I am slowly rereading Richard Rumelt’s Good Strategy, Bad Strategy. It’s the best discussion of business strategy I’ve ever read.
Rumelt on centralization vs. decentralization
I just read the section where he discussed the tension between good strategy and decentralization. He points out that good strategy requires a lot of coordination across an organization. In other words, good strategy requires clever central planning, the opposite of decentralization. Here’s what Rumelt has to say, starting on page 92:
The idea of centralized direction may set off warning bells in a modern educated person. Why does it make sense to exercise centralized power when we know that many decisions are efficiently made on a decentralized basis? One of the great lessons of the twentieth century-the most dramatic controlled experiment in human history-was that centrally controlled economies are grossly inefficient…
But decentralized decision making cannot do everything. In particular, it may fail when either the costs or benefits of actions are not borne by the decentralized actors.
His point is that centralized action can have extraordinary benefit when applied in a limited way. One of the key challenges of good strategy is finding the limited ways that centralized action can create huge wins. With this knowledge in hand, you can focus your strategic action, and let decentralization work its magic everywhere else.
We can think of a business as a centrally planned economy
Let’s explore this idea of a business as a centrally planned economy. In some ways, the connection is quite clear. The senior executives are the planners. They dictate a productivity quota for each group in the company. Manufacturing will produce X units. Human Resources will hire Y employees. Research and Development will deliver Z new designs.
While it’s clear how executives plan for internal productivity, it’s less clear how executives plan for internal “prices”. Imagine that, instead of an integrated company, each function stood as its own entity. Manufacturing would need to purchase new designs from Research and Development. Operations would need to purchase products from manufacturing. All the entities currently within a company would have to execute transactions with each other. These transactions would come with corresponding prices.
In a single, integrated business, these entities are forced to transact with each other. They can’t set their own “prices”, in part because all margin is aggregated at the corporate level. Each group then operates at its own cost. Rather than focusing on planning internal “prices”, executives plan for reducing internal costs.
The hope is that the margin each group could generate in its transactions with other internal groups flows along the value chain. We reach the end of the value chain when we close a transaction with a customer. At that point, the accumulated profit accrues to the company as a whole.
Why not let all the parts of a business compete as independent entities?
Now we can start to see the tension between strategy and decentralization. Imagine a large business with a bunch of different functional organizations (Manufacturing, Human Resources, Research and Development, Finance, Sales, Marketing, et cetera). Now imagine a competing universe where each of these organizations was a standalone business.
What advantages do you get by integrating all these organizations under the umbrella of a single company? A big one is that you get to keep all the margin in house. If they were separate, these entities would have to negotiate prices amongst themselves. Some would be more profitable. Others less so. When you integrate, all the margin is kept under one roof.
Why is it helpful to keep the margin under one roof? Because you can make more efficient investments. The most profitable entity might not require the most investment. You can then use profit dollars from the most profitable entity to make attractive investments in another entity. In other words, you can use profit dollars from the most profitable entity to ultimately make the other entities more profitable.
How a business invests in itself is a key element of strategy
Investing capital across an integrated organization is one important manifestation of strategy. It requires clever central planning. The senior executives need to know where the most attractive investment opportunities are. They need to know how many dollars to steer toward these opportunities.
How do we know if the senior executives are doing a good job? You can compare the company’s results against the market. The obvious choice is to compare them against the results of competitors. But you can also compare them against the results of a portfolio of smaller companies that, when added together, mimic the offering of the larger company.
In a sense, that’s what an executive team is competing against…the known value of decentralization. You can break a large company apart and let each entity fend for itself. Market forces will eventually determine which parts rise, and which parts sink. As Rumelt points out, we learned that centrally planned economies are grossly inefficient. Decentralized market forces are simply that powerful.
That’s a tall hurdle for executive teams to clear. And the most powerful tool executive teams have at their disposal is strategy. That’s part of the reason centrally planned economies fail so miserably. They’re way too large, with way too many moving parts. There’s little hope they can cobble together the collection of good strategies necessary to win.
Access to good strategy is the true justification for large integrated businesses
Large integrated companies get other advantages, besides just keeping margin in house. They get economies of scale. They get predictability of demand. They get access to cheaper capital. All of these things can happen, even with bad strategy. That’s why large integrated companies don’t necessarily fall apart when they execute a bad strategy.
Still, without good strategy, the best case is treading water, where the advantages immediately above balance the losses from centralized planning. If you can make central planning an advantage…something you can only do with good strategy…that’s how your company becomes transcendent. That’s how you outperform the market, and do a great service for all your stakeholders.
Here’s a somewhat sloppy way of putting it: good strategy is the exception that proves the rule that central planning is inefficient. Empirically, it’s difficult to design and execute good strategy. As your company, or your economy, gets bigger and bigger…it’s even harder to design and execute good strategy at the necessary scale. The inefficiency of central planning overwhelms the accrued benefits of integration.
Rumelt is right that decentralization can’t solve every problem. Typically, the most interesting and complex business problems require a centralized solution. Ironically, the true magic of central planning emerges with the embrace, not the suppression, of market forces.