Confronting the entropy inherent in managing a business

imageEntropy is an expression of the degree of disorder which accumulates in a system over time. Entropy has a natural tendency to increase within a particular system over time, though nailing down exactly what this means can be problematic. Conceptually, it is the amount of additional information needed to determine the exact state of a system at any particular point in time. Unfortunately, as the second law of thermodynamics postulates, it is not possible to measure this quantity in systems without impacting the system being measured.

Entropy is also an unfortunate fact of life on projects, and despite our futile desires for predictability, the interventions we introduce can be futile. Systems theory tells us that as we try to rearrange a system, the system often pushes back, usually in unforeseen ways. For example, the more energy we inject into projects in attempts to measure their status, the harder it may become to try to determine exactly what is going on.

Within most businesses, the amount of work that could be done within a particular time horizon often exceeds the amount which actually can be done. But businesses must explicitly decide which work will be focused on, rather than pretending to allocate without following though. These decisions are never easy, and can be overwhelming when the work portfolio is complex, distributed, irregular, and always changing.

That's why decisions about resource allocation require a careful consideration of options, clarification of commitments, clear decision gates, and timely (re)-allocation of resources as changes unfold.  There are many levels at which this planning, analysis, and execution must be considered:

  • First, there is the context of managing the fabric of the business itself: identifying the mission and strategy to support customers, securing the required financial and staffing resources, managing opportunities, risks, and knowledge, and incorporating historical performance into decisions about new endeavors.
  • Second, there is performing the top-level activities essential to any systems engineering effort; these activities form the 'outer loop' for managing the overall flow of work so that it is performed at the cadence that the business requires
  • Third, there is determining exactly what the work will consists of, how it will be performed, and how long it will take to be 'done'
  • Finally, there are the individual work cells of contributors who must execute the 'inner loop' details required so the plans of the organization can be successfully implemented

Integrated Product Teams, with their organizing principles firmly embodied in integrated planning, are often proposed as a way to solve the problem of resource allocation across these perspectives. The idea behind such IPTs is that they have the decision authority to reconcile and integrate information across multiple perspectives. Yet even when such a construct is adopted, the underlying teams may still struggle to gain access to scarce resources which are only 'on' the team for prescribed periods of performance (and ideally can focus and finish these  assignments) ; addressing resource scarcity involves triage and coming up with band aids for the gaps which exist or arise with time. When an endeavor of multiple projects requires cooperation across such integrated product teams, and they each require orchestration and cooperation many disciplines in order to be successful, the overall endeavor's leadership must find a balance between efficient resource utilization and the time and effort required to constantly re-balance resources in response to changing conditions.

For example, projects nearly always need access to specialists at specific points in time, in order for the critical path of the endeavor to be protected. These specialists by their very nature will have limited availability, high-value knowledge, and higher-than usual unit costs. Yet individual projects who also have budget constraints to adhere to want to avoid bearing the cost of these resources when they aren't using them. As a result, allocating these specialists to the right projects just in time is always difficult, but can be especially problematic at the fuzzy front end of new projects, since the fidelity and timing of when these resources will be needed is often roughly right at best.

To further complicate this multi-level coordination effort, it is well recognized that work tends to expand to fill the time made available to it (a phenomenon known as Parkinson's law. As a result, when slack time is allocated to tasks to account for this coordination, the result is often that it takes more to complete that task than was estimated under 'ideal' conditions. Unless all critical dimensions of the required planning are considered, including those of each performing organization and their interdependencies, disciplines, domains, work locations, and types of funding - the efforts involved in this coordination may be wasted, or even make the problem worse rather than better. 

Too often, project sponsors often invoke overly prescriptive goal-setting, when these patterns are suspected or detected. Yet this may just result in piling on additional work, no matter how well intended, and can often result in chaos, unless there is sufficient 'adult supervision' in place who can discern this dynamic and has options to address these needs within an agreed-to servicing strategy.

The objectives of integrated planning are to provide this adult supervision, by:

  • organizing jobs into work packages that are sufficiently defined, so that the activities necessary to realize target outcomes are actionable, and can be worked in parallel
  • managing the disciplines, domains, and roles required to perform this work at the job level
  • rationalizing the timing, dependencies, and resource requirements for all of these activities, so that a balanced solution is achieved
  • assuring that the input stream of work is actionable before resources are assigned and expected to deliver results

In order to manage any business, the projects which make up the business's portfolio should each have a plan, and the team members on those projects must help evolve these plans over time so that they are useful in steering the collective towards its objectives. Core disciplines are also necessary for reliable execution of these projects, but may not yet be in place when the project is launched. But sufficiently robust requirements are always an important element of securing commitments to perform the work, and provide accountability for honoring those commitments. It is essential to strike a balance across the desired set of features, quality, and schedule requirements, given a resource and business forecast that leaders have determined to be affordable.  

New projects should not be created within businesses (or begun in secrecy) just because someone thinks they have better idea than others do. Instead, candidate statements of work must be managed by a change management process so this balance is maintained within an overall portfolio of projects. Such change management processes aren't meant to throttle work for urgent, important work such as restoration after an outage that is affecting customers. When something is broken, it needs to be fixed, but fixes should take days or weeks, not months. If it takes months, something deeper is wrong. These 'interrupt-driven' tasks should focus on just the restoration of service, rather than trying to re-engineering a new approach in real time without additional authorized resources. Once operations is re-established, necessary 'improvements' can be developed within the context of the organization's other priorities and commitments, rather than just generating more work at the worst possible time.

A project repository should be used by all team members to plan and status their work. Jim McCarthy describes the purpose of sharing this information in this way:

Slips happen a little bit at a time. Slips don't happen at the end of the milestone or project. Slips just show up at the end. But they happen every day. They happen every hour. every time someone has to make a fresh pot of coffee, answer unexpected email, reconfigure a machine, or track down a maddeningly intermittent but catastrophic bug, slips happen.

You have to manage the granularity of development tasks in such a way that you emerge with visible deliverables over short intervals.

The initial successes which such investments can introduce can serve as a powerful example for others to follow. Once the controlled information in this repository is actively being employed as the basis for decision-making, the motivation level of everyone to keep this information up to date will be increased. As these successes accumulate across the organization, team morale will also improve. Performing organizations can then come to trust in these integrated planning approaches, and take pride in their ability to honor project commitments, while achieving a more reasonable work/life balance. As the fidelity of available information in this repository improves, leadership will also be motivated to expand the use of this integrated approach within other parts of the organization. And in this way, order can be restored to the kingdom.

PDF icon Managing the business.pdf27.17 KB