Promises and Prescriptions
Part 5 - Set a Priority to Eliminate Multi-Tasking

I agree with Frank Patrick whole-heartedly on this. Multi-tasking is a the primary source of waste in professional services firms. We have our attention on keeping people busy all the time. And then we go measure people on that. It's no wonder that those same people are starting and stopping tasks. We've got to stop multi-tasking before we go broke!

At the portfolio level, prioritize and launch projects only at the rate that the system can absorb them. If you try pushing ten pounds of project through a five-pound pipeline, you won't even get five pounds of successful projects through to the end.

Here's what you can do: adopt a hard and fast rule. Staff are not allowed to start a task that is not in a condition to be finished. That way, the only stopping is when the task is completed. Now, the action for you as project manager is to see that tasks are queued in a ready state for work. That's no small task. However, profits will soar and customers will be delighted!

posted by Hal at 10:02 PM | link  |   |   | Google It!  | Email This  |

I so am enjoying Frank Patrick's series on Projects and Promises. Part 4 - Multi-Project & Mixed-Function Multi-Tasking is a gem.

(P)rojects are usually delivered as one of a portfolio of efforts and often share resources with others in the pipeline. It is not unusual for scarce, highly skilled contributors to support multiple projects. The usual response to having a lot of work in one's inbox is to use the squeaky wheel method of prioritization. Whichever project is squeaking the loudest in the morning gets attention for the day, whether the previous day's task is completed or not.

Professional service firms have adopted multi-tasking as a strategy for profitability. Keep engineers very busy and profit will be optimized. Turns out, there's a limit to that.

(B)ouncing back and forth between unfinished tasks in an effort to show progress merely delays all the handoffs involved and wastes valuable capacity in unnecessary set-down, set-up, and "Where was I?" questions at every restart.

Not only do those starts and stops add to the time it takes to do the project tasks, they take away from the quality of the work performed. "Where was I?" is every engineer's lament. Not convinced? Have a look at Franks' simple graphic of multi-tasking. You'll be motivated to eliminate the practice.

For more on multi-tasking read Head-Banging in the A/E Industry.

posted by Hal at 9:20 PM | link  |   |   | Google It!  | Email This  |

This is a continuation of my comments on Frank Patrick's series on Projects and Promises, Part 3 - Work Your Way Out of Rework.

Frank proposes 3 actions that we can take to reduce our contributions to the rework on projects:

• Get rid of task due dates
• Develop a clear understanding of the major interdependencies between tasks
• Question assumptions about "necessary" handoffs between tasks

I'll focus here on the first prescription. How could we possibly get rid of task due dates? Of course tasks have due dates. Right? Maybe not!

Frank's prescription is in the usual context of project schedules. If a task is on the schedule, then it has a due date. He's not recommending that we change that. Instead, only put those tasks on the schedule that represent promises to the customer. He says, The only dates that count are those promised outside the project. He urges us to leave all other tasks off the schedule to avoid the predictable missed task starts and finishes. I certainly agree with Frank; the only dates that matter are the ones promised to the customer. But the schedule does more for us than just represent customer promises.

Behind Frank's prescription is the phenomenon of projects Goldratt called dependence and variability. By leaving tasks off the schedule essentially allowing the task dates to float, then the project manager isn't creating dependence. Or so the argument goes. But project tasks are inherently dependent one on another. When we tasks off the schedule we don't have visibility as to what we want to do.

Let me offer an alternative:

This is a kind of just in time planning. Plan in levels: the master schedule includes promises to clients and milestones, the phase schedule answers the question, "How will we do the project?" the six-week look-ahead plan resolves the constraints in the schedule at activity levels, and finally the weekly work plan includes the tasks that are promised for the coming week. Those task promises are made by the performers and include completion dates. This allows others who are dependent on a task completion to plan their work and make their promises.

Curious about how well this works? The Lean Construction Institute did a study of the reliability of task completions one week in advance. The average was about 50% -- about half of the tasks that people said would be completed were actually completed during the week. After putting people on the last planner system of production control™ reliability soars to greater than 80% to the day promised. With that level of reliability you won't need to be concerned with rework due to task due dates.

Read more on committing at the last responsible moment.

posted by Hal at 9:57 PM | link  |   |   | Google It!  | Email This  |

Reference Papers

Project Resources

Books

Bookmark These Weblogs

Yahoo! Groups

RecipRoll