The Emerging Science of “Doing” Development

Photo by Tom Chen on Unsplash

There’s a standard formula governments and non-profits use to bring about sustainable development: Experts research the problem, analyse the context, and prescribe a set course of action. Practitioners advocate for its implementation. Some years later they check if the solution is helping those it intended to. It’s a slow, uncertain, and often inefficient process .

Now, a breed of academics, policymakers and non-profit leaders are challenging it with a compelling alternative designed for today’s fast-paced, data-rich, and resource-constrained social sector.

It’s an approach that relies on experimentation more than planning, and on real-time learning rather than on scholarly research.

“The Lean Start Up” by Eric Ries, a book I read last year is the building block for this new approach. I believe Ries’s methodology will profoundly influence the non-profit sector. This article explains why.

The Pressure to Measure Impact

Data collection in developing countries is no longer an expensive, time-consuming, and risky affair.

Advances in technology put reliable and cost-effective data at our fingertips. This convenience comes with mounting pressure for non-profits to measure their impact.

Concrete targets embodied in the UN Sustainable Development Goals (SDGs), coupled with tight national budgets and calls for government accountability amplify demands to demonstrate impact — in a rigorous and quantifiable manner.

So, in 2018 I spent some time learning about how to measure impact. Along the way, I discovered a simpler and far more compelling notion: Build and nurture impact before you try to measure it. This sounds obvious but if like me, you’ve spent many years implementing social change or development projects, you’ll know that designing projects and solutions for serious impact is easier said than done.

The Challenge of Designing for Impact

Why? Well for one, seldom are development projects funded by the people they serve. Unlike in the for-profit sector, we can’t guage customer satisfaction using sales and profit figures. This absence of a water-tight, real-time yardstick makes it impossible to figure out which non-profit services or products are performing well, and which ones aren’t. In fact, due to resource constraints and because some solutions only deliver benefits after several years, most non-profits wait 2 to 5 years before assessing whether their projects are working. By the time the results are known, it’s often too late to course-correct ongoing efforts.

What can be done? Can the non-profit sector overcome such intractable problems? Yes. By gearing governments, donors and non-profits up to being more responsive and less ideological. The good news is we now have a rigorous scientific methodology for doing this.

Why a Rigourous Scientific Approach to Building Impact?

Despite the term “startup”, your organisation doesn’t need to be young, tech-savvy, or profit-oriented to adopt lean principles. In fact, this method is for any organisation dedicated to building something new under conditions of extreme uncertainty. In the world of sustainable development, almost everything we do falls under this definition. We face uncertainty from lack of data, conflicting data, changing political and economic realities, and because everything we do is about human beings. There is huge variance in the way that people respond to new ideas; especially when they’re living in tough and unpredictable circumstances. Will farmers adopt a new technology? How can workers be suitably re-skilled for the automation age? What’s the best way of delivering basic healthcare to remote communities?

Photo by Alex Boyd on Unsplash

Despite the huge unknowns that come with answering these questions, development projects address such issues using methods designed for conditions of moderate to high certainty. Typically, projects have a 3 to 5 year horizon with clear objectives and targets, and an accountability system to safeguard achievement of those set targets. As a result, most don’t fully deliver on their primary goal of helping the people they intended to. By using Lean Start Up, organisations working in the social sector can become far more impactful. A recent article in the SSIR argues the same.

The basic idea behind the methodology is that succeeding with an idea is less about the idea and more about discipline in implementation.

An Eternal Build — Measure — Learn Loop

Operating under conditions of uncertainty is the foundation for model’s core principle: Validated Learning.

“If you’re not 100% sure your beneficiaries will value and use your new idea, or you’re not even sure who your beneficiaries are, or if your product or service will grow and become sustainable (conditions are uncertain), then your main job is to learn these things ASAP.”

But this isn’t a fancy academic exercise culminating in ‘interesting insights’. No — Validated learning is about relentlessly confirming (or refuting) key assumptions about your business model using real people and a real product or service. Everything you do should then be based on the results of this (in)validation.

The first step in Validated Learning is to write down your “Leap of Faith Assumptions”. These are the key assumptions (hypotheses) that you believe are true, and that the success of your idea depends on. If you’re right, success beckons. If you’re wrong, you need to re-think. The key is to prove or disprove your assumptions as quickly as possible.

Writing down Leap of Faith Assumptions is particularly powerful because non-profits can get bogged down in the technicalities (think about how long it takes to develop a Theory of Change)— losing sight of the most critical aspects of a project. So for example, rather than spending weeks developing a Theory of Change for a magazine aimed at getting farmers to adopt organic farming practices, we would have the following Leap of Faith Assumptions:

We believe reading a monthly magazine on sustainable agricultural practices will:

  1. Persuade farmers to switch to these using these methods instead of conventional practices that harm the environment;
  2. Improve farmers knowledge and expertise on sustainable agriculture

The next step is to start running rigourous experiments (read real farmers and people)to understand whether this idea plays out in the real world. You might call it a pilot. But the key here is the focus on validated learning through a formal and perpetual Build-Measure-Learn loop.

Build a minimum-viable-product (MVP) — a version of a new product that allows a team to collect maximum amount of validated learning with the least effort. Here Lean Start Up differs considerably to how we think of Pilots — it advocates a MVP that is far from perfect because you’re not building for mainstream customer use. Rather, you’re trying to create an experience for early-adopters to prove a fundamental hypothesis. At the same time, MVP is the first of many micro steps to building the ultimate successful product.

You now need to Measure whether your MVP (or other experiment) supports your leap of faith assumptions. We do this through Actionable Metrics, setting an expected result upfront. An Actionable metric is one that can show true cause-and-effect of your MVP (Split or A/B testing is one technique that can be used to show causality).The Actionable metric is measured before (the baseline) and after pushing out your MVP.

Pivot or Persevere?

Of course, the process of learning is only meaningful if it drives decisions and gets you closer to the ideal solution. Actionable metrics allow you to learn and make determinations objectively. If things went as expected, then persevere: go through the Build-Measure-Learn loop again, this time further fine-tuning and improving the MVP (and getting one more micro-step closer to an ultimate solution)

If not, you now know an assumption is flawed, and you need to pivot. This doesn’t mean throwing everything to date away. A pivot is an informed decision to change direction. They can be minor adjustments or major overhauls — a different revenue model or customer segment for instance. Critically, it involves re-stating your leap of faith assumptions.

Executed properly, the BML loop enables you to Learn facts (validated by real customers) and Build a sustainable product at the same time — progressively taking many micro-steps towards the ultimate successful solution.

BML measures innovation efforts. By formally and repeatedly documenting assumptions, actionable metrics and pivot or persevere decisions, we can assess plainly whether progress is forthcoming. Everything is recorded for all to see. It sets a disciplined, rather than “let’s build it and see” approach to innovation. It takes politics and emotion out of decision-making.

Many of the components of the Lean Startup method (MVPs, scientific experimentation, cause-and-effect metrics, split testing, pivoting) are not new. It’s the wholeness of the entire framework and the interplay of these components that makes it unique, effective and powerful.

With 2030 just around the corner, we need new ways to move the needle on many of the SDGs. Waiting 2–5 years to figure out what where we went wrong is not an option!