A client reviewing the history of a troubled technology project portfolio asked, “Why are schedules always wrong? What challenges emerge between planning and implementation?” While there are infinite project-specific answers to that question, there are also general patterns helpful to understand.
The word “plan” was adopted from French, it means “prepare” or “map”. In a project sense, when we plan, we are preparing to perform a project. The products of planning often include a schedule, a map of the steps we believe must be performed and when we predict they will occur. The schedule describes a route through time (a map) that we believe takes us from where we think we are now, to where we think we want to go… and then the work starts, and the adjustments and replanning begin. Why?
No Facts About the Future
There are no facts about the future. Everything you think you know about the future is an assumption, something uncertain that you take as true based on your best understanding of history and the information available today. If you are planning a picnic that will occur in two weeks you might make reasonable assumptions about the weather, based upon knowledge of the region and season, and a recent long range weather forecast – but there is a chance that the prediction will be wrong. There is a chance that any assumption will be wrong.
Planning requires many assumptions about the future. Will team members be available when needed? Will ordered parts arrive on time? Can we get the vendor’s contract signed by the first of the month? Will external events (fire, flood, pandemic, pollical unrest) disrupt the effort? Each assumption introduces a potential error – slight or catastrophic – to your model of the future. Some of these assumptions will prove to be wrong and the errors accumulate to undermine your plans.
Estimates are NEVER “Accurate”
Estimates are forecasts about the size or quantity of something. When we are planning, task estimates are predictions of the amount of work, the quantity of resources, and the amount of time needed to perform a task. Because estimates are about the future, they too are based upon assumptions.
Estimates can be well researched, documented, and credible, but they are never completely accurate. You are well qualified to estimate how long it will take you to go to your favorite market and can likely predict it within a narrow range of error, yet each trip will have some variance. All estimates have some variance from their prediction. Experience suggests thoughtful estimates of well defined, familiar tasks typically vary ± 20%, but outliers can be significantly worse.
To build a schedule we identify tasks, estimate their approximate duration, determine their sequence, commit arithmetic, then graph the results on a calendar or Gantt chart. This gives us an approximate idea of what will happen and when. It helps us know when we will need to arrange for people, equipment, facilities, materials, and funding to accomplish the work.
When work begins, incorrect assumptions and estimate variances emerge. Some tasks take less time than expected and some take more. The complex network previously mapped to the calendar begins to expand or contract with each change, invalidating prior predictions about when tasks would occur and when the resources needed for those tasks were required. This can lead to cascading consequences.
Imagine it is 8:40 p.m. You are preparing a late dinner for guests expected to arrive at 10 p.m. so that you can go to a midnight show. You realize you are missing a key ingredient for your dish. The market is about 15 minutes away and closes at 9 p.m. You hop in the car for a quick grocery run and notice thankfully that you have just enough gas to make it to the market and back. Unfortunately, you miss every light between home and the market, arriving just as they lock the doors. There is an all-night market 15 minutes away, but you really will need to get gas if you are going to make it there and home, and the stop for gas means you won’t get home to start cooking until 10 p.m. Worried that there won’t be time to cook and consume the meal before you must leave for the show, you find yourself with a couple of options:
When tasks are delayed it can cause significant ripples through a schedule that involve trade-offs, replanning, reductions in quality, further delays, and budget overruns. These accumulate and can be amplified as they are mapped forward in your schedule. Imagine the weeklong task scheduled to start next Monday can only be done by one person on your team. Fortunately, her two-week honeymoon doesn’t begin until the following Monday – but then a predecessor task runs two days late and the two day slip may well become a 2 week slip.
Are Goals Clear (and Unchanging)?
The challenges above exist for all plans, even if the definition of the project’s outcome is crystal clear, agreed upon by all stakeholders, and will not evolve during the project. In my experience with complex technology projects, this has never been the case. Objectives evolve. Technology changes. Stakeholders change. Regulations change. With the passing of time, project needs and priorities change.
I’ve seen research that suggests technology project requirements change at a rate of 1%-4% per month. When requirements changes are detected, they may necessitate replanning, revisiting the overall project definition, or even cancellation of a project.
Why Bother Planning?
Our exploration thus far might suggest that building plans at all is a waste of time, but that’s a silly overstatement. Planning and schedules can be valuable even if they are imperfect. If we can build credible plans and monitor progress against them to adjust our approach as needed, most of the time we can accomplish most of what we set out to do within some reasonable variance of cost and schedule. Plans help us anticipate the future and predict and thoughtfully respond to the consequences when variance is discovered.
In addition to helping us anticipate and prepare for the future, planning and schedules help support informed decision making. If we cannot build credible plans, maybe the project is not feasible and shouldn’t be pursued. Without credible (if imperfect) predictions of project cost and schedule, senior leadership can’t know whether a project is worth the investment.
Over the last hundred years we have learned much about the mechanics of planning. We have developed planning tools to help visualize the future, refined our ability to build schedules, learned to better manage assumptions, and improved our ability to develop and document estimates. The fact that our resulting plans aren’t perfect is not an indictment of planning but an acknowledgement that we can’t predict the future with certainty. Problems arise when plans are mis-represented as perfect maps of the future or held to an imagined standard of precision that ignores the uncertainties of forecasting the future.
“No battle was ever won according to plan, but no battle was ever won without one.”