CRITICAL FACTOR 5: FOCUS ON PROCESS IMPLEMENTATION ASSETS

This critical factor was briefly addressed in the section, Work Product-Based Approach to Process Improvement, in Chapter 4 ” Process Improvement Strategies that Work and here we expand on this concept.

The topic is really a discussion of philosophies; it s a discussion of what process-focused people tend to do versus what delivery-oriented people do. Hard- core process people, such as the people you find on SEPGs and at the SEPG Conference (including this book s author), tend to care a lot about how things are done, but sometimes forget that performing a process is supposed to yield something, like a product. It is the means that s important, not so much the end result. Delivery-oriented people, which include people in project management jobs, sales and marketing, and engineers are producers . They want to create and deliver something and how they get that done is not as important as the results. These people are sometimes happy creating things no one needs or wants; they just need to create and produce. These are broad, sweeping generalizations which you may think are unfair, but pop your head up and look around and I think you ll make similar observations.

The importance of this observation of philosophical approach is that it is one of the reasons process people (like you?) in the organization often fail in their mission. It is also one of the root causes of the frequently observed contention between people in process roles and the rest of the organization. It is critical that you think about these differences in the way people see the world if you want your organization s CMMI process improvement effort to succeed.

And while you re at it, think about what business you re in ” assuming you are involved in CMMI or process work in your organization ” and answer this question: If the business of process is defining the way people work and the business of process improvement is helping people improve the way they work (i.e., helping them be more efficient, effective, and perhaps happier ), which business are you in?

Organizations can and often do build processes based on CMM or CMMI. In doing so, they define the way work gets done, but they do nothing to improve the way work gets done. Yet other organizations improve the way people work even though they ve never even heard of CMMI. Again, in which business are you? In which business do you want to be? Choose and choose wisely.

At this point, some of you and especially those of you who hate philosophical discussions, are asking, what does any of this have to do with the critical factor of focusing on process implementation assets? Keeping to one of the principles in this book, I ll show you, not just tell you.

Give the Deliverers Deliverables

If you are in the business of process improvement, then presumably you know your customers and what they want. And if you know this, then you know they want to get their work done as quick as possible, with as much quality as possible, collect their pay, and go home. What are you doing to satisfy those needs?

If you give an engineer a procedure for analyzing requirements (RD), all you ve done is define the way you think CMMI wants her to work. If you give her a checklist or Excel spreadsheet in which she can plug in a few answers and the tool spits out a determination on whether the requirement is acceptable or not, now you ve helped her do her job as quickly as possible, with as much quality as possible, collect her pay, and go home.

If you give a project manager a long, detailed procedure for developing project effort and cost estimates, all you ve done is define his work in terms of CMMI. But give him an MS Word template with macros which enable him to answer some questions about the project and then the project plan practically writes itself, now you ve improved the way he works!

If you give the quality assurance specialists a procedure that waxes philosophical about the merits of CMMI-compliant process and product quality assurance, all you ve done is either irritated them or put them to sleep. But put one-page quality assurance checklists in their hands and now you ve helped them do their job more efficiently and accurately. And, you ve turned them into strong supporters of the CMMI process improvement initiative.

Let the process people build their processes and procedures if they must, but give the deliverers their deliverables. Build process implementation assets, which are deliverables that, by their content, are inherently consistent with the organizational processes so that when people do their work and produce things, they are automatically compliant with the processes without even thinking about them.

Show, Don t Tell

In the same spirit of giving the deliverers their deliverables, there is another critical concept to incorporate into process definition work: the concept of showing people what to do and how to do it, not telling them. This doesn t mean simply replacing process words with pictures; it means demonstrating for people what they re supposed to do, and how they re supposed to do the task. So let s practice this principle together by illustrating the concept.

One of my favorite PAs in CMMI is Measurement and Analysis (MA) because it s such a gorgeous PA, yet it is so often poorly implemented by organizations. The organization s process definition team can easily write a procedure which more or less mimics the MA practices. The result will be a nice looking process definition (GP 3.1; a process) to show to an appraisal team, but which does nothing in terms of helping people perform in this area.

Conversely, you could take the approach which has been enormously successful for Natural SPI s clients . We conduct a two- hour workshop based on Goal-Question-Metric, ^[19] which we have trademarked as GQM Lite. This workshop walks people step-by-step through defining and prioritizing the organization s business goals (MA SP 1.1). The workshop explains how to translate goals into measurable concepts or indicators, which in turn are supported by derived and base measures. We then give them a blank form (implementation asset) to complete for each goal. We also give them a sample form that has been completed to use as a model (implementation asset) so that they can easily complete their forms. The Natural SPI GQM Lite form, which usually takes people less than one hour to complete when they have an example to reference, asks the following questions in natural language:

• Given the goal, what questions would the organization have to ask and answer to know if the goal is being achieved? (MA SP 1.1)

• Given the questions, what are the measurable concepts or indicators which would quantitatively or qualitatively provide the answers? (MA SP 1.1)

• Given the measurable concepts, what are the derived measures that would provide the indicators? (MA SP 1.2)

• Given the derived measures, what base measures need to be collected? (MA SP 1.2)

• Who will collect the measures, when, and how? (MA SP 1.2)

• What are the analysis techniques or methods to be used to analyze the measures? (MA SP 1.4)

• To whom will the measures, indicators, and information be reported and for what purposes or uses? (MA SP 2.4)

Immediately after these forms are completed, the organization can begin collecting the defined measures. Again, we don t tell them how, we show them how. We build a fairly simple Excel spreadsheet ( implementation asset) that incorporates the responses to the GQM Lite form as outlined above. One spreadsheet in the workbook is blank to be used for collecting actual measurement data and a second spreadsheet in the workbook shows examples of rows, columns , and cells filled in with data so that users can see what is being asked for.

Within days ” not weeks, months, or years ” the organization is implementing processes consistent with most of the practices in MA. And here s the kicker : they re implementing the process without a process description document! They can write that later; they ve got real work to do and the process focus people have helped them do it more efficiently and effectively. They can get their work done, deliver products and information, collect their pay, and go home. Ah, success!

^[19] Basili, Victor R., Software Modeling and Measurement: The Goal/Question/ Metric Paradigm, Technical Report, CS-TR-2956, Department of Computer Science, University of Maryland, College Park, MD 20742, September 1992.