Change programs are notorious for coming up short in achieving the results that were expected from them. Research indicates that only 16% of such change efforts are successful. Yet despite this poor performance, we rarely confront the constraints that cause us these stumbles. As a result, we're stuck with technical debt that is a drag on subsequent performance. It is almost like we enjoy pouring money down the drain.

Change initiatives usually promise revolutionary outcomes rather than evolutionary refinement. When we consider these alternative paths in politics and hockey, a pattern emerges:

When a company starts losing money, or a whole industry starts losing ground because of a new technology, most of us follow leaders who call for revolutionary change—even if no one really knows what change is needed. Leaders who advocate the status quo look like dinosaurs.

This is why tough times produce radical measures, radical leaders, and radical change. The call for revolution sounds weird in good times, but when things are bad, upending the status quo feels irresistible. We rarely think about selling stocks when their value is rising (when we could lock in our gains) but are enormously tempted to sell, sell, sell when their value starts to fall.

Goalies facing penalty kicks... are heavily predisposed to dive to one side or another to save a goal, even though their best odds of saving a goal are when they stay in the center. In one analysis of 293 penalty kicks in elite championship soccer, researchers... found that goalies had a 14 percent chance of stopping a goal when they dived to the left and a 13 percent chance when they dived to the right. The chance of stopping a goal when they stayed in the center was 33 percent. But, like voters and people stuck in traffic jams, goalies facing penalty kicks are drawn to action, not inaction. The analysis of the championship penalty kicks found that goalies stayed in the center only 6 percent of the time.

Under pressure, we are wired to take explicit action in the hope that by doing so, we will achieve better results, or at least will be seen by our leaders as trying to do so. Then, if these results evade us, we often pretend that we achieved them anyway, and move on. Integrity requires us to try to do better than our emotional tendencies.

Any change program involves a period of discovery, learning, and mastery by those that are involved in the change. Each of these phases have predictable patterns which impact our performance in the area in which change is being attempted over time. Attempts to pretend these phases are unnecessary, or to fail to manage these transitions, will only stretch the time that you end up spending in each phase. It is only after teams have successfully navigated through both of the initial phases that the desired change may finally begin to turn the corner and achieve positive net results.  Yet just because a team has then reached the point of basic competency in incorporating such changes into their work practices, it will still be some time before additional benefits accrue sufficiently to pay off the investments required during this period of reduced performance. Any projects which are cancelled before this payoff has been actually recovered must therefore be considered a failure.

Too often, when we suspect the change effort is troubled, we treat the perception of potential failure as an opportunity to turn up the motivational pressure, either implicitly or explicitly. We may not want to admit that our capability to implement required changes is as low as industry studies show it to be. The only way to survive the many mandates for change which seek to close needed performance gaps, and deliver the meaningful results that our businesses need, in the face of escalating demands for improved performance, is for our organizations to become mature in pursuing and implementing these needed improvements:

• the framework used to select from all possible areas for change must be comprehensive and relevant to the organization's work
• the organization's change planning must be practical, effective, and realistic in choosing where, how, and when to make investments
• the infrastructure which the organization puts in place for implementing improvements must reliably develop the competencies which will be needed to achieve the changes needed
• the execution of the change activities must be disciplined and focused
• the measures of effectiveness of the change activities must provide the organization with leading indicators, triggers, and means of recovery for the change efforts

Leaders and individual change agents who pursue change improvement programs are highly succeptible to underestimating the costs of change, and overestimating the benefits of change. There are many reasons for this. They may fail to account for the impact of system size and individual variation on productivity. They may not understand how much of the work would actually be impacted by the proposed change, or what risks might be introduced as a result of incorporating new methods into the work environment. Finally, they often have limited experiences about the area being changed, or their background may not be relevant to the environment in which changes must be incorporated.

Our own perspectives of our personal track record in implementing improvements is often biased, and subject to blind spots, rather than objectively being derived from independent assessments, themselves derived from proven assessment instruments. We also may look at assignments on such change efforts as temporary detours from doing the 'real work' of the organization, be it product development, support, or operations. Such rose-colored views are the consequence of not understanding how changes manifest themselves in systems over time:

The research reveals three reasons for the breakdowns: Time lags between causes and effects make it difficult to see how they're connected; fallible estimates color the chain of decisions that determine a project's outcome; and a bias toward the initial goals prevents managers from setting revised, more appropriate, targets when project circumstances change. Sticking to an initial low budget goal after a project grew in scope, for instance, led subjects to ignore quality assurance, which led to soaring defect rates--and costs.

