. Porter writes explicitly that the formation of clusters is predicated on very low barriers-‐to-‐entry, which corporate ownership of an “innovation center” explicitly prevents. So, it’s time for a re-‐think of our ICT cluster development strategy, and a great model has been provided by Kenya’s famed Ushahidi, who in March 2010 founded the “iHub”. Within six months the iHub had become the center of the ICT developer community in Nairobi, and tech firms were knocking down the door to try and access the market that its collection of high-‐power individuals provided. The iHub has more than two thousand members, and this collection of entrepreneurs, continuously throwing ideas around and collaborating with one another, has already led to the creation of companies, the expansion of existing companies, the creation of new and well-‐paying ICT developer job, and as a bonus, locally tailored ICT products designed to be used by governments and NGOs for development. Here is just one example of how the iHub works. Su Kahumbu-‐Stephanu, the founder of Kenya’s first locally certified organic farm (Green Dreams Ltd.), was struggling with how to manage the information required to maximize the economic output of her cows. There are optimal time periods for milking and calving, and due to the nature of cows there is an optimal 12-‐hour window every year when a female cow should be impregnated to ensure maximum profit to the farm. Su’s information management system was not robust enough to keep track of all the data needed to pin down that 12-‐hour period, so she came to the iHub looking for answers. There she met Charles Kithika, a developer from Nairobi that was looking to design a program that could accelerate his career. The two immediately got on, and together they developed a voice-‐based mobile phone application called “iCow.” The
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
organizers of the iHub helped Su and Charles enter the iCow application in the US State department’s first ever “Apps4Africa” competition, and they won. As a result, Su’s farm is more productive, she can afford to hire Charles as her full time CTO, and Charles was inspired to form a start-‐up (called “Nerd”) that he runs out of the iHub. Why did this story unfold the way it did? Because the iHub had become the center of a new ICT cluster in Nairobi. Su had to know where to go to find help in creating her application. Charles had to know where to go to look for a coding challenge that could make him successful. Technology companies had to know where to provide hardware and software platforms to reach their target market (developers like Charles). The US State Department had to know where to go to advertize their contest. All four instinctively knew to go to the iHub. And better yet, Charles, despite his newfound success, maintains ties to the iHub. He is frequently there, interacting with aspiring developers and making sure the channel that transmits “positive production externalities” stays open. And Charles is not alone. New success stories such as this one continue to pour out of the iHub nearly every day. In its eight months of operation to date, the iHub has arguably had more success in developing an ICT cluster than all the development initiatives that came before it. So why is the iHub so successful where other interventions have come up short? The key was that the iHub was created to be, and has remained, an independent, community-‐owned entity, and that everyone is welcome there. This includes all the big tech companies, but also every tech entrepreneur that had a project he or she wanted to work on. The companies came because the entrepreneurs came. The entrepreneurs came because the iHub is such a wired place. Through a connection with Kenyan ISP Zuku, the iHub provides entrepreneurs with 20mbps of free connectivity, connectivity that iHub users would either have to pay for at home, or couldn’t even access. Further, the iHub’s founders knew their audience, and simply created a cool place to hang out. There is a coffee bar, a great view, and a constant series of programming and events that is of great interest to the local ICT developer community. This includes training sessions on coding skills, discussions with big players in the tech industry, incubation services for new companies, film screenings, games and parties with like-‐minded people, and opportunities to compete in application development and business plan competitions. It is a big social club with a purpose, and it’s leading to a glut of economic growth and job creation. Open and innovative Local ICT Developer networks like the iHub (I’ll generically call them LID networks from now on) are starting to crop up in other places in Africa. These include the HiveCoLab in Uganda and Activ Spaces in Cameroon. But their development is slow. To create a successful LID network you need a very unique set of skills. First, you need an intimate connection to and understanding of the local tech community. Second, you need access to outside resources and opportunities. Third, you need motivation. The iHub’s founder, Erik Hersman, brought both of
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
