The Case for No Estimation
Drupalcon Europe — Online Oct 2021
No-estimation is not some kind of guerrilla resistance against good governance. To win the support of stakeholders, we need to frame the question to make the business value evident, demonstrating that practices which optimize for throughput result in improved predictability of outcomes.
The heart of science is measurement.
— Erik Brynjolfsson
An estimate is a prediction of future outcomes; it’s a sketchy business for the software game. There are few things that developers like less than to feign certainty about uncertain outcomes. The problem isn’t so much that our estimates are unreliable, but that they a converted by stakeholders into commitments. Stakeholders need governance, but estimates are not always about the time and money they claim to be, but sometimes they are really a sort of power-play, a tactic to shift the burden of responsibility downstream, from the shared responsibility across the business, to the development team.
The no-estimation movement came along as a sort of rebellion against the tyranny of standard estimation practice. Proponents, knowing that they face the monolithic opposition of convention, read out their manifesto like Martin Luther nailing his 95 Theses to the church door. This makes for good theater but it does little for those of use who have to continue to work inside the boundaries of the business. What’s needed is a case for no-estimation that helps dev teams deliver, and also appeals to people who fund the work.
Predictability in outcome is our common ground. In this talk, we’ll frame the question of no-estimation in the context of standard estimation practice, and describe the principles that underly no-estimation, illustrated with example scenarios.
You can think of no-estimation along the lines of “serverless compute”. We all know that serverless functions run on servers, and you shouldn’t be shocked to learn that no-estimation practice doesn’t do away with estimates, it just moves us beyond time and money as the sole object of estimation. This talk will put the principles and practice into a meaningful context.
No-estimation is not some kind of guerrilla resistance against good governance. To win the support of stakeholders, we need to frame the practices to make the business value evident. Time and money are derivative of throughput, so estimation practice that optimizes for throughput results in improved predictability of outcomes.
The skills that make you good at the no-estimation game will bring improved productivity and greater confidence for freelance developers, and can be a real game-changer in enterprise productivity. We’ll show how core practices of Kanban get us what stakeholders really want when they say ask for an estimate: predictability.
Let's agree to define productivity in terms of throughput. We can debate the meaning of productivity in terms of additional measurements of the business value of delivered work, but as Eliyahu Goldratt pointed out in his critique of the Balanced Scorecard, there is a virtue in simplicity. Throughput doesn’t answer all our questions about business value, but it is a sufficient metric for the context of evaluating the relationship of practices with productivity.