Sunday, November 27, 2011

Scrum, a new framework for modern slavery - part 1

The company I joined recently as a software developer is applying Scrum in a very rigorous way. The usage of this project management methodology has spread like fire in the startup world. Indeed, it increases significantly the productivity, specifically in software development. One can claim that it achieves this result through its innovative and intelligent project management implementation, and because its cycles of incremental change approach is particularly adapted to the startups needs. But does it do so also at the expense of the workers? Does it put a hidden cost on them?
Through a series of posts I will describe my own point of view and try to expose the Srum drawbacks and hidden pressures. The 1st post is about the Scrum planning poker.

The Scrum planning poker.
The scrum poker takes place at the beginning of every sprint. It consists in assessing the difficulty of each story in the sprint in order to create the planning. The difficulty is measured by an abstract unit of points.

Procedure: Each team member picks up a poker card with the number that represent his estimate for the story and keeps it face down. Then everybody show up their estimate by turning simultaneously their card over. If everybody has the same estimate there is an immediate consensus. If not, the team discusses until a consensus is reached.

Justification: it avoids "anchoring" (one member, usually the project manager in standard waterfall methods, biasing the thinking of the other members by giving an estimate before everybody else).

Reality: firstly, although the method claims to measure difficulty specifically to avoid giving estimates of time, in the real world people do an implicit conversion between difficulty and time. Secondly, because nobody wants to look like a slowbrain by giving a higher estimate than the others, everybody will tend to give a lower estimate than it would normally do. So, on average, the group will usually agree on a underestimated difficulty. Thirdly, because it is a consensus, it will take authority over, or at least have a higher degree of credibility than any individual subsequent estimate. Consequently, the assignee that requires more time than was estimated (unless there was an issue unforeseen during the planning) will look as a dumbass. So everybody will strive to get the task done under the estimated time/difficulty and everybody will therefore work in overdrive. To summarize, the method is using people's pride to squeeze more from them.

Hmm, maybe I should put my IT-pride aside sometimes.

1 comment:

  1. Whatever method is used, the company's goal is to try and squeeze the most out of its employees. Sometimes, the intention is good (when the company realizes happy employees = loyal and productive one), but the goal is always the same...