To Estimate or #NoEstimates, That is the Question
The #NoEstmates twitter hashtag was intended by Woody Zuill "..for the topic of exploring alternatives to estimates [of time, effort, cost] for making decisions in software development. That is, ways to make decisions with ‘No Estimates’." Based on twitter traffic it has been successful at generating activity. It's a bit debatable as to whether it has really spawned much exploration. In this talk Todd will actually do some exploration using real data from over 50 projects at companies ranging from startups to large enterprises. In addition to the analysis of the data, Todd was able to build a simulation model of the software development process to both replicate the data to and explore the conditions under which estimates add value and when they do not. Based on the findings from the data and the simulations, along with an analysis of the types of business decisions that organizations need to make, Todd will provide some pragmatic advice for estimators and #NoEstimators alike.
Outline/structure of the Session
I have given this session a number of times this year, with the inaugural kickoff given by Sean Dunn at Agile India 2016. It is based on research that Chris Verhoef and I have been doing which got initiated by my introduction to Vasco Duarte at Agile India 2015. Vasco has been collecting data and Chris and I decided to do some analysis on it. We are in the process of writing a paper for IEEE Software.
- Introduction to #NoEstimates
- The use of Velocity or Throughput as a predictor
- What does the data show us?
- How do we simulate this?
- What does the simulation show us?
- What are the key business decisions that teams and organizations much make?
- Given our findings, what is the pragmatic advice to estimators and to #NoEstimators?
- What does the data tell us about estimation?
- If our conditions are different, what does the simulation tell us?
- Given the decisions we need to make, how to we pragmatically proceed?
schedule Submitted 6 months ago
People who liked this proposal, also liked:
Let’s explore the purpose and use of estimates in the management of software development efforts, and consider possible alternatives. Why do we estimate and are we making estimates that are actually useful? We say we depend on estimates to make important decisions, and yet we’re often disappointed by the results.
Why are we so challenged at estimation? Are estimates for cost, time, or effort always needed? Is it possible there are other ways to approach our work? If we didn’t estimate, how could we still be successful at making businesses successful?
The default use of an "estimate-driven" approach is pervasive in software development efforts, and while estimates can be useful, it is worthwhile to scrutinize our use of estimates for cost, time, and effort, and to seek better ways to manage software development projects.
There are a number of things to explore and many questions to ask. For example, do we really need estimates for all the things we are currently using them? Are we getting a reasonable benefit from them? Is it possible to manage software development projects without these estimates, or at least with fewer estimates? Is there a way to prove that estimates are helping us make good decisions?
In this session we’ll participate in some interactive information gathering exercises to see if we can gain a shared idea of our current understanding of the purpose and use of estimates. We will examine the nature of software development projects and explore some real data to shed light on the art and science of software estimation. Our exploration goal is to see if we can work together to come up with some ideas about improving on the traditional approaches to using estimates.