A couple years back, a client came to me with a challenge that I was only partially able to help him meet at the time. A significant portion of his software-development projects carried a higher degree of technical complexity than is typical, and much of that technical complexity regularly presented his top-notch team with hurdles they had never faced previously. As a result, a continuous supply of innovative approaches was required to overcome these hurdles and deliver projects within plan. The central question he put to me was, “Can innovation be planned and executed reliably, and if so, how?”
In constructing my partial solution at the time, I drew upon my experience at NASA for how we were taught to manage such challenges:
1) Break the problem down into more manageable chunks (the basis of any solving any engineering problem).
2) Build more schedule and budget buffers into the project plan (again, pretty basic).
3) Build a more diverse, cross-disciplinary team, and facilitate environments that encourage innovative thinking both individually, and as a team (nothing new, but easier said than done; Agile/Scrum became part of the solution here).
4) Aggregate & balance the risk of “lumpy” innovations—that is, the tendency of innovations to show up on an unpredictable subset of projects, and at unpredictable times (a familiar portfolio-management challenge, but not a common PPM practice; CCPM became part of the solution here).
These are all perfectly sound approaches, and many successful R&D organizations have been applying them in a variety of ways for decades, so project portfolio managers can look to R&D portfolio-management practices for help. In the case of my client, he was already doing #1, was doing as much of #2 as his stakeholders could afford, and had made great leaps forward on #3 and #4. But while all of this helped, it wasn’t enough—his portfolio reliability was higher than the industry averages of 50-60%, but far short of the 90%+ target that CCPM has helped generate for many project portfolios.
Fostering Innovative Environments—Lessons From History, and From Natural History
I suspected that any further improvements would likely be found in the “fostering innovative environments” thrust of #3, but much of the advice here is based on anecdotal evidence, and is often conflicting. Then earlier this month, I stumbled across Steven Johnson’s illuminating book, Where Good Ideas Come From: The Natural History of Innovation (2010). Johnson has taken a close look at the history of the world’s major innovative advances, and has distilled seven common elements of environments highly conducive to innovation. Interestingly, all seven are also common elements of highly biodiverse environments in nature. Here’s a quick summary:
- The Adjacent Possible—breakthroughs aren’t really as “breakthrough” as we are often led to believe, but rather tend to consist of logical, progressive, incremental steps that open up new doors to explore new possibilities. For example, the creation of Youtube attained “adjacent possible” status only after some prerequisites had become established, such as standards-based web technologies, pervasive high-bandwidth connections, and Adobe’s Flash platform.
- Liquid Networks—just as water provides a rich, free-flowing environment for carbon-based molecules to mix and match in a multitude of ways that lead to expanding biodiversity, there must be a similarly “liquid” environment that allows building blocks of good ideas to combine in new ways. Interestingly, this suggests that secretive, proprietary R&D labs are less effective at spurring innovation than more open models.
- The Slow Hunch—most innovations are not the result of sudden “Eureka” moments, but the result of cultivating a hunch or kernel of an idea over months or years. While this suggests that breakthrough ideas generally need ample amounts of time, it also indicates that progress is far less random than simply waiting for inspiration to strike.
- Serendipity—many innovations have resulted from “fortuitous accidents,” such as the weak adhesive of Post-It Notes (the inventor was aiming for a strong adhesive). This suggests that, while serendipity by definition can’t be planned, the environment that fosters trial-and-error—and the exploitation of errors—can be very deliberate.
- Error—this logically follows from serendipity, but Johnson takes it a step further, asserting that there is an optimal amount of error that the most innovative environments must leave room for. This seems to fly in the face of reliability, which we’ve always been taught is maximized by eliminating as many errors as possible. But it turns out that leaving some small room for error may actually help make innovation more regular and predictable, just as genetic mutations drive a consistent pace of evolutionary change.
- Exaptation—referring to items that may have originally served one purpose, but which can be applied to serving another purpose in a completely different domain. A great example here is the giant screw press used to squeeze the juice from grapes for making wine, and which ended up being “exapted” by Gutenberg for the printing press.
- Platforms—specific constructs that open lots of new doors to the adjacent possible. Johnson uses coral reefs and rain forests as examples from the natural world here. What’s interesting about Johnson’s notion of platforms is that he draws particular attention to the balance of competition and cooperation required for such a platform to serve as an effective catalyst for innovation. For instance, a smart-phone platform encourages both competition and collaboration, and the result is thousands of innovative apps.
So, if we can replicate these seven characteristics to create ideal environments for innovation, we can expect more breakthroughs, and that’s pretty great. But the question remains: Will that higher volume of breakthroughs be predictable?
According to one major study that Johnson cites in his book, the amazing answer just might be “yes.” The study explored the relationship between innovation and city size, using a composite metric to gauge innovation, comprised of data points such as patents, R&D budgets, number of inventors, and number of “supercreative” professions. In city after city, a consistent pattern emerged—innovation was not only higher per capita in more populous cities, it was exponentially higher, and the study even pinpointed the exponent (1.2). This “superlinear” pattern means that a city of 1 million will be 17 times more innovative than a town of 100,000, while a city of 5 million will be 130 times more innovative.
