I’ve written a new post on balancing exploration and execution for the Lively Work blog… and had the audacity to name a ratio after myself. Check it out here:
and for archival purposes, here is the text:
The S Ratio: Finding the Optimal Balance of Exploration vs Execution.
For the last decade or so, I’ve spent a fair bit of time thinking about optimal structures for institutions, organizations and societies.
My thinking in these areas has largely been skewed toward an emphasis on the factors that lead to a competitive advantage. My reason for focusing on that particular question is quite simple: structures that do not lead to a competitive advantage tend not to persist. The institutions that make use of non-advantageous structures lose access to resources, talent and opportunities – and eventually, they transform themselves or are replaced.
Over the long term, the organization that prevails will be one that can best sense and respond to the needs and opportunities in their ecosystem. This means that they both have the ability to perceive the world around them and have the ability to actually take action based on that information.
When looking at the structure of societies and even of biological ecosystems, diversity has been identified as a prime driver of competitive advantage. This is because, with diversity comes access to a wide variety of solutions – and this in turn increases the chance that a working solution will emerge.
However, though diversity — or disagreement — bestows an advantage over the longterm, there are occassions where the “more diverse” group gets their clock cleaned. A prime example of this occurs on the battle field. In such situations, coordinated execution tends to be of greater importance than diversity of ideas.
What does this mean for the structure of organizations?
Basically, it means that there is no one size fits all answer. The optimal structure of an organization is determined by the circumstances that group is faced with. For instance, an established enterprise in an industry that faces little change is going to find itself in possession of a considerable amount of wisdom about what works, what doesn’t and how to get things done in the real world.
Because their environment is relatively static, that wisdom will prove extremely useful in guiding decision making. These organizations tend to be fairly hierarchical in structure – and this hierarchy works for them. It helps them execute on the lessons that they have already learned, while minimizing the amount of effort and energy that they waste “re-learning” what someone else in the organization had already figured out.
On the other hand, a high-tech startup faces an entirely different set of circumstances. They are working on new technology, have little accumulated wisdom about what will and will not work, and may be faced with significant uncertainty for every area of their business – who to approach, how to position their product, what channels to sell through, what price to charge etc.
At first, their organization will be small and so lessons learned will be communicated easily across the different members of the group. A relatively flat structure that maximizes exploration and learning will give them a competitive advantage.
Of course, as the high-tech startup grows the circumstances that they face will change. Their knowledge of the market will mature. The size of their organization will increase. And the way that decisions get made and executed upon within their business may need to change.
At the same time, the established enterprise that has grown accustomed to delivering a specific value proposition in a particular environment and has optimized its structure in order to do so may find themselves vulnerable to disruption. An overemphasis on execution, accompanied by insufficient investment in exploration may leave them divorced from changes going on in the world around them and thus unable to react to those changes until it is too late.
For the last couple of years, I’ve been puzzling over the optimal balance of exploration versus execution in an organization. I propose that this balance can be expressed as a ratio. Since I’ve never named a ratio before, and I can’t think of an appropriate name at present, let’s call this optimal balance for a particular context the “Schutte ratio” or S.
The formula would look something like this:
S = exploration / execution
In general, this can be expressed as a fraction, though a percentage may at times make comparisons easier to interpret.
From this we can see that, in theory, S could be anywhere from 0 (0/100) to a number approaching infinity (or an undefined number with zero in the dividend).
In reality, the optimal balance for an entire organization never reaches 0/100 (zero exploration, all execution) and it never reaches the undefined state 100/0 (all exploration, no execution).
Let’s take a few examples and think through the optimal balance for each.
1) The App maker: in the midst of digital disruption
Lets start with an example that is focused on a software based startup – one where there is plenty of transformation happening within their ecosystem. New devices are arriving on the scene regularly (new phones, new wearable devices etc). New applications that they might be able to integrate with, support or that may become competitors are emerging constantly. In short, the environment that their company’s product exists within is changing shape – and fast. In an environment like that, the optimal amount of time, effort and money spent on exploration might be relatively high – perhaps 20%. 20% exploration = an S ratio of 25%. (20% exploration / 80% execution = 20/80 = 1/4 = 25%).
