For successful automation projects, lean on subject matter experts and historical data to develop the most accurate project estimate possible.
Many who have been in automation have likely encountered a project that ran far astray of estimates. This can cause the project team undue stress, derail the project schedule, and harm working relationships with clients and co-workers.
A successful project demands accurate estimates. Consider these tips to develop accurate project estimates and to make the project go more smoothly.
Decompose the project
The first task before estimating a project is to decompose the project into smaller parts. Identifying each of the major deliverables typically does this; however, this is where some projects fall short.
After defining the major deliverables, the estimating team or project manager will sometimes apply a number of hours to the deliverable. This approach poses inherent risks, especially with large deliverables. To create an accurate estimate, it is necessary to break the work down into the activities that it takes to create it.
For example, to build a house, first the foundation is built and then the walls are put up, and so on. However, to understand what "building the foundation" entails, the activities required to complete the task (selecting the bricks, procuring the bricks, mixing the mortar and then laying the bricks) must be detailed. These smaller tasks are easier to estimate.
How far should a project be decomposed? The 8/80 rule is a good starting point. The rule states that the lowest level of defined project activities be no less than 8 and no more than 80 hours in duration. If a task will take less than (around) 8 hours to complete, then it's likely there's too much detail. If a task will take more than (around) 80 hours, consider decomposing that section into multiple, smaller parts.
Once a list of deliverables and activities is created, each activity can be estimated. Several different techniques exist for estimating the required labor, and our house-building example helps to illustrate them:
- Expert judgement: Using a subject matter expert (SME) to tell you how many hours it usually takes to mix the mortar for a house that size. If an SME is asked to estimate a pre-defined task, he or she will usually provide the best-case estimate. To enhance this estimate (because we know projects do not always follow the plan), ask the SME to offer his/her best-case, worst-case estimate, and most-likely estimate (the "three-point estimating technique"). There are various ways to refine into a single estimate, but a savvy project manager or estimator will know to factor the estimate given the other variables at play in the project. Vague or ill-defined tasks often lend themselves to the worst-case estimate while the most-likely estimates often work for straightforward tasks.
- Historical data: Using past experience to help predict labor hours. If you have built three houses in the last year, you can review how many hours it took to mix the mortar on those three houses and then apply it to the new project.
- Combination of methods: Use multiple techniques to crosscheck estimates. For example, I used the expert judgment technique to estimate all activities and then compared it to previously completed projects to validate the estimate.
Taking the time upfront to properly decompose and estimate smaller parts can greatly increase the accuracy of the overall project estimate—and decrease the stress on the project team, clients, and yourself. What techniques are you using for project estimating?