Why Copycat Strategy Doesn’t Work

A Story on Strategic Complexity

Introduction

Jesse Weaver recently wrote that “emulation is not a product strategy”, and he’s right — copycat strategy is not a strategy that can succeed in any moderately complex space. His analysis of YouTube Premium, a content subscription product, provides the basis for several qualitative, perspicacious observations:

  • Copycat strategy is not actually easy. Because of the number of capabilities that create a product and customers’ perception of it — trying to copy all of those capabilities is risky and a distraction from alternative opportunities that better fit the capabilities you have already developed successfully.
  • Products have momentum that can’t simply be reversed. YouTube has years of operating as a low-quality, free content product, and its customers and suppliers (including that seven-year-old millionaire) have built expectations and processes based on that. Adding higher quality content and putting it behind a subscription won’t change their minds.
  • Switching from free to premium is hard. It has two cognitive heuristics working against it: the Endowment Effect (free YouTube is ‘mine’ and I don’t want to give it up) and the Zero-Cost Effect (we greatly overvalue free things, even when alternatives are very close to free).

Building on his work, I developed a story about copying a classmate’s paper to further show why copying strategy doesn’t work.

Copying a Paper

Suppose you have a five-page paper due for your History of Strategy class, and although you have delayed working on it until the weekend before it’s due, you are not worried. Your teacher has given you permission to copy, entirely but with attribution, the first four pages of any of your classmates’ papers. The final page, the conclusion of the paper, is the only original work you are required to create. In fact, she has given everyone in class this permission and will post the first four pages of all of the papers as they are submitted. Since two papers could have an identical first four pages and nearly the same conclusion, your teacher gives “ties” to the originator: if you copy someone else’s work and you both reach the same conclusion, the person you copied gets credit for anything your conclusions have in common.

In your History of Strategy on Friday morning, your teacher reminds everyone that the paper is due Monday. She also announces that so far, 25 people have submitted a paper. With so many papers to use as a starting point, you think to yourself, one weekend is more than enough time to read through them and add a conclusion. In fact, you probably just need Sunday night. Your plan to spend the weekend at your friend’s house fits in easily.

She then outlines three models of strategy she is going to teach today:

  • Strategy as Product(Design, Skill)
  • Strategy as Independent Capabilities
  • Strategy as Interdependent Capabilities

Strategy as Product(Design, Skill)

Your teacher begins the day’s lesson on Paul Romer’s design-skill production model from his seminal paper on Endogenous Technological Change.

The model is simple: producing a good has two inputs, A, which is available to everyone, and X, which is available exclusively to you. She writes F(A, X) on the board.

“This equation is about a production function using two things, A and X. Think about the admonition to avoid ‘reinventing the wheel’,” she says, and points at the equation.

“Don’t reinvent the design of the wheel, that’s A here. Everyone has a wheel design available. The wheel design is nonrival — we can all use it simultaneously. And it’s nonexcludable — none of us can keep each other from using it.”

Then she points at X. “But this part is a skill. It’s the skill of taking that wheel design and actually making a wheel, or making a wheel and putting it together with something else, like a belt or a gear. The X part is rival, since you might have that skill and can’t use it to make two things at the same time. It also can be excludable, if you develop a tool to make wheels that you don’t offer to others. A is a design. The ability to do something with that design is X. Skill transforms design.”

“Strategy is a function of both of those inputs. It’s not just about finding designs to use but also about developing the skills to use them. It’s why two burger restaurants can have the same grill and frying technologies, the same customer segments — all designs — and have completely different outcomes. Maybe one has better-trained employees who make more and better food using the same technology; or a master chef who creates better food items.”

“But don’t make the mistake of copycat strategy and think all you have to do is copy a competitor’s technology, or a business model, or a customer segment. Copying strategy looks easy only if you think about designs. It’s much harder when you think about the skills you need to make those designs work.”

“Now think about your papers.” She says. “When you go to write your paper, don’t think that because you found a great A, a great design, in one of your classmate’s papers that your work is done. That design is one input. Your skill is the other. You can copy one, but you need both, in a paper and in a strategy.”

