Using Facebook’s ‘Move Fast’ Approach at an International NGO

Facebook’s motto used to be “move fast and break things.” But during a recent interview, Mark Zuckerberg updated it to something slightly more refined: “Move fast with a stable infrastructure.”

At any given point in time, Facebook engineers are testing thousands of different versions of the platform on users and measuring those versions’ performances against Facebook’s key metrics. If a new idea works, they scale it. If it doesn’t, they add it to the company’s list of lessons learned. “The strategy of Facebook is to learn as quickly as possible what our community wants us to do,” Zuckerberg said.

Image c/o Maurizio Pesce/Flickr

Zuckerberg does a great job laying out why Facebook tries to learn quickly and how they go about it. This jumped out at me because I’ve also been thinking about how to do this better at the international NGO I work for. What does “move fast with a stable infrastructure” look like for us?

Below are some thoughts about how to develop the infrastructure for learning quickly at an international NGO. But first, why should NGOs focus on learning quickly? Why does it matter? As simple as it sounds, the answer is in the phrase itself.

Learning means turning data into information and knowledge, figuring out what’s working, what’s not, and why. Unfortunately, while many NGOs and their donors talk about it, it’s not always put into practice.

Quickly matters because we are working with vulnerable people, and weeks, months, and years matter. If we can marginally improve an intervention in a matter of weeks, and keep doing that month after month, we will drastically improve the quality of the program within a couple years. To put it simply, if you made a program just 1% better every month through a “learning quickly” approach, there’d be an 80% improvement over the course of five years.

So where do we start? Learning quickly comes down to building the right culture and systems.

Culture

To experiment and learn, an organization must be willing to fail, and to learn from that failure. But the more important cultural issue might be how an organization views the people they’re trying to serve. Most organizations talk about “beneficiaries.” While it’s true that we hope they’re benefiting, I propose we call them “clients.” (Hat tip to Scott Todd and the work he’s doing around this at the Initiative for Integrated Philanthropy).

Once we start thinking of them as clients, or even customers, we intrinsically place a higher value on what they’re saying. International NGOs should actively listen to their customers, just as businesses do.

Systems

As the right culture is being developed, encourage different teams in your NGO to test new ideas. These tests should be initially done at a small scale, supported by a testing framework. I believe NGOs can create this framework, or “stable infrastructure,” as Zuckerberg puts it, by following these steps:

  1. Set clear goals. You have to know what you’re trying to achieve before you an test better ways to get there. If you don’t have a Theory of Change as an organization, start there.
  2. Have a “playbook” that outlines your NGO’s “base product.” Facebook knows what its base product is at all times by looking at the code. It’s not as simple in international development work, but there should be some documentation of a base product that is clear across the organization. If that’s not possible at a global level, it should at least be done at a national or sub-national level.
  3. Determine what data to collect. What are the “key performance indicators” you’ll measure? In Facebook’s case, these might be pretty obvious: How many times posts are shared, how many friend requests are sent, etc. In the social sector, it can be a bit fuzzier, but there should be a handful of key metrics you use in all of your testing.
  4. Figure out the best technology to use. Facebook has the advantage of being able to collect data automatically without users even knowing it. In the social sector, we often have to ask people questions to get data. So figure out how you’ll do that in a way that doesn’t require a lot of money or a lot of staff capacity. (I like the lean data approach being promoted by Acumen).
  5. Spend time analyzing and documenting insights. Too often evaluation reports get completed and then just sit on a shelf. Make sure you are creating space in your project timeline to review the data and collectively determine insights. If the new approach worked well, figure out how to incorporate it into your “playbook” and scale it up. If it failed, try to understand why and put it in a lessons learned document.

Even though writing software code for a social network and designing interventions to help the world’s poorest people are two very different things, the process for “building a better product” is remarkably similar. In a nutshell, it starts by building the culture of learning quickly. You have to be clear about your organization’s goals, allow people to innovate, provide a testing framework, and be willing to scale ideas that work. It’s as simple—and as hard—as that.