I think I’ve made some allusions over the past several months to my not-so-new-anymore gig at Grameen Foundation, but I’ve never really laid out what it is that I’m working on… and now seems like as good a time as any to do that.
When I moved back to Seattle after spending most of last year in Rwanda, my nominal job was to continue working on the Village Phone project with a focus on the ongoing effort in Rwanda and new replication efforts in other (likely African) countries. Then, around the first of March, I switched roles completely to take over management of our Mifos project. Mifos is sort of an acronym-like moniker for the initiative: Microfinance Open Source. The basic idea is fairly simple. Microfinance institutions have a bunch of information about their clients and loan portfolios that they need to track, and it’s a lot easier to do that with an automated computer system than to track the information on paper or, as they grow beyond a few thousand clients, an Excel spreadsheet.
There are a number of commercial software packages on the market that attempt to address this need, and some of them aren’t too bad. The difficulty, though, is that it’s a weird market to try to address. The practice of microfinance across the industry varies pretty dramatically. Microfinance institutions (MFIs) are often lacking in sophisticated IT skills on staff, and they don’t have boatloads of money to spend (sometimes donors give them big checks to spend on IT systems, but to be honest this is often part of the problem rather than the solution).
The overall market isn’t huge - there are anywhere from 3000 to 10,000 microfinance institutions around the planet (depending on how you count) but the majority are very small operations with limited money. This means that the total revenue available to systems vendors is fairly constrained… and in turn, that means that the business model probably isn’t there to build a highly structured, proprietary support structure that is able to provide support to the majority of MFIs in their language and near to their own time zone. Instead, it means that a vendor in Honduras might have a bunch of Spanish- and English-speaking staff trying to support an MFI deploying their product in Tunisia… many hours ahead and mainly conversant in French and Arabic.
As a result, something on the order of 90% of MFIs aren’t using commercial off the shelf software packages to manage their business. Of that 90%, about half are using manual (paper) and Excel-based systems. The remainder - nearly 45% of the industry - are using custom-built systems that are either developed in-house or by (highly) paid consultants. What this often means: mediocre systems (not all, but many - these are often built with limited software engineering talent in the house) that are expensive to build and expensive to maintain. They give the MFI flexibility - they can always just customize it further - but this flexibility comes at a cost to both the MFI and to the industry. The MFI pays a lot for software that may or may not do what they need, and pay even more for the ongoing upkeep of the system.
The industry, though, is really in a hole. Even with the need for flexibility to adapt to a wide variety of methodologies, the core functionality of microfinance systems is reasonably similar - probably 80% or more of the functionality that’s being built on a custom basis is the same. This means that as an industry we’re spending enormous sums of money building the same functionality over and over again. This, friends, is dumb.
What the microfinance industry needs is an open, shared software platform that can provide both the core functionality to manage portfolios, clients, and other key information and the extensibility to adapt to regional needs, variations in methodology, and new operating concepts and needs. Colleagues at Grameen Foundation started several years ago to see a pattern here, and eventually we landed on the following key needs:
- Flexibility to adapt to local requirements and conditions
- Reduced redundancy of effort across the industry
- Capacity to broadly share innovation across the industry
- Local (where the MFIs are) IT capacity
Enter Mifos. The general Mifos concept (originally called MOAP: Microfinance Open Architecture Platform) was developed by James Dailey (now focused on architecture and data standards on the Mifos team) several years ago and has since evolved into a broad initiative at the Grameen Technology Center. At its core, Mifos is an open source software platform (you can see the current build in action on the test server) that MFIs can use to manage their loan and client portfolios. The functionality in Mifos will grow over time to include support for things like micro-insurance products, surveys that can be used to gather and track information about how microfinance clients are moving out of poverty, and integration of data gathered using handheld devices in the field.
The Mifos initiative includes the software platform at its heart, but to address the needs identified above we need to go far beyond the software to build an ecosystem of microfinance institutions, technology providers, volunteers, and industry networks that will use, build, evolve, distribute, and support the Mifos platform.
In other words: lots of moving parts.
There’s lots more to it, suffice to say, but that’s the basic idea. Mifos is available on the web (don’t try installing at home - our deployment story still needs a bit o’ evolution), anyone can contribute to the code, and there’s nascent community building up around it. We’re working to develop a Mifos Specialist channel of local IT providers who can customize and deploy Mifos for MFIs in their region, and plan to use the platform and initiative to drive data standards efforts across the industry. Our first pilot, at Grameen Koota in Bangalore, India, has Mifos installed in one of their branches and is doing field testing (probably as I write).
We’ve got tons of work ahead - refining the architecture, nailing the business model, building in cutting-edge functionality, finding more users, building the open source community… not to mention fundraising - but it’s a great initiative and I’ve got a killer team working on solving some really hard problems. My job is to manage the overall initiative and team, direct traffic, whack moles, find money for the team to spend, and generally keep the train moving. I won’t be doing this same thing five years from now, but at the moment it’s a damn fine fit.
If you’d like more information, check out http://www.grameenfoundation.org/technology_center/mifos/, or for the technical types who want to see the code http://mifos.dev.java.net.