Strategy as Independent Capabilities

The next model your teacher presents is due to Porter. He starts by representing strategy as a set of capabilities, and then he gives a competitor a likelihood factor for successfully copying a capability.

“Copying a strategy requires successfully copying each of a competitor’s strategic capabilities.” Your teacher begins. “Romer’s model helped us see that strategy has two inputs — designs and skills. You can copy designs, but you have to develop skills. That’s where Porter comes in. For him, everything is a ‘capability’. Having a design is a capability, and using that design is a capability.”

“What Porter does is help us see that copying a capability has some less-than-1 chance of success. Let’s go back to the two burger restaurants. Burger Guys is already in the market, and Kahuna Burger is building its first restaurant. Kahuna Burger knows all of Burgers Guys’s food production technology (grills, friers, refrigerators, sinks, cooking tools), food items, ingredient vendors, employee training, and layouts. That’s five capabilities in Porter’s model.” She writes the capabilities on the board.

“Now, what’s the chance Kahuna Burger succeeds on any one of those? Start with restaurant layouts. Burger Guys have layouts partly determined by the space they’re leasing. Kahuna Burger can’t exactly replicate that space, they have to work within the space they have available. Let’s say it’s 80%, or .80. How about vendors? Maybe negotiating with each of them, Kahuna Burger gets seven of the ten, so that’s .70. And everything else, we’ll say is .80 or .70.” She writes the likelihood next to each capability:

  • Food production technology * .80
  • Food items * .80
  • Ingredient sourcing * .70
  • Employees * .70
  • Restaurant layouts * .80

“Getting any one of those capabilities right doesn’t look easy for Burger Now. To actually copy those requires some skill, and we said they don’t have 100% of the skill in each of them, they’re between 70% and 80% on each of them. Getting the whole strategy right is the chance they get each capability right. That’s all of the capability likelihoods multiplied together.”

She writes “80 * .80 * .70 * .70 * .80 = .251” on the board.

“Kahuna Burger has a one-in-four chance going the copycat route, and that’s only if Burger Guys has five capabilities.” She says. “I think that’s a significant underestimate if anything. Copying strategy means matching capabilities, and even if you’re good at matching each capability on its own, the product of your copycat strategy will fail.”

“With your papers, what capabilities do you need? An introduction, middle, and conclusion, sure. But how about researching, explaining, writing, citing, and connecting? I think you need at least those capabilities too. The best papers will show them throughout; copying the start of a paper won’t build those capabilities, and I doubt you can build them if all you write is the conclusion.”

Strategy as Sets of Interdependent Capabilities

The final model your teacher presents comes from Rivkin. He starts by counting the number of possible capabilities a competitor could have, writing that as the number N. Then he counts the number of capabilities that are interdependent, writing that as the number K.

Your teacher writes “NK” on the board and turns to the class.

“Porter’s model introduced us to capabilities. Rivkin’s model takes capabilities and uses them in two ways. First, he represents capabilities as a set, that’s N.” She points at the “N” on the board. “For our burger restaurants, there is the whole set of capabilities they could use, and then there’s the subset they actually use. Picking a subset is difficult because knowing how one subset works doesn’t tell you about how well a different one could work. Burger Guys doesn’t have a ‘fusion’ menu (blending different cuisines), it uses a Midwest-style menu. But that doesn’t tell Kahuna Burger that a fusion menu wouldn’t work. Kahuna Burgers has to think about all the different subsets of capabilities it could use, and without some insight about how to go about that research, it can take months, even years, to figure that out, depending on the number of possible capabilities. Porter only modeled capabilities that matter. Rivkin shows us that the step of deciding what those capabilities are is difficult.”

She taps her chalk on the K. “Rivkin also brings in interactions between capabilities. That’s K. Interactions between capabilities mean that a choice on one capability increases the benefit or cost of other capabilities. Burger Guys might decide on a complex food item as part of its menu, which requires more costly food production technology or more costly employee training (or both).” She picks up a stack of papers and divides into smaller stacks, handing them to the student at the front of each row.

“Take a look at this visualization of Rivkin’s NK strategy model.” She says over her shoulder as she walks toward the chalkboard.

