I wrote the essay and seems it sparked a good discussion in here :-) Can't comment on it all but the point I tried to get across was actually very closely related to the style of thinking most of you promote in here - focus on the outcome not the output, iterate fast and expect uncertainty in both cost and outcome. The principles I present is trying to do exactly that (IMO) while providing the structure to make informed decisions in terms of prioritization and risk.
Basically - Make it short and simple for the team to estimate stuff so they can spend their time focusing on what's important (which is NOT worrying about whether 4 or 8 hours remain before you hit you original estimate and the stress, drop in motivation and even potential blame games that might follow from that) - and be disciplined enough about data to answer all the other questions people will ask anyway. Estimating in time sucks!
And as I wrote, yes you can skip registering time on the individual user story if you find yourself in an environment where that is doable. You will loose a bit of data to make better informed decisions but I'll admit following the 80/20 rule that you probably should not do it if you can avoid it.
Once the structure is set up it actually is quite lightweight (especially for the team but also the person collecting the data, both with and without the per user story time registration) and having at least some amount of real data does in my experience provide better and faster decisions in most contexts.
But anyway, whether you agree or you don't thank you all for discussing and sharing. Never in a million years expected close to 13.000 users from 6 continents to visit my blog in less than 24 hours (and still rising)
Basically - Make it short and simple for the team to estimate stuff so they can spend their time focusing on what's important (which is NOT worrying about whether 4 or 8 hours remain before you hit you original estimate and the stress, drop in motivation and even potential blame games that might follow from that) - and be disciplined enough about data to answer all the other questions people will ask anyway. Estimating in time sucks!
And as I wrote, yes you can skip registering time on the individual user story if you find yourself in an environment where that is doable. You will loose a bit of data to make better informed decisions but I'll admit following the 80/20 rule that you probably should not do it if you can avoid it.
Once the structure is set up it actually is quite lightweight (especially for the team but also the person collecting the data, both with and without the per user story time registration) and having at least some amount of real data does in my experience provide better and faster decisions in most contexts.
But anyway, whether you agree or you don't thank you all for discussing and sharing. Never in a million years expected close to 13.000 users from 6 continents to visit my blog in less than 24 hours (and still rising)
BR Jesper