Chapter 6
A PI Case Study
Psychologists and knowledge management specialists tell us that people learn better and retain more when information is presented as stories. In this section, we put all the DLI phases in play and provide a narrative that describes how one (fictional) firm applied DLI to its particular problems.
This is the story of Vivian, a CEO at a small engineering-services house, who takes the plunge and implements a CMMI-based process improvement initiative.
6.1 Decide (Cycle 1: To Do Or Not To Do)
Vivian is not really surprised the first time she finds the term CMMI in a Request for Proposal. Her peers at other companies had mentioned the trend toward such requirements, so she guessed it wouldn’t be the last. But what she has heard about CMMI from some of the other CEOs in town worries her. Although some suggested that it was helping their business, others openly confessed that they were using it only to be able to qualify to bid on certain contracts. Vivian isn’t sure, but she has the impression that in the latter cases, CMMI wasn’t such a good thing for the company.
So, not wanting to hypothesize her way into a problem that might or might not exist, Vivian calls Bruce—a friend who is, in her opinion, a pretty practical manager who had recently spoken about CMMI at a chamber of commerce Technology Council meeting.
Bruce gladly shares his experience and provided Vivian a few lessons learned. First, he warns Vivian that the experience of adopting CMMI is different for every company and that it is not a “process in a box” that you just buy and implement.
Initially, Vivian voices three concerns. First, is CMMI compatible with the ISO 9001 work her company was already doing, because it had been quite helpful in stabilizing her operational infrastructure? Second, is it flexible enough for her different customer and project size contexts? And third, can it really help with project management practices?
Responding to the last two concerns, Bruce tells her that CMMI covers a broad number of process areas, including project management, but that it is important to plan carefully and not try to do everything at once. She should choose the things that are most important to her to begin the improvement.
Bruce then recommends that she adopt CMMI because of her company’s primary business: providing engineering services for large system integrators. He points out that with CMMI, she and her customers could speak the same project management language.
Vivian responds that it sounds to her as though CMMI has a set order in which you have to work on your improvements, without concern for the particular kind of business.
Somewhat sheepishly, Bruce responds that indeed, a set order was the traditional way of approaching CMMI, but his authority-averse organization used an alternative method that was based on a business analysis of its needs and where CMMI addressed them. As it turned out, those topics that seemed to have the greatest value for his company also turned up in the traditional list.
In response to Vivian’s question on the cost of his initiative, Bruce offers to provide the data and encourages her to make sure her cost tracking is sufficient to capture the costs, because it was one of the first items his management wanted to see.
Vivian thanks Bruce, hangs up the phone, and considers what she heard:
1. CMMI seems flexible and has broad coverage.
2. CMMI would meet her need for improving project management.
3. She could estimate her cost by extrapolating from Bruce’s data.
4. She could tie the improvement activities to business value.
5. CMMI is becoming more common within her customer base.
After reviewing Bruce’s data and coming up with an initial cost estimate, Vivian decides to start a process improvement initiative using CMMI.
6.2 Decide (Cycle 2: What To Do, Where, And When)
Having decided to go forward, Vivian finds a local SPIN (Software Process Improvement Network) group and attends a meeting. Even though it started out primarily as a software-focused group, as CMMI has expanded its focus, so has the SPIN group. She is intrigued and heartened by what she hears at her first meeting, so she brings Lou, her vice president of engineering services, to the second one. His enthusiasm seems genuine, so she designates him the focal point for the improvement effort. A review on Amazon.com leads them to A CMMI Survival Guide. They both read it and, after discussing the pros and cons, decide to try DLI.
The next week at work, Vivian gets her direct reports together to discuss process improvement and their possibilities; she has purchased copies of both CMMI Distilled and CMMI Survival Guide for them to read as an introduction to CMMI and process improvement knowledge.†
† A very wise decision, at least as we authors see it. We do hope you will follow Vivian’s good example.
Concerned about his depth of understanding, Lou suggests bringing in one of the consultants they met at the local SPIN to do the business analysis and RFA (Readiness and Fit Analysis) suggested by our book. They call Judy, an independent consultant and SEI transition partner, who works with them to put together the right team and schedules two specific events.