- Non-Zero

Striving for meaningful change, such as initiatives designed to improve the productivity of an organization, requires us to come to grips with how we build and support our products and services. The perceived benefits from achieving such productivity improvements is  obvious at the highest level of the organization, where it becomes subject to slogans  - reducing waste, pursuing added value, overcoming adversity. But translating such high-level direction into explicit performance changes that can deliver bottom-line results will usually require translation, trial and error. That's why a proven framework can help the organization begin to make its decisions based upon facts, rather than politics. When navigating this territory in areas that have been under constant cost pressures for long periods of time, there may not be much 'low hanging fruit' that can be harvested without investing additional time and resources, which themselves are likely already in short supply, due to these competitive pressures.

When we consider the question of how to demonstrate that we have actually achieved improved performance with facts and data, we must also confront that it is not always better to just crank out more of what we have been producing, without considering the quality ramifications of that increased production. For example, in a software organization, changes that produce more code in a given period of time, but provide releases with lower quality, are misguided, as they just generate more rework in the future. If this additional code that is written is not useful to the stakeholders who need it, or does not have sufficient quality to provide a suitable basis for customer applications or the project's future product evolution, we are not really being more productive when viewed across a lifecycle of system costs. We may be measuring the wrong thing in producing a synthetic benchmark such as workforce productivity in the short term since it may not be a good predictor of benefits for future business results. Measuring 'productivity' in the absence of considerations of value does not make a lot of sense. The most productive engineers may really be the ones who produced products that were most widely used, and had the greatest impact on the organizations that are using them, rather than those who produced the greatest quantity of code.

As a result, it can often be useful to build and calibrate a model of how improvements are expected be realized, since such an exercise can force us to be more realistic about what change will actually take to pay off over time. Deciding how to measure target outcomes, and which factors will actually be required to achieve those outcomes, is crucial to planning the change effort so that it can be implemented with confidence.

In practice, even a metric that appears to be as outwardly simple as lines of code per month will likely be measured quite differently across groups. Unless there is a clear definition of how to measure a line of code and what labor is counted during those months, aggregating such data together may prove meaningless. Even if you were successful in normalizing such differences, and kept the development team stable over time, individual projects themselves will still have significant variability, project to project. This is due to factors such as the experience of the project team with the processes and technologies used, and the quality of the data they are given to perform their work from.

The extensive research that has been performed using the Personal Software Process (PSP) highlights some of these issues. In the PSP, individual programmers keep careful measurements on their performance across a standardized set of programs. As these programs are developed, they also incorporate a number of quality improvement ideas. The PSP experience indicates that at least in the short term, quality improvement ideas at best have no effect on productivity, but can have a significant impact on quality (and thus total system costs) over time.

This highlights a harsh reality: if you cut costs first, your ability to improve quality (and thus long-term costs) will likely be seriously diminished. It also means that any absolute measure of production rates without accompanying measures of quality is meaningless. This is why it can be dangerous to pursue opportunities to increase productivity, without understanding where the leverage points really are. It also reinforces that it is difficult to predict the future, until meaningful data from past performance has been collected. Such foundational work can be very difficult to do if estimates for total resources are scrubbed to unrealistic levels by stakeholders, or are based upon optimistic projections of how work could unfold under the best possible circumstances, rather than being based upon realistic assessments of risk.

There are many decisions and tradeoffs that must be made in order to ensure that the collection and aggregation baseline measurements are suitable for use over time. These include:

• the convenience of instrumentation, collection, and measurement strategies.
• the ability to accurately characterize results for different audiences
• the means of handling incomplete data (including data on project attributes)
• the use of data in actively monitoring and controlling the phenomenon under study
• how hypotheses are developed and validated regarding the data and projects which are under evaluation

On the other hand, if you need a quick win that sounds impressive in the short term, you can exploit Pareto's law, and increase your focus on the 80% of your work that takes 20% of your time. A checklist can help you filter and select that 80% in a pretty straightforward fashion. Additional benefits can also be realized by reduced multitasking and improvements in how the work flows through the work group. But none of these 'work shuffling' exercises will provide required solutions for the remaining 20% of your work, and this will still take 80% of your time, once you confront it. To solve this bigger problem, the costs of poor quality will likely need to be confronted... but at least you may be around long enough for that opportunity to play out.