NEWS AT SEI
This library item is related to the following area(s) of work:Process Improvement
This article was originally published in News at SEI on: September 1, 1998
Congratulations, you've just been promoted. You get to lead the new project we just won. You will have six engineers, but two of them are half time for a month or two. You can hire four more. The delivery date is nine months. What do you say? Most engineers would say, "Gee thanks boss, I've always wanted to run a project and this sounds like a great opportunity. I'll give it my best shot, but the date looks awfully tight." If that's your answer, you lose!
Who owns the nine-month date? When your boss offered you the promotion, whose date was the nine months? It was the boss's date. But, when you said, "Boy that's a tough date, I'll do my best to meet it," whose date was it then? Yours! And don't ever forget it! You have just bought the ranch. Even though you had no intention of doing so, and you didn't even have time to think about it, bang, it hit you out of the blue. And there you are, the proud owner of a budding disaster.
So what else could you do? It turns out there is plenty you could do.
Projects usually get in trouble at the very beginning. They start with impossible dates, and nobody has time to think, let alone do creative or quality work. All that counts is getting into test, and the rush to test invariably produces a hoked-up product, a poor quality program, a late delivery, an unhappy customer, and disgruntled management.
While the promotion looks good at first glance, it is only good news if you handle it right. To have any chance of a successful project, there are things you must do right now. Before we talk about what to do, however, let's discuss the causes and consequences of this all-too-common situation.
The only way to manage development work is with pressure. Managers know that relaxed projects rarely succeed. Projects can get into trouble by endlessly studying alternatives, not making decisions, or loading up a product with nice features. These all seem like good ideas at the time, but without pressure months can go by and nobody notices.
So expect pressure. Managers know they must push, and you can expect them to keep pushing, at least if they are awake and doing their jobs. Their only questions are
Schedule pressure causes all kinds of counterproductive behavior. People don't plan their work, they rush through their designs, and they don't review their products. The big push is to demonstrate progress to management, and the only thing management recognizes as progress is getting into test.
Unfortunately, few managers really understand that this is the worst possible thing they could push for. Most software engineers know that racing to throw a product into test is a mistake but they don't know how to fight the pressure. Then they feel compelled to rush through requirements and design and to skip everything else but code and test. When they do, they know what will happen. And of course it does.
While you can always say management was unreasonable, you will be responsible for being late and producing a poor-quality product. Everyone can easily blame somebody else, but do you really want to spend your life this way?
You know intuitively when a schedule is too aggressive. You have a queasy feeling in the pit of your stomach. You can sense the dates are wildly optimistic, and you know this is a disaster just waiting to happen. So you tell the manager, "That schedule looks awfully tight to me," and the manager responds, "But that is the date in the contract, or that's the date the customer demands, or that's the date the boss committed." So you say, "Ok boss, if you say so, I'll try but don't be surprised if it takes longer."
Now that you told the boss, if there are problems, you can always say I told you so. Unfortunately, that won't help. The minute you walk out of the room, it is your date. You may have been worried, but you took the job didn't you? Why did you take the job if you didn't think you could do it? If you don't settle the issue right now, you will be the goat, no matter what you say about the date.
So you must take a stand. Most engineers are so focused on the job that they don't think about what the manager is saying. When managers say the delivery date is nine months, they are making a bid. And you bought it without a counter offer. You'd never buy a house or a car or a boat this way. You'd debate the number.
Think about it this way. Management has just said, "We have this key project, and the best date we think you can make is nine months." If you don't counter with another date, they will hold you to nine months. Unfortunately, if you just guess a later date, they will ask you why. And if you don't have a good answer, they will either ignore your date or get somebody who won't argue.
Management doesn't know how long the job will take, and neither do you. If you knew, and if you could convince them you knew, they would accept your date. If the project really will take 12 months, the last thing most managers want is a commitment to deliver in 9 months. You work for them, and they will also be held accountable for your schedule. They could easily lose their jobs if your project fails, and all you would lose is a chance to have another disaster. That would probably be a relief, at least after this project.
You must start by convincing management that you know how long the project will take. To do this, however, you must know yourself. This, it turns out, is not very difficult. It takes time and some hard work, but when engineers make careful estimates, and when they use historical data to make these estimates, they are generally pretty accurate.
The way you determine the date is to make a plan, and to make a very detailed plan. Since this can take a lot of work, and since you want your team to be part of this planning process, you need to get your new team to help you. This actually is the best possible approach. It will not only produce the best plan, but the team will then be in a far better position to do the work. They will also be committed to the schedule.
Then, when you have the plan, go back to management and tell them what the date really is. When they argue with you, as they will, take them through your plan. Show them as much detail as they will sit still for. Walk them through the numbers, and the task lists, and the historical data you used for comparison. Talk about product sizes and productivity rates. Show them enough to really convince them that you know what you are talking about.
Once you have made the sale, and management accepts your plan, stop selling and move on to the next subject. They will probably want to talk about alternatives. What would the date be with more staff, or reduced requirements, or a phased-version delivery plan? They may want you to present the plan to higher management, or to the customer. In fact, you will probably have to walk more people through the plan, so keep it dusted off, and make sure your backup is solid. Expect people to find any chinks or inconsistencies. Remember though, these are estimates. So tell them what you think and why, but remember that no one knows as well as you do how long the job will take.