The first event is similar to the kickoff described in the pilot toolkit referenced in CMMI Survival Guide. Pilot projects and improvement Process Areas are tentatively selected. Additionally, improvement goals and success criteria (mostly of the “solve this problem” variety) are established.
The second event brings in the project teams that would be involved in the pilot. The first project, led by Jim, has been dealing with unstable, changing requirements. Lynette, the team leader for Project 2, is facing a type of project the organization has never done before and so has no corporate experience to draw upon. At this second event, they do an RFA and confirm that these two projects are decent fits, but that there are obvious risks that need to be addressed. They create a top-ten risk list and identify preliminary strategies for monitoring and mitigating the risks. They agree to revisit the list weekly.
6.3 Try (Cycle 2: The First Pilot)
The pilot begins! Judy works with each project to do a gap analysis using an informal SCAMPI C level assessment. Each team walks through the practices for the target Process Areas, and for each PA, the team lists both the things it does and the things it doesn’t. The team members brainstorm about alternative practices but find only one or two that might really fit the requirement.
When the gaps are known, the teams use the one-hour process description technique to understand the gap between their current process and the PA requirements. Because it is unprecedented, Project 2 uses the “closest” projects it can find to talk about general project management practices and then extrapolates from those the areas where it believes new practices are needed. All use CMMI as a general guide, but Project 1 augments it with some techniques from agile methods that the team members believe can help them with the requirements volatility they have been experiencing.
Judy then works with each project to identify simple diffusion and level-of-use measures to use in the pilot. What she chooses is not a full-scale diffusion profile, but one tailored to the needs of the 20-person project teams.
As luck would have it, during the course of Project 2, Bette, the key engineer, wins the lottery and moves to the Caribbean to open a bicycling shop. The project is disrupted both by jealousy and by the arrival of the engineer’s replacement, a fairly new employee named Mickey. Luckily, the work the team did in the two initial events and the gap analysis provides direct support for bringing Mickey up to speed. Having come from a company that had used CMMI before, Mickey is comfortable with the terminology and is able to contribute almost immediately. The change in staffing does slow the effort, however, and the project manager is worried that by the time new practices are defined, the project will be beyond the point of being able to use them. He brings up his concern to Lou at the biweekly meetings they have to discuss his project. They telecon with Judy to decide what, if any, changes they should make to their approach. Because the first area that they tackled was configuration management, Judy recommends that they continue with the improvement track, because much of the activity will occur in the testing phase of the project, which is still more than two months away.
Both projects have periodic meetings focused on the process improvement goals and status. Project 2 has also added PI progress and issues to its daily standup meetings. From these meetings, several issues arise, including difficulty in understanding the new processes by those who had not attended the one-hour workshop.
6.4 Analyze (Cycle 2: The First Pilot)
About halfway through the life cycle of each project, Judy gets the teams together with the management sponsors to talk about what’s working and what’s needed.
One thing that comes up is that there are no templates for helping the teams do any of the things Judy is asking them to do. In response, Lou charters a subteam to work on this with Judy. Keeping in mind that they are also thinking about ISO 9000 registration, he charges them to make sure that whatever they do is compatible with the ISO standard. Unfortunately, Judy doesn’t work much with ISO 9000, but she recommends someone with both CMMI and 9000 experience who can help them set up a PAL (process asset library).
The process asset team sends one of its members, Athena, to a class on information mapping for procedures. She comes back with a bunch of templates to use for Web pages that appear to be applicable in Project 1.
As the teams continue, the process asset subteam takes the work done so far and converts it, using methods from the “developing useful process guidance” section of the Survival Guide book and the informationmapping knowledge that Athena has gained from her class.
At the second progress meeting, the adoption progress measures are reviewed, and two more transition mechanism gaps are identified. After some initial investigation, it’s decided that the teams should solve these individually. The management team, now calling itself the PI Steering Group, authorizes the expenditures needed. At this point, both teams are able to provide some qualitative benefits.