Relevance to Project Portfolios
But how relevant is this to a project portfolio seeking a steady, predictable flow of innovation? After all, there aren’t many project portfolios with resource pools as big as entire cities. If Johnson is right, however, the predictability of the city numbers may well be relevant, because cities tend to be excellent examples of environments that exhibit all seven of his characteristics. In other words, if we can replicate the seven characteristics for our project portfolios, perhaps we can achieve a pattern of predictable innovations similar to a big city—without having to BE a big city, or even a big company.
A colleague (Koichi Ujigawa) offered an interesting example here—the rock tumbler. For those readers who may not be familiar, a rock tumbler can be smaller than a coffee can, into which one places a few dozen small rocks, a few tablespoons of sand or grit, and water. After tumbling this simple mixture overnight, you can reliably expect a canister full of nicely polished rocks by morning. I was always pretty amazed by this as a kid, but what amazes me even more now is that only two of Johnson’s seven characteristics are being replicated here—Liquid Networks, and Serendipity. In other words, in the rock tumbler we have created an environment in which the grit flows easily throughout the liquid medium, and in which the rocks are colliding in millions of random—yet patterned—ways with each other. The “liquid grit” caught in the middle of all those collisions creates the accelerated polishing effect.
Here are some compelling takeaways for our project portfolios from the rock tumbler example:
1) While the concept borrows heavily from nature’s example of how an ocean environment polishes rocks, the replicated environment can be much smaller in scale and still achieve the desired result—as long as the liquid network still has hundreds of thousands or millions of objects flowing around and colliding.
2) The desired result is achieved within a predictable time window. The tumbling action in the contained environment of the tumbler serves to boost dramatically the volume of serendipitous collisions. The higher the volume of serendipitous collisions, the faster and more predictable the desired result. We may not be able to predict the exact sizes and shapes of each rock in the end, but we have a very good idea of what results to expect for the collection as a whole.
3) If we can create a liquid environment containing the right rocks (ideas), the right granularity of sand (to help refine those ideas as they collide), and lots of tumbling action (lots of serendipitous collisions), we can expect a consistent result within a predictable duration.
The Scale Challenge
So, while the small scale of the rock-tumbler example demonstrates that “the scale challenge” can be overcome, the “big city” study tells us that a larger scale drives exponentially better results. So as project-portfolio executives we must ask ourselves: Which of the seven characteristics can we replicate at the scale of a typical project-portfolio resource pool, and which require the scale of a large city? It seems that just one must have some scale in order to spur innovation: Liquid Networks. While the Adjacent Possible, Serendipity, and Platforms clearly benefit from scale, it’s Liquid Networks that seem to help those characteristics work their magic. For example, how would an innovator even be aware of newly emerging Adjacent Possibilities without Liquid Networks? How often would Serendipitous events occur with no Liquid Networks? Even a small rock tumbler must have a very active liquid network with perhaps millions of agents at work.
Following this line of reasoning, if we can create an environment with the other six characteristics for our project portfolios, how might we secure access to the large-scale Liquid Networks that seem so essential for achieving exponentially high results? The Internet definitely helps make a wide variety of networks more liquid, but to what degree do Internet-based communities resemble the Liquid Networks of a city? Even more intriguing, are there aspects of Internet-based Liquid Networks that might actually exceed the scale advantage of a city?
Encouraging Anecdotal Evidence
We don’t yet have conclusive answers to these questions, but we do have a growing body of encouraging anecdotal evidence. Examples abound in open-source software-development networks, crowd-source initiatives in a wide range of industries, and with the many successes of highly diverse, geographically distributed teams. I can offer my own anecdote here as well—in coming up with one new idea recently, I discovered through a combination of personal networks and Internet searches that two other individuals (Koichi Ujigawa in Japan, and Wolfram Müller in Germany) had already developed similar ideas; by connecting with them and bouncing ideas off each other, we were able to generate a breakthrough technique almost immediately. No two individuals involved in this Liquid Network example were physically located in the same city, and in fact, the closest two individuals were more than 3,000 miles from each other.
This suggests that relatively low-population environments such as our project portfolios may be able to experience a high volume of reliable innovations just like a big city. As long as an organization can replicate six of the seven characteristics for its project-portfolio environment, and as long as this organization does so while immersed in large-scale Liquid Networks, it can expect the pace of innovation to be both higher and more predictable. As a result, any project portfolio wrestling with high technical complexity and the regular appearance of unfamiliar hurdles may finally have its long-sought-after “holy grail.”
I’ll be testing out this approach on any willing clients facing such challenges, and will report out results as soon as I have a meaningful set. In addition, I will share my ideas for how to replicate “the other six” characteristics in my next blog posts. In the meantime, I welcome you to share your own insights and experiences on this issue!