Pay me now or pay me later...

Recently sat in on a planning meeting with a client… the technical lead on the development of new packaging for an advanced semiconductor design broke-down a macro task called “Packaging Feasibility Study” into ten micro tasks. He then proceeded to define their names and the scope of work for each task, we then determined the dependencies between tasks, then we asked, “What is the duration of each task?”

Big pause… then, “I will need more time to give you the answer to that question.” He went on to say, “I need to look at the target Tape-out date (i.e. the point at which the design is done for the device)” and “Work backwards from that date in order to determine how much time I have available to do each step in the Packaging study.” Further he said, “I really have no idea how long each step will take, because it is based on the learning that I discover from the initial steps.”

We stopped and explained that we don’t plan this way… FTTM Planning works from the other direction. We determine the end state outcomes of a macro task like “Feasibility Study” and then work out the steps to achieve the end state deliverables. We plan based on what is needed to get the job done right, not from the time that is made available to us. If we plan from within the constraint of the available time, we will never see the gap until it is too late. Further, this sort of planning does not work when you are doing pathfinding discovery work when much is unknown.

Rather, we explained that we work forward - we estimate how much time each task SHOULD reasonably take, or in other words, “How much time do you need to get it right?” This is a different questions than, “How much time do I have available?” We said…

“We want to encourage everyone to estimate based on the time they think they will really need to do the right job on a task, not based on the time that is available.”

The later will initially look shorter, but it will eventually take much longer, because of rework to do what was really needed to get it right in the first place. It is the, “Pay me now or pay me later” problem. FTTM Planning wants you to “pay for it now,” and if you run out of money, let’s find that out now instead of later when we have also run out of time.

The reason for this thinking is that we have to find out now what the schedule gap is, which is why we need to know the real durations, not the “work backwards from the tapeout date” durations. These “work backwards” durations will always hide the problems, which will eventually show up later and cause the big unexpected schedule slip.

One of the other managers chimed in and said, “Let's see the damage now and then make adjustments to pull-in.” This kind of thinking is the difference between FTTM Planning and traditional planning.

Also see:

Accelerate Pain
For the problem of not knowing future durations, use the Learning Cycle concept: Learning Cycles