A Rugged Landscape (Randy Olson [CC BY-SA 3.0 (https://creativecommons.org/licenses/by-sa/3.0)], from Wikimedia Commons)

Standing at the chalkboard, she continues. “What this shows is how N possible capabilities and K interdependencies together produce a landscape. Each peak is the value produced by a subset of the N capabilities that have K interdependencies. Have you ever heard about ‘local v. global optimization’?”

Heads nod.

“Well this is a visualization of why that’s hard. When you are developing a strategy, of course you want feedback as soon as possible. If that feedback is positive, you continue going that direction. What can happen is that you reach a local peak. Now, you could start over, or adjust your strategy. But with K interdependent capabilities, you can’t adjust them incrementally. You have to adjust combinations of them. If you’re Kahuna Burger, you can’t just copy Burger Guy’s Midwest-style menu and then change to ‘Asian fusion’. You’d have to change your menu, your training, your ingredient sourcing, and your marketing. And that’s much, much harder to do than getting one capability right and moving on to the next one, as Porter’s model allows.”

“Applying this to your paper, I suggested that researching, explaining, writing, citing, and connecting are all potentially important capabilities your papers must demonstrate. But now you have to think about what capabilities this leaves out — what other capabilities are there? And how do capabilities interact — if you include diagrams in your paper (a capability I didn’t mention before), does that mean your writing need to be better, can it be worse, or is it an independent capability? What you have to do is have a theory, a heuristic, an intuition, all of which can be wrong but useful, to guide how you select and develop capabilities. Starting with whatever capabilities are in a classmate’s paper and incrementally improving on them: that’s a surefire way to get to a local maximum. And in a landscape of many local maxima, you’ll be average at best.”

Strategic Complexity

Your teacher looks over at the clock on the wall, which shows only a few minutes left for class.

“Ok, our time is nearly up. Let’s go through each of our strategy models and what they tell us about copycat strategy.”

“Romer demonstrated how strategy depends on designs, which can be copied, and skills, which can’t be. Copycat strategy can’t work because it’s incomplete. The questions to ask given Romer’s model are: What designs are we copying, and what skills do we need to implement them?”

“Porter showed that copying a strategy means copying capabilities, which is difficult even when you are good at copying strategies individually. A less-than-1 chance of matching each of your competitor’s capabilities results in a low likelihood of success for your strategy overall. The questions to ask in Porter’s model are: How many capabilities are we trying to match, and what is our chance of matching them?”

“Rivkin adds to Porter’s model of capabilities. His model shows the difficulty of finding the right subset of capabilities, and the difficulty of changing capabilities when they are interdependent. The questions Rivkin’s model raises are: How will we search for capabilities that matter, and which capabilities are interdependent?”

“Together, all of these models demonstrate that more, more difficult to develop, and more interdependent capabilities increase strategic complexity. When strategy is complex, it doesn’t matter that competitors can see what the strategy is. Strategic complexity is itself an entry barrier. Apple doesn’t own all of the world’s smartphone production facilities, and it still has a competitive advantage in smartphone production because of its design, distribution, and marketing capabilities.”

“Strategic complexity is an entry barrier for the focal firm as well. As Steven Sinofsky recently discussed, Apple cannot just lower its smartphone prices without also changing the interdependent capabilities of smartphone production, distribution, marketing, and service delivery. In other words, the entry barrier Apple created for others to the premium smartphone market is also an exit barrier to it leaving for the mid-level smartphone market.”

Towards Strategic Novelty

The class bell rings, and your teacher pauses. Students leap to their feet and gather their belongings.

As conversations far distant from the History of Strategy erupt, your teacher shouts over the din:

“Remember, papers are due Monday morning by class time. And think about whether copycat strategy could work for you — or ever!”

Walking down the hallway as students rush past, you think about your weekend plans and your paper. Copying is something that all of your other teachers forbid, and you thought your History of Strategy teacher was cool, but crazy, to allow it. It gave you the chance to use someone else’s paper, add to it, and still spend a weekend at your friend’s house. You shake your head. None of the models you learned today say you can’t use a copycat strategy. But all of them show it won’t work.

You message your friend, letting him know that you can come by but not until your paper is done. Then you head for the library, to do the hard work of creating a novel strategy for your paper.