Coming up with good strategies can be really hard. In 1997, Daniel A. Levinthal wrote a paper called “Adaptation on Rugged Landscapes” explaining a bit about how life works and why it makes good strategy so difficult. Whether you’re creating a strategy for your business, your chess game, or your life, you can improve your success rate by getting a grasp on the concept of Rugged Landscapes.

Strategy

First, let’s get a basic understanding of strategy. Although there are about as many definitions of “strategy” as there are strategy experts, we’re going to use the word “strategy” to refer to “the alignment of the component parts of an entity in a way designed to achieve a goal.” This is not unlike a common definition for strategy that simply states that it’s “a goal and a plan to achieve it,” but we’re going to go a step further and explicitly include the idea of “alignment.”

“Alignment” refers to setting up the operations, objectives, incentives, and all other component parts of an entity in a way that it supports and enhances the other components. If the goal of your strategy is to beat the Nazis, then your combat units (infantry, artillery, navy, air force, etc.), logistics units, commanding officers, and political systems all need to work together in a way that moves the whole entity toward the same objective. If the air force attacks random targets that provide no help to troops on the ground nor gain political advantage, then the air force is out of alignment, and the strategy will have a lower chance of success.

In a business, you might be aligning finance operations with marketing, logistics, product development, sales, accounting, and everything else. If product development is working on products that run against the goals of the company (creating an airplane when the company sells cars for example), then the product development team is out of alignment.

Rugged Landscapes

Visualizing the Concept

Levinthal called his paper “Rugged Landscapes” for a reason. Visualizing strategic success as a landscape helps to understand the various complexities. When we think about strategy, we conceptually want to be on the highest part of the landscape; it’s a metaphorical game of “King of the Hill.” If we put this mountain on a 3-dimensional plane, you can think of climbing a mountain as increasing your value on the Y-axis. To do this, you need to move yourself along the X and Z axis until you get to the highest peak.

In this concept, the X and Z axes represents different attributes of the entity or strategy. As an example, the attributes of your strategy against the Nazis might include the balance between anti-tank and anti-personnel training that you give to your infantry, and the balance of anti-mechanical and anti-infantry missions that your Air Force executes. This is a highly simplified model because true strategies will need to align far more than 2 attributes, but it does give us an idea of how the concept works.

In the visual example of a ” rugged landscape”, you can see how your anti-Nazi strategy tends to work well when your infantry and Air Force have a relatively equivalent balance, and works best when they give approximately equal time to both. You can also see a sharp drop-off when one division focuses differently from the other, such as when the Air Force focuses on mechanical targets while the infantry focuses on enemy infantry. The real world will look very different from this chart, but hopefully this example can help solidify the concept in your mind.

With the basics established, let’s consider how this concept helps us understand why strategy can be so hard.

Interdependency

The concept of “interdependency” partially explains why strategic alignment can be difficult. In its simplest form, interdependency refers to the idea that a change in one attribute of a strategy may impact the efficacy of another attribute of a strategy. This principle severely complicates the real world because attributes often interact in ways we either don’t expect or can’t predict. Even when we can predict that two attributes will act interdependently, we often can’t predict the degree of impact that one will have on the other with much accuracy.

For example, say our air force had previously been focusing its efforts on disrupting enemy logistics for infantry support, but  changed its focus slightly more to disrupting enemy logistics for mechanical and manufacturing support. Although we might be able to predict that this could impact our troops on the ground, we might not be able to predict just how big of an impact it will have. This might put you more or less in alignment with the infantry, which may increase or decrease your success against the Nazis. In our visual example, you can see that the slope of our landscape is very steep, indicating that whatever change we make, it will likely have a significant effect one way or the other.

Every attribute in an organization has a risk of being greatly interdependent with other attributes. Although we might be able to predict the consequences in our air force example above, many times we can’t readily identify how a slight change in one aspect is going to impact others.

Local Optimization

If interdependency wasn’t troubling enough, life throws another curve-ball to aspiring strategists in the form of localized peaks and valleys. Our anti-Nazi example looks relatively simple with its single peak, but what if the landscape looks more like this:

As you can see, this landscape really is a bit more rugged, and closer to what real life might feel like. Instead of just one peak, there are 4 big ones, with a spattering of smaller peaks, a few valleys, and moderate planes.

Depending on where you start with your various strategic attributes, you will probably continue to change and modify your strategy until you reach a localized peak. However, this does not guarantee that you will reach the highest peak on the field. You will only likely reach the closest peak to wherever you start from.

For example, say you were creating an airline company. You leverage two attributes: price, and the number of locations you serve. If you start your planning with attributes close to one of the group of 3 peaks in our visual above, you’ll likely develop and adjust until you reach the top of one of them. If you try to make adjustments from there, you’ll see decreased performance, and not go any further, not knowing that far away from our little mountain cluster is a hill even higher (and therefore more desirable) than our own.

Against this example, you might ask why someone wouldn’t recognize the higher, distant peak and adjust themselves to get there. The next two strategic challenges explain this: path dependency, and the fog of the landscape.