If anyone can convince you that your estimates are off, be willing to make adjustments. As long as they have actual historical data to back up their opinions, and as long as these data are relevant to your project, consider the new facts. Under no conditions, however, make any such decisions on the spot. Any schedule change requires careful study and team agreement. So don't change your estimates without data and the time to review them with your team. Remember, almost all initial plans are tight, so don't cut your schedule without a very good story that the team agrees with.
So, when management says the date is nine months, the way to answer is to say, "That looks like a great opportunity boss. Let me get the team together and we'll take a look. We'll make the best plan we can, and be back to you in a couple of days."
If you think the schedule will be longer than management wants, don't argue about it right now. You don't have the ammunition to win that argument, and all that would do is convince management that you have a bad attitude. If they think you are out to prove their date is wrong, they won't let you make the plan. Remember, they don't believe you can give them a good date. After all, nobody has made good schedules before, so why should you be first?
So start with a positive attitude, and really drive for a better date. In fact I once had a team come back with a five-week better schedule than management had asked for, and they're still holding to their plan. So give it an honest try, but then get the data to defend it.
I have coached many development teams on how to do this, and it always works. Of course, we have started by softening up management, and we have been there to help at the beginning. But teams are surprised at how well this approach works and, after a good start, they can usually continue working this way. Management also quickly discovers that informed and committed teams do vastly better work. While most plans come in with longer dates than management wanted and management always asks lots of tough questions, when teams have good plans they can defend them. And when they defend their plans, they always convince management. Best of all, they end up working to their schedule not management's.
So addressing the schedule problems up front really pays off, both for the engineers and for management. While it takes management a few days to get over the shock, they will end up with a software team that knows what they are doing. These teams can report their progress against realistic plans, deliver on the agreed schedules, and produce fine products.
So you need to know how to make a plan, how to present this plan to management, and then how to defend the plan when it is attacked. Then, once you have done all this, all you have to do is develop the product. But at least you will have a date that you and your team agree you can meet. And, most important, you will have a well-thought-out plan to guide the work.
While these methods are not difficult, they are not obvious. If you need help in how to do this, one place to look is at the various Personal Software Process (PSP) references that explain how to make individual plans.(The PSP references are my two books A Discipline for Software Engineering and Introduction to the Personal Software Process, both published by Addison-Wesley. Additional references are in the February, March, and April 1998 issues of Crosstalk.)
The SEI teaches PSP courses, and there are a growing number of university and commercial courses available. We are also introducing the Team Software Process (TSP) which shows teams of PSP-trained engineers how to handle this planning and commitment process. TSP introduction also walks teams through a launch process that produces their detailed plans and negotiates their schedules with management.
While this all sounds logical and simple, everybody's situation is different. If you remember two basic principles, however, you should be able to handle almost any case. First, you are paid to do what management wants you to do, so don't refuse a direct order. Second, always be willing to make a "best effort" but don't commit to a date without a plan. If management appears totally unreasonable, however, read up on negotiating strategies before you get in over your head. (Probably the best reference on this subject is Getting to Yes, by Roger Fisher and William Ury, Houghton Mifflin, 1981. I also summarize some key points about negotiating programming issues in Chapter 12 (Power and Politics) of my book, Managing Technical People, Innovation, Teamwork, and the Software Process, Addison-Wesley, 1997.)
Most managers will be reasonable and respect your desire to make a plan, but occasionally one won't listen. While he or she could be totally unreasonable, it is more likely that higher level managers are applying heavy pressure that your manager is unwilling to buck. You should certainly try to turn such situations around, but that is normally very difficult. If you can't make headway pretty soon, get out from under your gutless manager as quickly as you can.
In writing papers and columns, I make a practice of asking associates to review early drafts. For this column, I particularly appreciate the helpful suggestions from Linda Gates, Alan Koch, Warren Morrison, Mark Paulk, and Bill Peterson.
In Closing, An Invitation to Readers
In these columns, I plan to discuss software issues and the impact of quality and process on engineers and their organizations. I am, however, most interested in addressing issues you feel are important. So please drop me a note with your comments and suggestions. I will read your notes and consider them when I plan future columns.
Thanks for your attention and please stay tuned in.
Watts S. Humphrey
Watts S. Humphrey founded the Software Process Program at the SEI. He is a fellow of the institute and is a research scientist on its staff. From 1959 to 1986, he was associated with IBM Corporation, where he was director of programming quality and process. His publications include many technical papers and six books. His most recent books are Managing the Software Process (1989), A Discipline for Software Engineering (1995), Managing Technical People (1996), and Introduction to the Personal Software Process(1997). He holds five U.S. patents. He is a member of the Association for Computing Machinery, a fellow of the Institute for Electrical and Electronics Engineers, and a past member of the Malcolm Baldrige National Quality Award Board of Examiners. He holds a BS in physics from the University of Chicago, an MS in physics from the Illinois Institute of Technology, and an MBA from the University of Chicago.
The views expressed in this article are the author's only and do not represent directly or imply any official position or view of the Software Engineering Institute or Carnegie Mellon University. This article is intended to stimulate further discussion about this topic.
For more information
Please tell us what you
think with this short
(< 5 minute) survey.