those traits to the table through his own unique personal story. The son of American missionaries in Africa, he grew up in Sudan and Kenya and consequently feels African at heart. He went back to the US to start his career, got into the American technology community, and then came back to Africa and started blogging about technology there. His blogs gained a large following, and he got to know all the key players in Nairobi while maintaining his ties to the US. He became a TED fellow, and out of the “TED Global” conference in Arusha, Tanzania in 2007 (which occurred just before the outbreak of post-‐election violence in Kenya in December of that year) he co-‐founded Ushahidi, the world-‐renowned crowd-‐sourcing application. That gave him a pile of international recognition and local prestige, plus a small pot of money with which he and his colleagues could do as they pleased. He used all three to create the iHub. The first thing he did was to put together a leadership team for the iHub. In addition to Erik Hersman it included two of Nairobi’s best-‐known tech writers (one journalist and one blogger), as well as the founder of a group called “skunkworks,” a loosely affiliated group of ICT-‐industry people that met at regular intervals. It also included Kenyan ISP Zuku, who contributed all of the connectivity. So before the iHub opened its doors, Erik had arranged community buy-‐in and a superb level of connectivity. His issue was funding. He instinctively knew that the iHub was a good thing to do. He felt that, in his words, “if you put a bunch of smart people together in a room then good things happen.” But he couldn’t offer success metrics to donor agencies, and consequently they weren’t interested in funding it. He also knew he couldn’t brand the iHub and cede control of its image to any outside group: the neutrality of the space was too important. Consequently, he was not able to access any private sector funds either. After two-‐years of unsuccessful fundraising attempts he decided to throw in the towel and fund it with Ushahidi money. With that final hurdle cleared the iHub’s doors soon opened, and it was an immediate hit, so much so that donors are now lined up to fund the program and help it expand. For example, the iHub (and an affiliated consortium) just won a grant from the Finnish Government, via InfoDev, to host a mobile applications lab. This will include services to train mobile applications developers and to mentor them in business skills. (Note – this is likely to be successful, as it plans to use the iHub’s social and professional network as the transmission mechanism for its intended spillover effects). Additional funding comes from the members themselves. The iHub is so popular that it can charge more successful entrepreneurs a small fee for permanent desk and locker space. After 8 months of operation the iHub is on the verge of financial self-‐sustainability – an unbelievable success for an organization so young. Thinking about the iHub in Value Chain terms
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
Another way to think about why LID networks are effective, while traditional ICT sector development strategies have not been, is to think in terms of value chains (which is another term popularized by Michael Porter). In a value chain approach, a donor agency looks at every step along the line of production, where value is added to a product. This includes sourcing inputs, every step in production, distribution, marketing, etc. Bottlenecks are likely to appear in one or more places along the value chain, and targeted assistance to alleviate these inefficiencies will have very large effects on the chain as a whole. If we think about the products produced by software and mobile application developers, they have relatively simple value chains. The developer first needs an idea. Then he needs a few pieces of equipment, software, and Internet access to create it. Then he needs access to a group of people to help him develop the product, refine it for his intended audience, work out potential bugs, and perfect it. Once that is done he needs to commercialize: show the product to investors, get seed capital, run pilots, and market. So, a value chain approach to the software and mobile applications sector would likely focus on the one or two areas where the value chain is constrained. If we are to examine current interventions, one would think that the only problem the industry faces is in commercialization. That is what ICT incubators are for, and they dominate government’s strategies. If we think of the “division of innovation labor” as consisting of three phases, idea generation, development, and commercialization, the first two are largely ignored. Traditional tech incubators look for groups that already have a good product, and then provide them with a series of services designed to get that product out. We rarely see interventions designed to help potential entrepreneurs come up with ideas, or help people that have an idea turn it into a product. Yet the technology value chain is just as likely to face constraints in these areas as in the commercialization process. Why the focus on commercialization? Again, simply because it is the easy solution to fund. Business incubators produce a tangible output. Idea generation is either unsupportable, or privatized within a firm that will own the final product, artificially constraining the inputs (ideas) that will go into product development. Product development (going from idea to product) is underfunded even in advanced economies – it is simply too risky for most angel investors to get involved in. LID networks, by bringing people together is a shared social and professional group, will facilitate idea generation, and allow entrepreneurs to access a large peer group (and outside expert visitors) to help with product development. When products have been developed, the LID network can feed it into affiliated incubator for commercialization. In this way, LID networks correct an imbalance in value chain interventions, and make donor money more effective.
