We live in a lean era. In today’s economy, virtually every type of organization succeeds or fails in proportion to the speed at which it can learn and respond to change. Many NGOs have embraced this reality, joining the march toward steady, if not yet widespread, adoption of lean principles in the development of new social interventions. When it comes to scale, however, even the most innovative of these groups still routinely neglects the continuous innovation and learning methodologies that helped them develop scalable ideas in the first place.

We think this lapse is due to a pervasive notion in the NGO sector that the goal of continuous innovation and learning is simply to identify silver-bullet policy solutions faster, rather than to respond better to changing conditions over time.

Once promising ideas are identified, this notion holds that replication must rightly displace innovation.

We used to think this way, too. In the mid-2000s, we pioneered a four-year effort to reduce street homelessness in Times Square by connecting people to permanent housing and supportive services. We finally achieved an 87 percent reduction by doing four things: interviewing every person sleeping outside, mapping the places they slept each night, prioritizing the list according to social and medical risk factors, and reorienting local NGOs around this new approach.

In 2008, we took our successful methodology on the road throughout the United States. We expected that every place would be similar enough to adopt the Times Square model with minor tweaks. What we found was the opposite: The principles of the model held, but the details demanded new learning and continuous adaptation in every city.

Scale didn’t eliminate the need for innovation — it intensified it.

In 2010, we launched the 100,000 Homes Campaign to engage as many cities as possible in a national movement to find homes for 100,000 homeless Americans. One hundred eighty-six cities signed up, and we gave them explicit permission to incorporate new ideas, as long as they shared their learning with the whole movement. We also tracked every city’s housing performance centrally so we could test different strategies against each other.

This decentralized approach to scale would make some of today’s top NGOs and funders uncomfortable. It made some of our supporters uncomfortable. “You should only be scaling evidence-based practices,” many said. “It’s irresponsible to engage such a large network until your solutions can be longitudinally validated.”

The counterpoint, of course, is that few of the world’s problems have silver-bullet solutions, and even where such solutions exist, a continuous innovation and learning methodology can challenge core internal assumptions about minimum costs and viable target populations. In the 100,000 Homes Campaign, such challenges proved groundbreakingly fruitful time and time again.

In the end, the campaign also greatly benefitted from traditional research and policy thinking. We relied on these tools to inform a rigorous process of continuous innovation, but we never let them govern it outright. And it’s a good thing, because the 186 communities that signed up exceeded their collective goal, tripling the average speed of their work while finding homes for more than 105,000 of their most vulnerable homeless neighbors in just four years. One thing we do know from traditional research is that many of those people would have died by now if communities had not been willing to try new and unorthodox approaches to house them.

Since the close of the campaign, we have launched Zero: 2016, a new learning effort to help 71 U.S. communities end chronic and veteran homelessness outright in the next two years.

Many companies learned the lessons of continuous innovation long ago: The real-time process of evaluation and adaptation never stops, and it is never meant to stop. We want to help the NGO sector learn this lesson.

In addressing the world’s most urgent social problems, our goal should not be to figure out what works best and then simply reproduce it. Instead, it should be to figure out what works best today and to remain perpetually open to asking the same question all over again tomorrow.

This post first appeared on MediumPublication does not imply endorsement of views by the World Economic Forum.

To keep up with the Agenda subscribe to our weekly newsletter.

Author: Rosanne Haggerty and Becky Kanis Margiotta are co-founders of Community Solutions.

Image: A businessman avoids puddles at the International Financial Services Centre. REUTERS/Luke MacGregor.