For the last few weeks (interrupted briefly by the holiday break) we have been looking at executive coaching. We have taken a look at some of the big problems executives face and at some of the ways we can use agile tools to help resolve them. We have looked at resource planning, controlling financial spend and estimating ROI. All these things, though, are manifestations of a more fundamental problem - the problem of control. Control is a real issue for executives. They are responsible for a P&L. They have business goals to meet. They have people under them to meet those goals. They are expected to be in control.

In a traditional environment they maintain control through their position as central decision maker. Any significant decision will be funneled up through them. In an agile environment we recognise that centralised decision making is slow and inefficient, so we decentralise the decision making for efficiency. The problem is that, to the exec, we have taken away their decision making (and therefore their control) and not given them any other control mechanism to replace it. Without some alternative control mechanism, execs in an agile environment will continue to rely on their old control mechanism - centralised decision making - to the detriment of the agility of the group. All the unnecessary steering committees, status reports, executive briefings, financial controls, and so on are all manifestations of this fundamental problem - how does an executive maintain control when they are no longer the one making all the key decisions?

This is where agile coaches can start to sound a bit like cheap self-help gurus talking about the need to become leaders rather than managers, and to influence rather than dictate. While that is desirable, the reality for execs is that they do actually need to be in control. It's their job. The organisation expects it. No matter how much we talk about leadership vs management, they do need to be in control.

The trick here is that in an environment where decisions are distributed, centralised control no longer works. Centralised control really requires centralised decision making. In a decentralised environment we need decentralised control as well as decentralised decision making.

It's worth looking at the two types of control here. In a centralised model, information flows up to a central point where it is integrated and decisions are made. In a distributed model, information flows upwards but only as far as it needs to for a decision to be made - if a decision affects only one team, it is made at the team level. If a decision affects multiple teams it will flow up a little higher. If it impacts the whole portfolio, it will flow right up to the exec. Information flows around in a network where decision making occurs at every node in the network rather than being centralised at one master node.

The way an exec maintains control in this scenario is by setting the parameters for decision making at the other nodes in the network. They set the environment in which decision making happens. So if a decision impacts a single team, they can make that decision provided it fits within certain parameters - cost impact less than $5000 say, or schedule impact less than 2 days. If it falls out of those parameters, it gets escalated up to the next level in the network where the decision making limitations are higher.

So decision making occurs at the local level but still under the control of the person at the top. The other big thing here is that the results of the local decisions are propagated up the network and used to tune the decision making parameters. So if teams are making too many decisions to drop scope, or to extend dates and that is now have a material impact on the overall program, then the higher nodes in the network can turn down the decision making limits for the nodes below. Or if too many trivial decisions are being escalated, the limits below can be relaxed. So the exec maintains control by creating and maintaining the decision making network.

But how does the exec know they are in control? Sure, they can set decision making parameters and keep tweaking them, but how do they know things are running according to plan if they aren't making all the decisions? We've covered this before when looking at ROI and financial controls - they know things are under control because they set limits on spending and they set and monitor progress towards business objectives. The details of the decisions being made don't matter as long as the results are driving things in the right direction to meet the business objectives and spending remains within limits.

This is hard stuff by the way. This is not the conversation you start off with when talking to execs (unless they are fantastically self-aware). This is the sort of thing you lead into by focusing on one of the other things we have looked at - ROI, financial control and resource management. All these are really about control so by fixing them, you are addressing the control issue. Or at least starting to address it.

To really address control requires a conscious mindset change from the exec (and often their management) and that is a difficult thing to do. Those other conversations will give them some tools to use and will raise awareness that a change in mindset is needed. As a coach, it's your job (and often your biggest challenge) to help and support them through that change.