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
The point is simple – no matter the terms in which one discusses them, LID networks are effective ways of kick-‐starting growth in the ICT sector. The role for USAID The social experimentation that was involved in creating the iHub can pave the way for the more efficient creation of other LID networks. Whereas Erik Hersman wasn’t sure what would happen when the iHub opened, thanks to him follow-‐on groups now have a better idea of success metrics. The key will be to judge whether or not the network is becoming the center of a new cluster, i.e. does it have magnetism? How many people sign up to use the LID network? How many products does it produce? How well do its members do in international competitions? How many offers of funding or in-‐kind donations from technology companies does the LID network receive? We also have an idea of how quickly these networks can become self-‐sustaining, and consequently are more able to judge if the network is successful or not. This clearer idea of metrics should allow donors to get involved at an earlier stage, and prevent the two-‐year lag it took the iHub to go from idea to execution. Donors can play two roles in creating new LID networks. First, they can facilitate coordination between groups that come together to create the network. They can foster discussions between potential founders: bringing journalists, community leaders, ISPs, and technology companies around a single table and helping them work through any issues of vested interest. The idea will be to inspire the right people to lead an effort to coalesce existing entrepreneurs into a stronger social and professional network. This is something that Erik Hersman did himself in Kenya, but he is a unique individual with a specific background that may not exist in every market that could benefit from a LID network. Second, if a viable coalition of partners seems to be taking shape, donors could fund the basics of a network for a set period of time until they become self-‐sustaining. This would be mostly infrastructure: the physical space and its associated utilities. This type of funding should have a sunset clause, and should be contingent upon passing a rigorous Go-‐No Go test with the coalition as established. The sunset clause would likely kick in after two years. Two years of guaranteed funding would allow the leaders of the new LID network to ignore fundraising for the first year. They could instead concentrate simply on outreach, programming, and building the network into the center of the local ICT community. Then, during the second year, they would have a history of results to bring to their external fundraising pitches, and the LID network would require less day-‐to-‐day guidance from the founders. For example, much of the programming at the iHub is member-‐led. I recently attended
Local ICT Developer (LID) Networks A new understanding of ICT sector development
GBI Innovation Series Working Paper #3
by Eric White
an event called “Random Hacks of Kindness” at the iHub, where no iHub staff was present. If, after two years of operation the network was not able to finance itself donor money would be cut off and the network would dissipate. In this worst-‐case scenario, a valuable post-‐mortem would be produced, and the donor would not have spent much money to get it. That’s because neither of the two funding roles for donors would cost a lot. Hiring a small team to lead coordination and “inspiration” efforts, conduct outreach, and host meetings would amount to no more than $250,000. Funding two year’s worth of LID operations would cost about $500,000. All told, a donor agency could engage in this type of intervention for about three quarters of a million dollars, peanuts when compared to the millions and millions spent on development projects that have been far less productive than the iHub has proven LID networks can be. An effective cluster development strategy “Cluster Development” as a development strategy has a long history of mixed success. It can immediately bring to mind pictures of empty office parks and export zones. But, as I’ve tried to make clear, this is an indictment of tactics rather than of the strategy itself. Cluster development can in fact be successful (look no further than Bangalore, India for proof). The general consensus of the literature over the last twenty years is that support to existing clusters has a much greater chance of succeeding than do efforts designed to will their creation out of nothing. LID networks understand this. They are not so much about creating a cluster as they are about inspiring leaders to create stronger linkages between existing entrepreneurs. This behind-‐the-‐scenes work will pave the way for clusters to develop naturally. To conclude, Local ICT developer networks (or LID networks) such as the iHub are potentially revolutionary. They are cheap, effective, and in line with the principals that guided the formation of national development policy in the first place. It’s time for USAID to offer them its support.