By Chris Batchelder, Creative Director of Bon Education, with Mary Ames
In the programs that we design and build at Bon Education, we worry a lot about structure. How much structure is the right amount? What kind of structure should we adopt? Structure is, by definition, a very important part of any programmatic design. Without it, you can’t run a “program.”
Sometimes we try to control as many details as possible, creating a very carefully crafted experience for our program beneficiaries - perhaps an in-depth business internship or a quick activation at a conference booth. This kind of structure yields very predictable results (usually).
But something we also play around with is loosening the structure, creating structure through unstructuredness. How can we structure a learning experience by creating very unstructured parameters? How lightweight could our intervention be while still achieving a desired outcome?
Effective solutions that are highly unstructured are extremely attractive for many reasons. You can reduce costs. You can reduce time and energy of management. Your solution might be much more scalable. You might even be able to give more power to your direct beneficiaries.
We love reducing structure!
So when I was listening to the Planet Money podcast, and the hosts started talking about a highly effective program that used randomness as a key design principle, my ears pricked up.
In this blog post, we will explore the structure of randomness and how that structure can be used to reduce other structural components that might be undesirable for one reason or another. Let’s dive in!
Randomness is nothing new in the realm of scientific experimentation. A randomized controlled trial (RCT) “is a type of intervention study that aims to reduce certain sources of bias when testing the effectiveness of new treatments; this is accomplished by randomly allocating subjects to two or more groups, treating them differently, and then comparing them with respect to a measured response.” In the health sciences, randomized trials are the gold standard, used in the development of new treatments, medicines and vaccines. Each USFDA-approved COVID-19 vaccine underwent multiple randomized trials prior to being accepted for widespread use.
When it comes to learning, empowerment and outreach programs, however, selecting participants at random seems at odds with our core principles. Surely we’d want to know that our program participants are the ones both most in need of an intervention and most likely to succeed because of it - thereby not squandering precious resources? Structure means accountability, measurability and sustainability.
That’s certainly how I usually think about it.
The program I heard about on Planet Money is called YouWiN! YouWiN! is a business plan competition in Nigeria, where winners - selected at random - received up to US$64,000 in cash toward their small business or business idea.
The YouWiN! competition was designed in partnership with the Nigerian government, the Department for International Development and the World Bank to tackle a big economic puzzle: How do you get a small business to grow into a medium-sized business? The competition invited young Nigerian business owners to submit their business ideas in a bid to win up to $64,000 - no strings attached. But with nearly 50 million youth between the ages of 18 and 40, who do you give the money to?
The YouWiN! program used randomness as a key design principle. After an initial screening process, the top applicants were selected for a four-day business plan training course, and then winners were chosen to receive awards averaging US$50,000 each. In total, 24,000 Nigerians applied for the program and 1,200 received a cash prize. Crucially, the program designers created a “control group” to measure the effectiveness of the program by randomly selecting 729 of the 1,200 winners from a group of 1,841 semi-finalists. The results of the program speak for themselves:
“Three years after applying, new firm applicant winners were 37% more likely than the control group to be operating a business and 23% more likely to have a firm with 10 or more workers, while existing firm winners were 20% more likely to have survived, and 21% more likely to have a firm with 10 or more workers. Together the 1,200 winners are estimated to have generated 7,000 more jobs than the control group, are innovating more, and are earning higher sales and profits.”
Even though the winners were chosen largely at random, the results are undeniable. So, what can we learn about the structure of randomness from this case study?
The randomness built into the design of the YouWiN! program solved difficulties faced by many program designers, specifically the difficulties of being fair, knowing who is better than other people and working at scale. It may sound counterintuitive, but randomness has real benefits.
Randomness is fair.
We like to think that we are selecting the people most likely to benefit, and succeed, as a result of our programs. But applications and test performance are not reliable indicators of future success, and bias has a way of creeping in. Not only does randomization reduce the influence of bias, it affirms the narrative that anyone in the population group is capable of success. In a randomized program, everyone gets a fair chance. Strange as it seems, choosing program participants at random is very empowering!
Randomness is measurable.
As the YouWiN! program clearly demonstrates, embracing randomness does not mean abandoning structure. Using a Randomized Controlled Trial approach, program managers can measure meaningful data on the impact of their program against multiple metrics over time by tracking the performance of program participants against the randomized control group. In fact, data from randomized programs is more indicative of impact potential compared to selective programs.
Randomness makes scale possible.
In a randomized program, participants tend to be more representative of the general population compared to a selective program. Therefore, a randomized program that is showing success for a small group of participants can reasonably be assumed to continue to show success for a much larger group of participants. This evidence-backed generalization is one of the key benefits of randomized controlled trials. (Hey, if it worked for the COVID-19 vaccines, it can work for programs!)
Randomization also promotes efficiency, another key consideration of scale. Selecting program participants is simple enough when working within one organization or community. But when programs expand to a national or international scale, randomness, again, proves its value. Randomly selecting participants - or, in the case of YouWiN!, winners - not only spreads fairness, it also increases the overall efficiency of the program, with fewer resources spent on the selection process.
Certainly, I am not suggesting that randomness is a panacea to solve every problem related to running programs. But it is definitely worth considering as an important tool in every program manager’s toolkit.
The YouWiN! program began in 2011 and ran for eight years. Similar programs have launched in Kenya and Senegal. Perhaps the long-term success of the program isn’t so random after all?
Wishing you luck with your next random program!
Chris Batchelder is Creative Director of Bon Education and Co-Founder of interstory. He has designed and led programs for clients across the Middle East, Asia and North America. Chris is a dedicated, lifelong multiculturalist and art lover, and he resides in the UAE with his wife and three children.