So let’s assume that S = 20/80 for this use case.
2) Railroad, Long History, Slow to Change On the other hand, let’s look at the S ratio for a capital intensive business like a railroad. Making changes to the physical locations that they serve is incredibly costly, possibly even requiring government intervention to gain access to land. There are certain areas of their business where innovation may be happening (ticketing, customer interaction etc.) but relative to the software startup, the railroad is 1) in a competitive arena that changes relatively slowly, 2) equipped with decades of experience in handling competition in the space. Too much effort spent “experimenting” will leave them stuck with high costs and will drive customers to the competition — whether in the form of trucking, shipping or air transportation. As a result, the railroad’s optimal balance is likely far lower – perhaps at 5/95 or 1/19 (S = 5.26%). Likely, it is even lower than that.
Nonetheless, let’s assume that S = 1/19 for this use case.
3) A portion of a business
The previous two examples looked at the S ratio for a business as a whole. But S ratios don’t tend to be uniform across the organization. Let’s take a moment to open one of them up and peer inside a bit. Let’s take the railroad example and examine two areas of their company: a) online reservations and b) track configuration.
a) Online reservations
In this portion of the railroad company, the ecosystem that they exist within is closer to that of the software company. It is changing constantly, service providers are emerging regularly and a larger amount of exploration is necessary. Changes are easy to make, and easy to undo. In essence, experimentation is cheap, and necessary. In this portion of the company, the S ratio is likely in the 15 – 25% range – not too far off from the software company.
b) Track Configuration
On the other hand, in the area of track configuration, experimentation could prove highly problematic. The tracks don’t operate independent of other entities. The successful movement of train across track is dependent on the tracks being kept within a particular specification – a pre-set thickness of gauge, a specified width etc.
Any “experimentation” in this particular area could result in a possible train wreck. In fact, we try to prevent “experimentation” in this area by making it illegal for people to tamper with the tracks.
This seems obvious, but it is important to distinguish areas like this – where we depend upon infrastructure, or machines to execute a decision that has been previously made in a precise and highly repetitive fashion. That is not to say that people can’t experiment in the laboratory with new track configurations, but with regard to those tracks that have already been laid down, it is pretty clear that S is approximately 0/100.
This is not to say that where infrastructure or machines are involved there is no room for experimentation – in fact, much of the work on computer simulation (and on Artificial Intelligence) is focused on enabling cheaper, safer machine enabled “experimentation,” but in many contexts, the role that machines will play within an enterprise has been (and will continue to be) to improve the precision and repeatability of accomplishing mundane tasks (since humans tend not to be particularly all that good at precision or repetition).
On the other hand, in areas where people are involved, keeping people engaged over the long term virtually requires that they be given some degree of freedom to experiment — otherwise they tend to lose motivation and productivity goes down. S ratios for workplaces that are not entirely automated, will always be greater than 0/100 for that reason.
To wrap up, particular areas of an organization may see large amounts of experimentation or alternatively, an entire focus on repeatable execution of a previously decided plan. On the other hand, for organizations as a whole, S ratios will usually include some degree of experimentation (they won’t reach zero) will include a large degree of execution (they won’t reach 100/0), and the optimum balance for a particular organization will tend to follow these basic guidelines:
1) rapidly changing environment = higher S ratio
2) fairly static environment = lower S ratio
3) cheap / easy to test out new things = higher S ratio
4) expensive / difficult to test new stuff = lower S ratio
5) lots of human involvement = higher S ratio (if only to keep your people engaged!)
6) heavy dependence upon machinery = lower S ratio
Got any thoughts on the S ratio as a concept (or my audacity to name something after myself)? Please contribute to the conversation in the comments below.