6.5 Commit (Cycle 2: The First Pilot)
By the third progress meeting, the PI Steering Group and the projects are feeling that this is worth taking farther. They look at how much money they’ve spent so far on consulting and in-house effort (that’s the one thing that Judy got them to track on spreadsheets, because their internal effort-tracking system wasn’t set up to handle detailed effort allocations) and figure that they can afford about twice that amount per year as their PI budget.
Lou is set up as head of the steering group, and Vivian agrees to sponsor and resource the broader institutionalization of the assets developed in the pilots.
6.6 Reflect (Cycle 2: The First Pilot)
Judy runs project retrospectives with each of the projects. For the top five problems identified, the teams use the CSI technique from the Survival Guide to get a better idea of possible root causes of the problems and some possible changes that could prevent them next time.
One of the main things teams talk about is the need for a repository of artifacts and lessons learned. This would initially serve the people involved in the improvement projects but would ultimately support the projects using the new processes. The subteam that worked on process documentation standards volunteers to work on a PAL project. They again look to the Survival Guide to help them understand the requirements they need to establish before looking for a solution.
Lou suggests that the steering group also conduct a retrospective to think about what worked and didn’t work from the group’s perspective. They decide that an internal staff focal point for improvement would be useful, and they ask Athena, one of the most respected systems engineers (who went to the info-mapping class), to take the job.
She agrees, but with the caveat that she can have access to some of the company’s organizational development staff, because she’s primarily a “geek,” in her words, and doesn’t understand how to do all the change management stuff that Judy always talks about. Although she’s read many of the books that were recommended in the Survival Guide, she doesn’t feel comfortable attempting the activities without some experienced help. Lou agrees and adds Sue, an HR person, as a permanent consultant to the team in charge of internal professional development. Sue is ecstatic. She has a strong background in OD and has complained about not getting to use much of that background in her current role.
The other request Athena has is to add Sean to the team, at least part time. He’s the IT/finance guy who is responsible for all their accounting and effortrecording systems. Athena has some ideas about making those systems friendlier to measurement and wants Sean to have a good understanding and commitment to what’s needed. Lou agrees to this as well. The minutes from the reflection meetings are reviewed, agreed to, and slated to be added to the PAL when it is available.
6.7 Decide (Cycle 3: What’s Next)
Judy helps the projects figure out how much they can expect to get done, based on the kinds of issues identified in their original business analysis. Vivian decides that most of the identified problems fit with CMMI Maturity Level 2 (ML2), so she tentatively sets achieving that as the first model-based goal. She also commits to lead by example and solve their problems related to CMMI with CMMI wherever possible.
Judy and Lou set up another CMMI-based Business Analysis with a broader set of the engineering and management staff. They use it to validate the first analysis and to see whether any different priorities come out. Sure enough, a few things they hadn’t known about need more attention than they thought, although they’re still primarily ML2 issues.
New pilot teams are selected, and one person from each previous team is included to help with the transfer of learning from one project to the next.
Lou recommends that they go forward with establishing some kind of PAL; the team is chartered to initially organize what’s been done so far. As soon as that’s finished, they are authorized to look for an external solution that will serve the business as it grows in terms of improvement activity and staffing.
Vivian decides to review progress every quarter in general and in six months with Judy to see how close they are to being ready for an external appraisal. She admits to Judy that although the improvement is the foremost goal, she thinks attaining ML2 will be beneficial for the corporate image.
Vivian sees Bruce at one of the chamber of commerce meetings several months after their first conversation. He asks about what she’s done with CMMI and how it’s working for her. She admits that it was slow going at first, but that momentum has picked up, and the teams that are working in CMMI-enabled projects seem to be working more effectively than the ones that don’t have access to the assets they’ve been building. She’s most worried about the cost of consulting and the cost of the external appraisal she wants to do in the new year, if they’re ready. Bruce agrees that funding to get external expertise is one of his biggest continuing challenges, although he also says that as people inside the company get more understanding of the model, the reliance on external consulting does go down.
After the meeting, Vivian reflects on how different a conversation she had with Bruce this time in comparison to the first time they talked about CMMI. She can only hope that the next “new thing” to come along in their world is as fruitful as CMMI has been so far!