These notes are based on the talk #NoEstiamtes by Allen Holub
His take is to have a discussion, not to have a discussion
Estimation leads to disfunction, it leads to situations where one is working unecessary hours without needing to Leads to irrational schedules, people burn out which results in poor work As soon as you put estimates on the scene you cannot have sustainable pace because people work to deadlines
We guess on how long a piece of work takes, we then get held to it and businesses make decisions on this
Steve Mconnel puts percentages on estimates to keep them true, people don’t reallly understand what percentages mean
Chaos Report - 80% of projects either fail or are so late that they are considered to be failures Chaos Report leads to panic Chaos report is indicating that we are poor at estimating stuff We need to accept the fact that we don’t know how to estimate
The real software crisis is that we don’t know how to build anything that is useful Building the wrong thing on time and on budget is a failure
Agile was meant to introduce us to producing something useful But many of us got agile wrong and it was just a disguised way of describing a functional requirement The story has become the central way of describing software [9:00]
Story points were to obscure duration so that certain managers would not put pressure on teams But people would still reverse engineer story points to time The numbers are still leading us to doing estimates
Velocity was (ideal time / actual time) Velocity brings you back into the realm of estimation Points per sprint is as oppressive as any other time based estimate
How long will it take to estimate that warp drive… Estimation is a ritual We are hoping that estimation will achieve results but it will not The other motivation around estimation is fear, how can I spend a large amount of money in a large project without estimating Alternative is to do smaller things and re-assess, build small and revise
Don’t think about estimates, rather think of projections… It’s hard to predict the future…
Time spent in estimation of the top items is mostly waste because it is well known and not that much Time spent on the top 30 are wasteful because they change based on what we discover on the top 10 Time spent on the remained of the items is a complete waste
Fredrick Taylor - Scientific Management - this is where it all went wrong
with TPS give the watch to the people who do the work, get them to do the best job that they can do and figure it out
The only reason we do estimates is because our bossess force us to do them
Good things from lean
- The developers control time
- The business side decides what to build
- Nobody is ordering anyone around
- If you don’t have the schedule, do you need managers (managers control schedules)?
- If you are giving control to the people doing the work then the managers that remain are there to support people
What is waste, if it doesn’t put value into the customers hands it’s waste Estimation is waste, we need to eliminate waste whenever possible You need to have acurate projections, we are still running a business
We need to make projections to make two decisions 1) Shall we kill the project 2) Should we add more people
With estimation you don’t know you are in trouble until things slip so late that you are in trouble
Explains CFD’s - [27:13] We can make projections with CFD’s
We over complicate things - or we can do things simple - the results is going to be the same…
The point he is making that counting stories is almost good enough, it doesn’t involve any estimation at all
Where we need to focus is on prioritizing… Best way to plan is a story map… His explanation of a story map isn’t great, it’s different to how I’ve understood story maps - I believe mine is more useful The focus is still on slicing vertically - deliverable value
Summary Estimation is entirely waste There are simpler ways of projecting Estimation is actively destructive to the organization, it creates a disfunctional culture. Counting stories is enough for projections If you always do the most important thing firs then you always have something of value Build the most important thing first Call to arms: Refuse to do stupid stuff
The message I got out was that is a lot of things we do that are “waste” We over complicate things, there are simpler processes that are sufficient Identify the most important thing and focus on that
