Helping Your Team Estimate
Estimating is not easy. How much time will something take? Do you mean when it will be done? Are you including weekends and holidays? Or do you mean how much effort will go into it so you can figure out what the labor and other costs will be? My last article “Effort, Duration, and Elapsed Time” addressed the different meanings of these three concepts, which are usually confused.
Let’s assume you are the leader of a team and you want to help them learn how to estimate. You need to keep these concepts in mind and make sure that as you work with your team they know which term you want to address.
The best approach is to estimate effort, the amount of actual work to be done. Then address duration and finally derive elapsed time. Let’s talk about each of these in turn.
Effort
Estimating Effort requires domain knowledge in terms of the work to be done. Is this something we’ve done before, so we can derive our estimate from the prior work, otherwise known as “analogous estimating”. Or maybe this is a brand-new task. Can it be decomposed so that some pieces can be estimated via analogous estimating? For those that cannot be estimated, what does the work consist of? Is there a knowledgeable person (“expert judgment estimating”) who can help explain the work and help with the estimate? Note that this “expert” will probably do it faster (less effort, less duration) due to their expertise.
Or is the task something that can be compared with something else? This approach, sometimes named t-shirt sizing estimate, compares the work to be done with each other. Pick what the group thinks it is an “average” task then compare the other tasks to it. Is it smaller? Larger? How much smaller/larger? Is it 1/2 or 2X? Or 1/3? 4X? 10X? Sort the tasks accordingly. If you are using cards for each task, put the Medium one on the table, then as you compare each of the other tasks, put them to the left (smaller) or to the right (larger). Then make additional comparisons (e.g., is this small task the same as this other small task? Or is it bigger or smaller?) You’ll end with a sorted list.
In t-shirt sizing, the average task would be a Medium (M). Then you would have extra small (XS), small (S), Large (L), and extra-large (XL). You can even have bigger (XXL, XXXL) and much smaller (XXS). But there’s no need to have that many more. Most of the time XS, S, M, L, and XL are enough.
Your team can estimate the effort for the “average” task then multiply the other tasks by the size factor. For example, the team can use the following table. They can adjust the factors as appropriate:
There are other estimating techniques, such as three-point (optimistic, pessimistic, most likely) and parametric (use a formula to try an estimate the effort based on given factors).
Duration
But how long will it take? Each task is competing with other tasks for the people and equipment time. So adding the effort estimates will give you an estimate that is only true if the resources are available (unlimited?). It also assumes no inter-task dependency, a potentially major challenge.
It may also be possible to provide more resources to a task, reducing the duration accordingly. So, if a task takes 80 person hours, having one person work on it will take 80 hours. Having two persons will still have an effort of 80 hours but the duration would be 40 hours (this assumes that there’s no coordinating “tax” that would make the increased productivity non-linear. That is, coordinating between the two individuals could consume 10%, so the duration would be closer to 44 hours.)
“Resource contention” needs to be addressed. What this means is that a person can only work on one task at a time (multi-tasking does not mean that a person works on two things at a time. It means they work on one task until a point when they break from that task to work on another one. Note that there’s a “context switch” cost in the person coming up to speed on the work they were doing in the second task.) If a person is working on more than one task, then duration will extend for each task.
So, duration estimation requires the team to set the tasks on a timeline so that resources are available to do them and dependencies, if any, are considered. They must consider limits on resource availability (e.g., each person can only work 8 hours/day, unless overtime is allowed, and even that will be limited. Equipment cannot be used more than 24 hours/day). Typically, a person is only available 70%-80% of a day due to other tasks such as meetings, paperwork, bathroom breaks, etc. Tools such as Microsoft Project can help address these limits by providing reports that highlight overcommits as well as a resource leveling capability. Note that organizing tasks can be done even for independent tasks (tasks not in a project) as the resources doing the work may be used by multiple tasks.
Elapsed Time
Resources are not typically available 24 hours/day, 7 days/week, 365 days/year. Duration addresses the daily time availability. Elapsed Time then considers that people usually don’t work on weekends or holidays, and that they may take vacation. It also considers conflicting demands on the resources outside of the tasks they own. For example, if there’s a holiday shut-down for the group, this is considered as a holiday. Another example would be if a higher priority task needs to be completed before the team can work on theirs tasks.
This discussion is enough for individual task estimating. Just getting your team to have this knowledge and practice it will make them more effective at estimating how long something will take.
Project Estimating
If the tasks are part of a project, then more work is needed. The tasks need to be organized so that the overall project effort, duration, and elapsed time can be estimated.
The process then is as follows:
- Estimate task efforts
- Organize tasks on a timeline that addresses resource contention and dependencies
- Estimate duration
- Estimate elapsed time
Things to Consider
Estimates are better the closer they are made to the time the work will be done. Farther out, the less accurate they are. So, don’t just estimate once and be done with it. You and your team need to reassess estimates as they get close to doing the work, adjusting for new information and learning.
From a project point of view, the “Cone of Uncertainty” reflects this issue. Whether it is effort, duration, or elapsed time, there’s a +/-40% uncertainty at the beginning, +/-25% in the middle, +/-10% towards the end, etc. Actually, it is probably more like +40%/-20% at the beginning of the project to reflect the fact that there’s a limit (0) to how much a task will take while there’s no upper limit. The graphic below shows an example.
Conclusion
As a leader, you can help your team estimate how much something will take. Note that most team members are optimistic, and their estimates will not reflect reality, leading to friction and embarrassment. As a leader, you need to help them accept the fact of how much they can do without penalizing them for being realistic rather than optimistic.
Estimating is a bit of an art but it can benefit from some of the guidance above.
Once they have enough practice at estimating, you can leave it up to them to estimate while you plan the overall project/program using approaches such as Commitment-Based Project Management/Map Days. See “Simplifying the Planning of a Complex Program/Project”.
Want to know talk about project/program management?
Send me a note: jose@coachsolera.com.
Subscribe to my stories
Interested in more of my Medium stories? Visit and subscribe at https://medium.com/@josesolera/subscribe.
