BUSINESS AND CMMI ALIGNMENT

Here is an assertion many readers will not like: Process improvement, and perhaps process capability, is good for every organization; CMMI maturity levels are not good for every organization. Blasphemy, you cry! Yes and here s why.

CMMI is a model that provides guidelines for improving software and systems development and delivery, acquisition processes, work management, and process and product integration. It is a model for operational excellence. However, it is not a model for innovation or customer relationship management. Thus, CMMI is not a model for every business in every industry. It is not the grand unification process model, which is as much a fantasy as pixie dust, magic beans, and the universal business adaptor. ^[27] Like it or not, CMMI ” as with any model ” is limited in its application in the real world. This will be unwelcome news to many people who think that CMMI maturity levels are the solution to all their problems.

The lesson is this: Before someone in the organization decides that CMMI is the right model for quality or process improvement or before someone decides to pursue CMMI maturity levels, everyone must first understand the core business of your organization. Let there be no doubt that whatever business you re in, there will be some components (process areas or practices) that your organization will be able to use as guidelines for process improvement. However, adapting some components of CMMI to your business is a far cry from achieving a maturity level.

I remember attending my class at SEI to become a SCAMPI Lead Appraiser. During the course of a class discussion on the model scope for a SCAMPISM, there were students who repeatedly asked questions along the lines of well, what if my organization doesn t do testing, then isn t Validation not applicable? or what if we don t subcontract, then isn t SAM not applicable ? C mon, do you mean to tell me your organization never procures any goods or services from an outside source?

The answer to all such questions can be answered by resolving one simple, but very hard question: If your organization doesn t need to use all the process areas of CMMI or, conversely, if all the process areas of CMMI cannot be adopted by your business, why is your organization trying to achieve a maturity level? Why don t you and your leadership simply recognize that only some pieces of the model fit your business and try to improve the organization s process capability in those areas? And at the extreme end of that line of thinking: If there is very little or no relationship between the purpose of the CMMI process areas and practices and the purposes of your business, why is your organization using CMMI at all?

Let s prove the point. Let s take two not-so-hypothetical businesses and explore how well the CMMI process areas fit their business needs. For this exercise, we ll compare a landscaping company with an organizational unit that maintains integrated COTS applications for its customer, a municipal government. For this exercise, we ll only explore the CMMI Level 2 process areas.

Table 4.2 shows the relative translatability and adoptability of the CMMI Level 2 PAs to our two hypothetical businesses. For the Adoption Difficulty rating, we use an arbitrary one to five relative scale with a rating of one (1) representing the greatest difficulty in interpreting and adopting the CMMI practices into the business and a rating of five (5) representing the greatest ease of interpretation and adoption. Using this scale, a total score of 35 (seven process areas multiplied by a high score of 5/PA) would indicate a 100 percent fit or ease of adoption or interpretation between an organization s business and the CMMI Level 2 PAs. A total score of 5 would indicate a strong incompatibility between the organization s business and CMMI.

Table 4.2: CMMI Translatability and Adoptability

Process Area	Landscaping Company		Integrated COTS Maintenance Company
	Adoption Difficulty	CMMI Interpretation	Adoption Difficulty	CMMI Interpretation
REQM	4	Landscaping jobs (projects) differ only in a few requirements such as mowing, trimming, hedging, raking, pruning, etc.	2	It s difficult to determine if one problem or support call equals a requirement or if multiple batched problem reports or calls equal a requirement. People cannot shift their thinking to considering a break-fix as a requirement.
PP	5	The landscaping job (project is easy to estimate and plan. Size (acreage), effort (person-hours), schedule (total time), and cost (labor + equipment fuel + mileage + maintenance/repair) are all easily quantifiable. Virtually all risks are understood and easily planned.	1	Very difficult to determine what constitutes a project. Too many types of problems and their resolution result in an unusable collection of historical performance on which to base estimates and planning. The standard unit of work performed is difficult to determine.
PMC	5	Very easy to know exactly how much area is mowed or how many feet of hedge trimmed per unit of time. Performance indicators such as earned value are easy to produce and are accurate.	3	Progress and status reporting is easy, but not against vague plans based on the difficulty of defining a project. An almost infinite source of risks precludes risks being managed.
SAM	3	A limited and finite number of skills required for landscaping makes vendor/source selection and management relatively easy.	2	How does the organization define supplier or procurement of the maintenance of COTS. Are the COTS vendors actually vendors or customers or both? Can the COTS integrator get COTS changed? If so, are they a customer or supplier or both.
MA	4	Performance measures are very easy to plan and collect: e.g., linear or square feet mowed or trimmed per unit of time.	3	Ill-defined units of work or units of output make measures almost impossible and mostly meaningless. Some efficiency measures are possible: e.g., number of fixes delivered per unit of time.
PPQA	5	Objective verification of process performance can be performed simply by observing the mowed and trimmed areas to ensure growth was cut at the required level.	2	The organization s processes are determined largely by the constraints of its ability to change COTS products. Floating, ill-defined performance criteria makes objective verification difficult.
CM	3	An almost infinitely variable combination of people, hardware, and processes can accomplish the same results, so CM isn t really important or value adding.	5	Technical limitations of the COTS products and their integration limits the number of configurations to be planned and monitored .

Based on the rationale provided in the table, the landscaping company, with a score of 29, appears to have an easier time interpreting and adopting the CMMI Level 2 PAs than the organization that maintains integrated COTS packages, which gets a score of 18. Surprise! A non-IT organization would have an easier time adopting CMMI practices than would an IT organization!

Maybe the comparison between a landscaping company and a COTS maintenance organization is a bit of a stretch. However, there are plenty of real-world organizations which will have a very difficult, costly time trying to adopt CMMI practices to their work. Take, for example, an organizational unit which essentially provides integration and governance (oversight) of a delivered system. The organization outsources engineering, project management, and quality assurance. Even though the organization ultimately delivers the system to a customer, how can it be appraised at all the process areas in a CMMI maturity level? Its core business and core competency ” outsourcing ” only fits some of the model s PAs, primarily SAM and ISM. Sure, the organization could include its subcontractors in the organizational scope of the appraisal, and then you have multiple separate organizations walking away with a maturity level claim from one appraisal, with all of them separately having implemented only parts of the model.

Here s the lesson: You and the leadership of your organization need to ask and find answers to these three questions:

1. What does our organization do? What is our business?

2. What does CMMI do?

3. How well does the answer to Question 1 align with the answer to Question 2?

^[27] In 2002 and 2003, IBM ran a series of television advertisements which poked fun at the belief there could somehow exist a single, magical , universal integrated solution for any and all business problems and goals. The ads were powerful in their ironic comedy .