Path Dependency

If you have taken a finance class or econ class, you’ll likely have learned that “sunk costs are irrelevant.” Although this is true for financial problems, sunk costs must be considered when creating strategy. The reason for that is because you can’t instantly divest and reinvest in a completely new set of attributes.*

In a business, your attributes may include things like the number and type of manufacturing plants you own, supply chain operations, company culture, hiring/firing/recruiting processes, brand recognition and reputation, and many other things. Moving from one set of attributes to another can’t be done instantly, but is done over time, often incrementally.

Consider again the rugged landscape above. If you’re a business on one of the 3 nearby peaks and you want to bring your business to the highest peak, you can’t simply wave your fingers and have the necessary changes already done. You will likely have to slowly make the necessary changes, which in this case, might pull your company through a long, deep valley. As my professor called it, you must go through a “valley of the shadow of death.” Not all companies can survive the time and resources it takes to change its attributes enough to reach a new peak.

Where you are now depends on where you’ve been. Where you can go depends on where you are now. You must trudge through every point along the way if you want to make changes.

Foggy Landscapes

Another challenge to creating good strategy comes from the limited information available in the real world. In most cases, you will never know if you’ve actually found a peak. You will also never know if you’ve found the highest peak available, or if some major leaps of strategic change will put you close to an even greater peak.

This wouldn’t be strategy if it didn’t operate in uncertainty.

Changing Landscapes

The landscape itself provides a final challenge to strategy creation. Put simply, the landscape changes. Today’s peak might be tomorrow’s valley. Internet search companies are a good example of this: AOL, Yahoo, Lycos, and many other search engines spent a fair amount of time on local peaks in their strategic design. As technology advanced and other forces changed the landscape, their peaks dropped and left other organizations (Google) to plant themselves on higher, safer peaks.

The story behind Blockbuster Video’s eventual demise is a good example of this as well. Blockbuster had a great foothold in the movie business, sitting on quite a firm peak for a fair amount of time. However, tech advances enabling new ways to watch movies eventually shifted the landscape in favor of entities with a different set of attributes, sucking the floor away from Blockbuster until it closed.

What This Means for You

Having an understanding of the field can strengthen your intuition for solving strategic problems, but this quick glance at the theory yields some specific insights as well.

First, before you start building a strategy, try to learn everything you can about the “landscape.” Due to the “path dependency” issue we talked about earlier, much of where you end up will be determined by where you start. Most strategies can only make incremental changes, which often results in climbing the nearest peak. Anything you can do that helps you understand where the highest peak will be is beneficial.

Second, once you have established yourself on a peak, you need to find ways to continue to experiment. One of the best ways of experimenting on a rugged landscape is to create independent entities with new attributes, essentially creating a strategy from a new starting point. Innosight tapped into this idea relatively quickly, helping Deseret News build a new foothold on a growing peak while their traditional newspaper business saw its established peak melting away.

Third, if you must push an organization to a completely new peak, do so as quickly and completely as possible. A company heading through the valley of the shadow of death shouldn’t linger. This requires strong leadership and a dedicated team, as well as a whole lot of luck to accomplish. My post on “Grounding Radical Change” expounds on one [less than ideal] way of accomplishing this.

Fourth, look for ways to impact the selection environment. BCG pointed out that Tesla’s move to allow others to use its patents could be an advantageous move for Tesla. Success in the automobile industry is partially dependent on network effects, such as the availability of fueling stations. By making its technological advances available for more businesses to use, Tesla is creating an incentive for others to increase the usable network for electric vehicles, making the fully-electric vehicle a decent hill to fight for.

Finally, stop obsessing over finding the correct answer to strategy. Harvard Business Review published a decent article called “Why Smart People Struggle with Strategy.” Smart people often face the hurdle of needing validation that they arrived at the correct answer. In strategy, there isn’t necessarily a single right answer, and even if there is some optimal peak, you will never really know whether you found it. The landscape is foggy, changing, and rugged. A good understanding of strategy will enable you to make choices in the face of uncertainty with the best chance of success.

So as you develop strategies in your businesses, games, and life, keep these principles in mind, and you’ll find a healthy peak.

Keep seeing truth.

*I’m aware that this is a bit of an abuse on the idea of “sunk costs.” A true consideration of sunk costs will consider the challenges of changing the infrastructure and other attributes as future costs and account for them in an analysis. However, since much of where an entity is today is a the result of these often unconsidered “sunk costs,” my professor chose to use the concept to help explain the idea of path dependency.

Lagniappe: The Selection Pressure

I left out one important aspect of the Rugged Landscapes theory: the actual selection mechanism. This is the actual causal mechanism behind whether a set of attributes will be successful or not.

For example, in nature, one critical selection pressure is the ability to get food. If a bunch of birds live in an area where seeds are the primary food source, deep beaks might be an attribute that results in greater success due to its ease in opening seed shells. If the primary food source consists of bugs burrowed in trees, longer beaks will be the preferred attribute. The food source and availability create a selection pressure that causes one attribute or another to be successful.

The landscape itself will change for these birds as food sources change in availability. A season with lots of bugs but few seeds will see peaks rise on the long-beak side of the “beak” attribute, while a season thick with seeds will lift the deep-beak side.

It is not always necessary to understand the actual selection pressure to understand the concept of rugged landscapes, but once a strategist identifies the true selection pressure, he or she can often have an easier time predicting the ideal characteristics of simple attributes.