The scientific method is the basis of our modern world. The method involves the steps of developing a hypotheses, planning and performing experiments, and evaluating the results. Completion of these steps does not guarantee a successful outcome on the first try, but has the advantage of eventually converging on a solution (if indeed a solution can be found). Most are familiar (though not necessarily experienced) with these steps, but overlook the fact that unless a core set of disciplines are used throughout these processes, the results produced by the process may not be valid.
Within process improvement communities, Edward Deming taught his disciples how to apply this scientific method through his PDCA framework. The PDCA steps also are self-correcting, and require a similar set of underlying disciplines to be successful. These disciplines help ensure that when observations are made about a phenomenon or an outcome, the results are sufficiently reliable to be acted on, and that those actions will eventually produce what is desired. These disciplines are key to establishing a research organization's maturity in performing the scientific method consistently and produce results over time that can be reproduced by others, during peer reviews. The maturity of an endeavor is thus judged by their ability to consistently apply a process, make valid observations about it, and successfully act on those results, especially as techniques, people, technology, and hypotheses evolve over time.
It should be obvious that planned projects run better than unplanned projects. But how does one determine whether any practice like planning is done well enough? Each disciplines is described in more detail as a set of key behaviors which define the generally accepted practices of planning. This involves managing the statement of work, processes, and risks, and establishing an orderly approach to doing the work among the units responsible for doing the work.The planning core disciplines are Project management, Process management, Requirements management, Risk management, and Organizational alignment. Each of these disciplines receives different amounts of focus across the activities of a project. Yet due to their interactions, they collectively have greater leverage together than they do in isolation. Interestingly, these disciplines are equally applicable to improving how to do something better over time, as well as to doing it just once. These core practices help us treat improvement activities like any other project. They also assure that the underlying foundation for all processes is adequate for effective problem solving to occur. Without this foundation, many organizations exhibit analysis paralysis, and waste considerable resources on relatively unimportant decisions.
During implementation, there are other essential practices that are critical to implementing these plans successfully, including Requirements Development, Decision analysis and Resolution, Supplier Management, Training, Project Monitoring and Control, and of course, performing the work itself. The importance of each of these elements can be made obvious by asking the question "what happens downstream if this process isn't done well?", and following up with the question, "what is our evidence that we are doing this well today?"
Execution accountability is achieved through the 'check' part of the PDCA cycle, which includes Measurement and analysis, Verification, and Validation. Each of these can of course be done adequately or poorly, but this insight brings one other key element into the equation. That is what the definition of 'doing things well' really means. Often, that definition is itself subjective, since these words can mean dramatically different things to different people. That's why a configuration managed definition of terms is relevant to all of these disciplines.
Each of the pages reached by following one of the above links provide further links to still lower-level criteria, which describe the underlying basis of how an independent observer might evaluate whether you are 'doing this well' within the key practices of that process area. Note that these links are only available to registered users, and are based upon the CMMI framework. This should not be interpreted as meaning that the CMMI criteria are burdensome, overly complex, or bureaucratic. The determination of whether behaviors are compliant or not can be difficult, but that is typically the case when there are no written processes in place at all, or when organizations have become used to ignoring them.
Finally, one must be able to act on the information which is collected during project and process execution. This requires the disciplines of Configuration management, Process Assurance, and Causal analysis to be in place. These items are often overlooked, but are essential to being able to focus attention on the areas that need improvement, without worrying that everything else may be changing at the same time. They are the equivalent of controlling the variables of our scientific experiment.
All of this can be overwhelming for an organization to take on while concurrently in the middle of development efforts. That's why organizations often adopt an established maturity model which defines a level-based approach to defining and deploying these key practices throughout an organization.
People often see such models as dictating a particular structure for improvements, but this is untrue. These criteria are to be tailored to an organizational context - the scope of work performed within an organization, and the business goals of that work. The actions necessary to define and implement any solution to a particular problem can be enormously complex - say, when performing health care, continuing to deliver products that satisfy Moore's law within the electronics industry, or decoding the genome. But while each of these is, by itself, really complex, it should be recognized that the progress which has been made (and continues to be made) in each of these efforts is essentially the result of a disciplined application of PDCA, and these underlying disciplines.
No one 'central authority' has to oversee all this work (though accountability does need to be in place); what matters is that the associated organizations work together, with roles and responsibilities that are coherent and well integrated to assure that products have sufficient quality, and that accountability is provided to assure that the necessary activities of these organizations are achieving the desired outcomes. Progress can be made in systematic improvement in even multi-disciplinary ventures, as long as there is a commitment to such improvement, appropriate incentives to understand and pursue required changes, adequate resources to achieve results, and a proper roadmap and infrastructure for implementing the changes themselves. The underlying CMMI practices assure that these requirements are satisfied, which is why they work.
Since process improvement can be overwhelming, and often takes much longer than expected or is desired, organizations should adopt an improvement framework like the CMMI for guiding their long-term improvement. Such a framework provides a multi-year framework for reinforcing best practices, and helps sequence and reinforce the necessary foundations which must be built over time. The self-assessments (and independent assessments) against such criteria can provide a means of tracking progress in implementing improvements time, but only if these assessment method evaluates criteria similar to the CMMI, and if such criteria are applied consistently and accurately.
The CMMI suggests emphasizing adoption of level 1 practices first, before worrying about level 2 practices. Similarly, it suggests implementing level 2 practices before pouring a lot of energy into level 3 practices. This is because if you try to build a house without the foundations, you're just going to have to come back later and try to install those foundations, because the walls will collapse otherwise. You'll then discover how expensive rework can be. That's what makes the application of these disciplines so core to accomplishing scalable and sustainable improvements within any organization.