Jess Jurick | @jessjurick WordCamp Milwaukee 2015
The Importance of Estimating
Jess Jurick VP of Client Delivery
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Estimation TheoryHow do we think about estimating?
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Predicting the Future?That’s hard.
Humans are bad at it.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Predicting the Future?
Animals are bad at it too.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Instead…Science!
Informed hypotheses constructed to be tested,
revisited, and revised.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Estimation Theory• Understand the problem
• Define the details
• Develop an approach
• Identify the variables (controlled and uncontrolled)
• Present a Level of Effort
• Revisit and revise as needed
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Counting Jelly BeansCan you guess how many there are in these jars?
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
– Vasken Hauri
“The challenge in estimating is to achieve a reasonable measure of accuracy, while not spending an inordinate amount of time on
creating the estimate.”
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Preparing for an Estimate
• Before starting the estimation process, it’s important to determine what we know and what we don’t know. Anything unknown should be researched and clarified prior to estimation, if possible. If not, this must be noted in the estimate.
• More minds = more steps and effort accounted for. Don’t estimate alone! Gather the project team and assemble a list of required steps in order to be as complete as possible.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Defining the Steps• Create a written list of items that need to be
accomplished to complete the work being estimated. Do not attempt to assign levels of effort to these granular tasks.
• Workshop the tasks list with colleagues to improve the chance of catching missing tasks.
• Remember to involve appropriate subject-matter experts (e.g. designers) where appropriate to determine the granular tasks list.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Techniques for Estimating Tasks
• Program evaluation and review technique (PERT)
• Best/worst/average case estimation (simplified PERT)
• SCRUM estimates (small, medium, and large ‘buckets’)
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
PERT Advantages
• Provides an estimate of Level of Effort (LoE) as well as minimum timeline to complete the project.
• Identifies project dependencies, and therefore provides a mechanism to predict overall delays based on missed deadlines.
• Accounts for Optimistic, Likely, and Pessimistic projections. This can be a powerful technique for reminding all parties this is an estimate.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
PERT Drawbacks
• Requires a lot of effort to complete in the full, traditional sense. For example, a true PERT requires a Gantt chart, timeline, multiple people to estimate each granular task with Optimistic, Likely, and Pessimistic LoEs, etc.
• Does a great job of illustrating an ideal timeline, but is just as susceptible to delays as any other means of estimation.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Simplified PERT Advantages
• Provides for an Optimistic, Likely, and Pessimistic estimate for each granular task.
• Results in three estimates (O/L/P) that provide a best/worst/middle case for the overall project LoE.
• Performed by one or more engineer in conjunction with the PM and subject-matter experts.
• Mitigates overruns (or underruns) on the per-task level by averaging them across the entire estimate.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Simplified PERT Drawbacks
• Creates 3 estimates without a real sense of the likelihood of any of the 3 outcomes.
• Using this technique without also considering individual task weighting can skew estimates.
• Smaller group of individuals performing the estimate means it’s more likely that some tasks or LoEs will be incorrectly estimated.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
SCRUM Advantages
• Splits all tasks into Small, Medium, or Large. Commonly, this equates to < 4 hours, 4-8 hours, and 8-16 hours.
• Very simple and quick with some practice.
• Forces granular breakdown of tasks that might otherwise be considered as a single task. This in turn improves the completeness of the estimate.
• Despite the “bucket” approach, averages tend to work themselves out across the project.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
SCRUM Drawbacks
• Relies on a large group of tasks to achieve a reliable average.
• Does not provide an estimate of completion date, only of total LoE.
• Requires that all tasks (even large ones with relatively stable and known LoEs) be broken down into chunks of less than 16 hours each.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Intelligent Estimation
• Effective estimates identify pain points and low-hanging fruit in an effort to guide intelligent decision-making by the client.
• Clearly state all relevant dependencies (internal and external) to prevent confusion down the road.
• Remember: it is neither improper nor inappropriate to directly and clearly state the effect of any potential delay. It is our job as consultants to do so.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Intelligent Estimation
• Don’t be too granular in presenting the estimate to the client.
• Provide roll-ups of the granular tasks that identify effort by phase and deliverable (things the client can see and understand) rather than technical tasks (things we need to understand).
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Communicating Estimates
• Statements of Work
• Shared Task Lists
• Status Reports
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Communicating Estimates
• Shared Task Lists
• 10000ft
• Resource & Capacity Reports
• Project Reporting Dashboard
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
10000ftForecasting Tool
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Project Reporting DashboardInternal Estimate Tracking
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Why does all this matter?
• Estimates inform process and vice versa. If our estimates aren’t good, our processes can’t really be either.
• Estimates that clearly define risk, expected outcomes, and places for further discovery inspire confidence from our clients and, ultimately, in our own ability to deliver with quality.
• Approaching estimates as hypotheses feels better. When something feels better we have more confidence in our ability to do it successfully and, as a result, often do.
Jess Jurick | @jessjurick WordCamp Milwaukee 2015
Thank You
• Work for us: [email protected]
• Work with us: [email protected]
• Tweet me: @jessjurick
• Leave comments: simplecanv.as
Top Related