Reducing possibilities to decisions

Hierarchy has a purpose: to remove uncertainty.

Often the purpose of hierarchy is confused with meritocracy. In a hierarchy, some people hold authority over others. In a meritocracy, people deserve the authority they have because they use the authority well. Ideally the two are perfectly combined, and people in higher positions fully merit the authority that they hold. But as we all know from our own infallible opinions, not everyone who has authority deserves to have it; not every hierarchy is a meritocracy.

Even a hierarchy which is not a meritocracy can be useful.

Imagine four passengers survive a plane crash in the wild. Each has equal survival skills and each is equally uncertain about where the closest civilization can be found. Each guesses a different direction. What would be best for the group? Travelling one hour in a single direction under one leader before switching out and travelling in another direction under another leader? Obviously the group would be better off picking one direction and sticking with it.

This is the fundamental purpose of hierarchy: to remove uncertainty by reducing possibilities to decisions.

Hierarchy cannot accomplish this unless the hierarchy is known and respected. It is the function of the hierarchy itself that must be respected, not the capabilities of each person in it. That the person higher up the hierarchy makes the decision in any cases of dispute or uncertainty is intrinsically necessary for hierarchy to have any value in removing uncertainty.

When hierarchy is not followed, uncertainty proliferates. I watched one CEO hire managers but continue to give instructions directly to employees, without passing it through their manager. Employees learned that any instruction given by their manager could be overridden at any time with new instructions, so they regarded instructions as one possible course of action among many unless the instruction came from the CEO. Employees who had been around for a while were used to this pattern, so new managers never had a chance – all they could do was add more possible courses of action to the ones employees were already considering. The employees all had a notion of what it means to be a manager, so they would generally try to comply with manager directions if it seemed possible; but any time the demands were too many or too contradictory they would default to either what they knew for certain the CEO wanted or what in their own judgement seemed best.

A senior manager who bypasses a middle manager forces that middle manager to be worse than useless. A useless person adds nothing; a worse than useless person subtracts something. A manager who is bypassed by his boss cannot reduce uncertainty for his employees but adds to it by giving additional input that may or may not be let stand by the senior manager, and takes up his employee’s time asking them what they have been told to do by the senior manager.

In another organization, the general manager of a factory had reporting responsibility to a vice president of manufacturing and a vice president of sales. These two VPs had different priorities. For one, the highest priority was increasing on-time delivery; for the other, the highest priority was reducing cash tied up in inventory. A very high-performing organization can accomplish both, but the fastest way to make progress on either of those goals is to sacrifice the other. Over the course of several years the factory would whiplash back and forth, approximately once a quarter: a few months of squeezing out inventory wherever possible, which resulted in deliveries slipping out when the least unexpected event caused components to be unavailable; then a few months pushing for delivery on schedule, causing accumulation of inventory to account for unexpected disruptions.

A matrix organization is a fundamentally bad idea because it reverses the most useful function of a hierarchy. Rather than having a single superior to resolve uncertain situations, additional uncertainty is produced by the possibility of disagreement between authorities. The matrix organization was invented out of a recognition that sometimes engineering skills are needed to accomplish sales goals, and it was somehow imagined that the only way to have access to skills within an organization is to be the boss of a person having those skills. But this is immature at all levels; respectful people who communicate clearly can obtain assistance from people whom they do not control. This ought to be even more true of people at higher levels in the organization; if your VP of Engineering is useless at supporting your VP of Sales, why is he your VP of Engineering at all?

Furthermore, the matrix organization solution doesn’t scale, as there are quite often more than two relevant interests. If making everyone with a legitimate interest a “dotted line” supervisor helped, an organization would become a dense thicket of “dotted line” responsibility and most employees would find that they must treat anyone higher up in the hierarchy as if they were their own manager. Too often employees do feel this way; but in such organizations very little real change occurs, and the organization survives on the inertia it gathered in leaner days to keep producing enough value to survive.

Hierarchy has a specific and real value, and it is a mistake to think of hierarchy as merely a reward system: you do a good job, you get a promotion. A job well done can be rewarded with more pay, or in other ways. A hierarchy is not necessary to have a reward system. Likewise, if merit were equally obvious to all, a hierarchy would not be necessary because people would spontaneously agree on the best option. And a hierarchy does not require that all possible decisions be decided at a higher level than the question originated; generally, an organization will work faster and more effectively if decisions are made lower in the hierarchy.

But at some point uncertainty is inevitable; there will be significant doubt about which choice is best, and the consequences will be significant. It will be impossible or grossly ineffective to attempt more than one option. Perhaps more than one choice will be effective. The only way you reduce the infinite possibilities to decisions, actions and results is by allowing one person in a group to have the final say in decision making.

The Matrix Can’t Read Your Mind

One of the typical tools for facilitating product portfolio management is a decision matrix – some kind of multi-factor analysis that might include, for example, market size, market growth, and expected margin. Many also include subjective criteria such as “fit with strategy.”

I had the privilege of building one such matrix. I chose to use only qualitative factors since one of the main reasons for the project was to provide objectivity to a decision making process that was regarded as arbitrary and subjective. Although I did have a factor for strategic fit, in that particular case an established industry trade group provided a market breakdown that corresponded closely with the company’s own strategic classification.

After setting up some nice formulas in an Excel workbook I was ready to show my work. I demonstrated to the project sponsor that the criteria weighting could be easily adjusted: if market share were set to the highest priority, Option A emerged as the strongest; if profit margin were the top priority, Option B became the most compelling; and if long-term strategy were most important, Option C took the top spot

All that was needed, then, was clear input on the business priorities, and from that new product development priorities could be clearly, consistently, and rationally defined.

Consequently the tool was never used.

A major part of the reason the existing process was perceived as arbitrary and subjective was that the strategic priorities were not clearly defined in the first place; in fact the official vision statement of the corporation, if actually practiced, would have resulted in direct competition with the parent company of the corporation – with the end result of the child overthrowing the parent!

In a world in which we can measure sub-atomic particles, objective quantifiable data is available without limit. Data can never tell you what to do unless it is filtered through a purpose; you must already know what your intentions are before information can help you understand how to achieve them. Purpose makes the difference between bricks and mud, between surgery and assault, between chaos and progress. You cannot make data-driven decisions until you know where you want to go.