proactive stance, you might ask questions like this:
- What issues that cause production problems were a result of miscommunications or misunderstandings between the development group and the operations group? Give your assessment of the impact of the single biggest problem on a scale of 1 to 10 (1 negligible impact; 10 a major disruption to business).
- Which of those could have been discovered before the new features went into production?
You can ask free form questions to follow up on interesting answers as well.
5. Interview your list of credible sources.
Keep your interviews short, and record responses carefully. Make sure to account for all the people on the "credible sources" list, even if you were unable to interview them. Otherwise, your manager may assume you were selective in your interviews to bias the case in your favor.
6. Summarize and present the results.
Write a summary report no longer than one page. Start by reminding your manager that she agreed to the list of people, so the report is at her request. Account for all the people on your manager's list, even if for some reason you didn't interview them. Be sure not to omit negative results. Chances are good that your boss will learn of the negative result anyway, and if she hears it from someone else, your credibility is dead. If you show the negative results, it telegraphs that your case is strong. Explain the negative results, but don't dismiss them or argue them away. After all, your manager identified this person as credible, and it won't help to tell her otherwise.
Have the detailed data at hand so you can go over the details if your manager asks for them.
Here are the some of the results from the interviews for the pre-production turnover meeting, which was presented to the VP who wanted to avoid production problems. The interviewer asked the questions listed in Step four. She interviewed sixteen people, each of whom identified the single biggest problem found in a pre-production turnover meeting and then rated the impact of that problem on a scale of 1-10.
This represents each persons' assessment of only the biggest problems, so it's a subset of all the problems. Four people assessed the impact of the biggest problem at 10. One person assessed the impact of the biggest problem as 9, and so forth. For the most part, the development folks identified lower impact problems and the operations folks higher impact problems. (Very interesting!)
This was enough information for the VP to continue the pre-production turnover meeting. "Hold on," you may say. "This isn't hard data. These responses are correlated to any standard of severity." That's true. But the data is related to what the manager said was important to her. And that's what matters when you are building a case.
Acknowledgements: I originally learned this technique from Jerry Weinberg. You can read more about subjective impact analysis in QSM, Volume 2.