Downturns reveal a company's weaknesses. An organisation that seemed nimble and focused during a period of expansion may be sluggish and ineffectual when demand drops off. Survival can depend on quickly determining which products are making money, what customers really value, and where organisational bottlenecks are.

One major cause for this sluggishness is complexity—product complexity, organisational complexity, and process complexity. The costs of complexity are usually hidden, so executives often don't grasp the magnitude of the problem until a downturn hits and the business feels dangerously close to stalling.

The challenge is that some complexity is advantageous, even in a downturn. For example, country or regional business units are closer to the ground than headquarters and are more likely to know what customers want. It takes a complex organisation to provide enough local autonomy so products or services can be tailored to those customers while still taking advantage of global scale. But that kind of complexity can be vital to sustain sales through a recession.

A similar challenge arises when companies struggle to balance complexity and innovation. Adding new products, services, features and options creates complexity of all sorts. But companies become leaders by offering customers new choices, and in a downturn, innovation may be a company's salvation.

The key is not to eliminate complexity but to balance its benefits with its costs.

A useful way of analysing the level of complexity in your company—and separating complexity that's beneficial from complexity that hurts the business—is to begin from a base of zero.

Imagine, for example, that your company produced just one product or service with no variations, sort of like Henry Ford's classic Model T. A manufacturer with only one product would still need a supply chain, a factory, a distribution network, and a sales-and-marketing function. But it could greatly simplify its IT systems, distribution and sales efforts, and forecasting.

One plant manager with whom we discussed this exercise was flying in 15 planes' worth of parts almost every day to meet the next day's production schedule. In a Model T environment, he noted, "all those costs would disappear instantaneously".

The point of the exercise, of course, isn't to return to the days of the Model T, but to determine your zero-complexity costs, and then assess the costs of adding variety back in. Often the cost curve has a "knee", a step change triggered by adding one more model or level of variety, and you can determine whether moving beyond the knee is worth the additional expense. You can also assess the benefits of innovation, and determine the focal point where a given innovation overshoots what most customers want and are willing to pay for.

For example, you might decide to eliminate individual options and instead offer customers a small number of configurations that include the most popular features. Thus, Honda's CRV comes in just eight configurations and 13 interior/exterior colour combinations, for a total of 104 possible build combinations. This presents far fewer choices than most cars offer, yet the CRV is the hottest-selling vehicle in its class.

Similar kinds of analyses can diagnose organisational and process complexity.

We've found that companies get the best results by attacking product complexity first and organisational complexity next, and only then focusing on process complexity. The reason is this: Complex processes often reflect unnecessary product variety or poor organisational design. If you attempt to simplify a process without changing product or organisational complexity, you find even more complexity cropping up in some other process area, like pushing on one side of a balloon only to find it bulging out on the other side.

Unfortunately, most companies that do attempt to manage complexity usually begin with processes, often through efforts such as Lean Six Sigma. Typically, the emphasis is on how companies can execute all their current operations faster and with fewer resources. But that's actually the wrong place to start.

Reducing process complexity should be a company's last step, and it involves looking for the process improvements that add the most value and by eliminating unnecessary data collection.

One of the world's largest natural resources companies, for example, found that it had no fewer than 483 process improvement projects in the works, and that only 25 would deliver a significant impact. In tandem with product and organisational simplifications, the company was able to boost operating income by more than 20%. Meanwhile, the same company found it could reduce its volume of reports by 40% in one major business unit.

All these complexity management efforts help a company become lean and flexible enough to adjust to the changing market conditions in a downturn. It pays off again when the economy improves and a company has stripped out enough complexity to accelerate quickly out of the downturn.

Kevin Meehan is a partner with Bain & Company Southeast Asia. He is a member of the Global Oil & Gas practice and leads the firm's Southeast Asia Technology practice. Mark Gottfredson is a partner with Bain & Company and head of the firm's global Performance Improvement practice.