Everyone who has ever worked for a large company knows that they make really silly decisions. Completely illogical decisions. Decisions so monumentally ridiculous that you wonder how the company actually manages to survive as a going concern, let alone turn a profit. It's seemingly obvious to everyone in the organisation, except the senior executives who are making the decisions. Good projects aren't funded, bad ones are. Good teams or departments are restructured but poorly performing ones aren't. Opportunities are lost. How do they continue to make money with all these bad decisions? And why do smart executives continue to make them?

The answer of course is that big companies very seldom make truly bad decisions. What they make are a lot of very sub-optimal decisions. Decisions made are seldom illogical, there is a lot of reasoning that goes into them. Unfortunately, that logic and reasoning is based on very poor information. The decisions they make are good enough to stay in business and continue to make significant amounts of money. They just aren't the best decisions possible. The real question isn't "how can companies still make money while making poor decisions" but "how much more money could they make if they made better decisions". Looking at the reasons why companies make sub-optimal decisions can point us to ways to make better ones.

The main problem large companies have with decision making is centralisation. The way large companies are structured, most of the important decisions are made by a few key people. Unfortunately those key people are located a long way from where the action is. It's very much like a general hundreds of miles behind the front directing troops based on high level maps. While the decisions made can be good enough to win the war, there may be far better decisions that can be made at the local level to achieve the same outcome.

The reason local decisions are often better is that the decision makers have far more information available to them. Because there is just so much information involved in planning a large enterprise centrally, the decision makers simply can't process it all, so instead they work with summaries and rolled up data. So the extensive 80 page business plan someone puts together to explain their concept gets converted to a 10 page summary for the local exec team, then to a 3 page summary for the central PMO, then a half page executive briefing for the executive planning committee before anyone makes a decision. Some companies have entire departments whose only job is to make these project plan summaries for executives.

While this may preserve some key information, usually some ROI calculation or other financial numbers, mostly because they are amenable to being easily rolled up and summarised, so therefore must be important (which is a manifestation of the measurement fallacy), all the context gets lost. It just comes down to a raw financial numbers game. Unfortunately, the financial numbers are also quite easy to play with to make your case look good. I know managers who are very successful just because they are very good at juggling numbers to make projects look better than they are. An overstated benefit here, an understated cost there. Planning becomes a game of who can fake their financials better.

With no context there is no information on why the decision is being made, planning becomes disconnected from strategic goals. Someone's pet project that has no strategic alignment (but with really nice looking numbers) will get prioritized above a project with less impressive fake numbers but far greater strategic alignment.

The other disadvantage with centralised decision making is speed. Or more precisely, the lack of it. Executives are busy people so they tend not to meet very often. Centralised planning tends to be an infrequent event. At best quarterly but more often annually. It is not uncommon for project approval in a large company to take 12 to 18 months. Not only does this slow down the organisation's response to changing conditions (because you essentially need to submit a chunk of work large enough to fund you till the next planning cycle) a centralised system tends to encourage large batches of work rather than multiple smaller ones. Large batches have their own set of problems and tend to be less successful than smaller ones so the system really skews towards things that are likely to have a poorer outcome.

So, a centralised planning system favours large batches, works with minimal information, lacks any context around decisions and is slow. The decisions made are entirely rational and perfectly correct given the information available, they are just highly sub-optimal.

There is also a slightly more subtle problem here as well. The senior execs in an organisation are there to set strategy. By getting them involved in funding individual projects or other batches of work, they are being lifted out of the strategic world and being dropped straight into operational details. It's even worse when, as most organisations do, they have the same execs who make the funding decisions also getting involved in overseeing the spend of the money they allocate through some sort of governance committee. You have just turned your senior execs into glorified project managers and made them so busy overseeing spending in various projects, that they no longer have time to do real strategy.

The reason companies centralise decision making, obviously, is for control. Unless the executives who set the strategy make the funding decisions, how can they ensure that decisions align to strategy? Of course as we have seen, the centralised system, by removing decision context, actually tends to lead to poor strategic alignment. Most execs will freely admit that when asked about funding, but surely allowing funding decisions to be made by a wider group would result in even poorer alignment? And how can you control spend except by having some centralised body to keep track of who has been allocated what?

These are very valid questions and fortunately there is a way to satisfy them while allowing for better decision making. It involves changing the process to a mix of centralised decision making based on strategic goals and decentralised decision making based on the tactical decisions needed to meet the strategic goals. More on that next time...