Does it scale? That question is central to the discussion of social innovation, yet the answer to it might lead us to questions about why it is so important to us in the first place and answers that could surprise us.
“Does it scale?” or “how to we take [idea, product, service] to scale?” are commonly heard questions in social innovation circles; so much so that they are left unquestioned. The thinking behind these questions is that if something works well at one level (or scale) then taking it another scale larger and achieving a wider reach must be better. Who wouldn’t want to see the benefits of something that serves the needs of one population, community or user extended outward and upward?
This is a laudable utilitarian goal, but it is a deceptively problematic one when we look a little closer at what scaling something actually means in practice.
Jamer Hunt, the Director of the MFA program in Transdisciplinary Design at the New School in New York, speaking at last year’s DMI Fall Conference (which is available to view for DMI members), looked at the issue of design scaling through the lens of complexity and pointed to some of the problems with ‘scaling design’ in varied contexts. One of the examples he suggested is that of an ant compared with a human being taking a shower. For humans, the shower’s droplets of water are fine bodies of liquid that perform a particular task of facilitating cleaning, but for an ant those same droplets are enormous orbs of potential death. Water doesn’t scale the same for a human and an ant even though it is the same substance at both levels and the shower is identical in its structure.
In physics this is called scalar variance. What works ideally for humans is terrible for ants even though we are speaking of the same substance, same planet, same context. Water (most notably, a shower of it) doesn’t scale well in this case.
Yet, there is this insatiable desire among those working in social innovation to “scale things up” and “bring our innovations to scale” (even if we have little concept of how that would look or — as I will discuss — what that really means). The adherence to scaling as an ideology in social innovation (and applied social science in general) is bordering on “four legs good, two legs better” territory.
The Cult of Efficiency
International affairs scholar Janice Gross Stein attributes some of this fascination with scaling to a cult of efficiency, a political ideology that assumes that we can always rationalize human services optimally. What she found is that efficiency is used falsely as a stand-in for accountability, particularly in fields like education. Far from being against striving for optimal use of scarce resources, Stein nonetheless concludes that efficiency in human systems doesn’t always scale (my phrase, not hers) and that bigger and faster is often not better. Anyone who has taken a lecture with hundreds of others knows the difference of scale in learning between that and a seminar of five to ten people.
Taking Jamer Hunt’s argument: Bigger is just bigger…and whether its better or not is dependent on whether you’re an ant, a human and need to come into contact with water.
Designing for Systems and Scale: The Powers of 10
Designers and systems thinkers probably know the movie “The Powers of 10” by legendary designers Charles and Ray Eames. It’s a fascinating short film that looks at the universe moving out from a human being into the cosmos and inward towards what would now be quarks and everything in between. It is perhaps the best example of scaling ever produced. Beyond its educational and entertainment value, the Powers of 10 provide an illustrative example of where striving for scaling social innovations could be foolish and where it could have potential.
When traveling through the universe it is easy to see scales that are self-similar, thus they share properties that make them optimally relatable. These forms are often fractal in nature (thus, they share the same properties at different scales like that of a snowflake). Imperfectly, certain scales in the Powers of 10 are close to self-similarity where one scale looks and shows behaviour similar to those adjacent to it. These are spaces where it may be possible to transport an innovation from one to the other to good effect. Others scales look radically different from one another, suggesting a mis-fit in the scalar variance.
This is an idea, not an empirical point as we have little research on scalar variance in social innovation. Scaling innovation makes greater sense when the social systems have similar structures and ‘shapes’ and less when they do not. It is why in organizational science, certain models of management and decision making transport well from setting to setting and others do not. It’s why we’ve seen quality improvement processes like Six Sigma achieve great success in certain industries and firms and spectacularly fail in others.
Rather than adhere to an ideology that imposes scaling as a goal, social innovators need to generate the kinds of intelligence about the systems they are operating in (or seeking to operate or expand into) before making plans for scaling a promising intervention or product. As funders and policymakers this means setting performance targets that are appropriate or, perhaps better yet, working developmentally with innovators to co-create the outcomes of interest and the measures and metrics used to determine scalability and appropriateness early in the design and implementation cycle.
Without best evidence (which is almost always lacking in social innovation by its very nature), setting performance targets related to scale a priori is foolish. For innovators themselves, equally foolish is not gathering the kind of information about the systems they are operating in to know if they are the human or the ant and whether a shower is on the way.