Principle 2: Agile Development
Worldreader started with a simple goal—to fill the empty shelves of schoolrooms across Ghana with books. Until very recently, literacy programs to achieve such a goal ran like classic, hierarchical bureaucracies, with long-term plans determined in distant capitals, flowing out into the field to splash up against the messy reality of communities where illiteracy is widespread. Without even realizing it, Worldreader turned centuries of Plan-Fund-Do literacy projects on their heads. David Risher, Worldreader’s founder, said, “We were agile without even knowing it.”
Principle 1 replaces the elaborate assumptions implicit in the old Plan–Fund–Do model with educated guesses to be tested quickly. Agile development (Principle 2) replaces the Plan–Fund–Do method itself with a new way of building things. Agile production methods offer fast, iterative ways of building change. At more advanced levels, they offer a blueprint for how to generate guesses about your project that have the best chance of being right.
Worldreader’s experience started with direct experience with the problems of illiteracy and the absence of books in Ghana’s classrooms. The few books that were there were wildly out of date and out of context (A History of Utah was a title on one school’s bookshelf). You couldn’t make this stuff up, but it was probably the residue of some elaborate previous plan to help educate Ghanaians.
One of the ways agile development works is by creating a story about users, often called use cases. From there, you work backward to scenarios and archetypes—fleshed-out ideas about the place of your innovation in its broader context as well as multidimensional portraits of the people you are trying to reach or serve. In Risher’s case, he returned from Ghana with a clear vision of an entire generation leapfrogging paper books with digital ones, of getting the right tools and the right content to build literacy across the globe.
What he did next was the opposite of Plan–Fund–Do. He did not write up a long, detailed plan, and he did not seek funding. Instead, he asked Amazon for some Kindle tablets and smuggled them into Ghana preloaded with kid-friendly books like Curious George. The kids took to the books immediately and started looking for more culturally specific books, seeking new downloads.
Agile development is an ingredient to lean startups but a diverse field of practice in its own right with many streams and approaches useful to the social sector. What those diverse approaches have in common is that they can all be boiled down to one simple cycle: Build–Measure–Learn. Those three words capture the iterative nature of agile development. They are the way we make the product or service we are offering and how we test our hypotheses about the best ways to make it effective.
Agile development is the methodology that complements the customer-centered data gathering in customer development. As the name implies, agile development is also very much about speed. It is the method that can allow lean startup practitioners in the social sector to find the path to success quickly. Alternately, it’s the method that helps you figure out why you’re failing quickly enough to do something about it, or to revamp your approach as soon as possible.
The main imperative of agile development is to implement your innovation in ways that allow the fastest looping through the Build–Measure–Learn cycle. Contrast Risher’s approach to that used by the famed One-Laptop-per-Child (OLPC) project. OLPC was an idea, born at the MIT Media Lab, to mass produce a great laptop for low-income children around the world. While its main goal was similar (basic literacy), OLPC took more than six years of design and building before producing a viable machine. But by then it had become clear that cell phones were already well on their way to penetrating the same markets with essentially the same capabilities as a simple, basic, relatively indestructible laptop. Almost fifteen years since OLPC was first conceived, they’ve reached over 2 million children. Those are great results, but Risher and Worldreader surpassed that number in two years.
Figure 2.2 The Agile Way: Build–Measure–Learn
How fast is fast? To give you a sense of what “fast” in the agile world can mean, there are companies in the software business that release dozens of versions of their product every day. They take a little extra time up front and build diagnostics into the product itself so they can measure what each change has achieved in terms of product uptake, customer satisfaction, and delivery of essential functions right to the customer.
Most organizations, including nonprofits and government agencies, can’t meet this standard right out of the box. But Build–Measure–Learn is not an abstract concept. Organizations can set goals for the iterative process itself. Worldreader’s first few experiments took months: collecting Kindle tablets, smuggling them into different countries, experimenting with different content and ebooks. By steadily focusing on the rate of iteration itself, Worldreader can now prototype and test entry into a new community or serving a new type of user on a weekly basis.
As important as Build–Measure–Learn are the outputs at each step in that iterative process:
• The output of build is a minimum viable product (or MVP) based on feedback from customer development. It’s the smallest, leanest possible implementation of the original innovation that allows testing of the most basic guesses about how to make change happen.
• The output of measure is data, used to evaluate the experiment that you ran with the MVP.
• The output of learn is what lean practitioners call a pivot, a revision of the original MVP meant to retest the original hypothesis, to test a revised hypothesis, or, if the original hypothesis is spot on, to move the product or service to the next level in creating even more value.
In sum, agile development is the process by which you actually implement your innovation and iterate it until it’s explosively effective, or you give up. Agile is the toolkit for the experimentation that is key to lean startup.
To many in the social sector, the idea of trying multiple different approaches is scary. Innovation often has to happen in environments like jails or hospitals where there is already plenty of risk without adding an explicit focus on risky experimentation. And how are we supposed to fund multiple experiments in environments where people have barely enough to scrape by today? That’s where the third principle comes in.