Lato Sensu Management

Last update: 02/06/2016 13:53

The “No-Estimates” Paradigm

We more and more come across questions about what is now called the “No-Estimates” paradigm, which we are defenders of. People either think it's nonsense or that they will never succeed having this paradigm accepted by Top Management, leading to more problems than solutions. Some even think it is so stupid that it does not deserve any bit of thinking, any afterthought. What they generally do not grasp is the rationality behind the paradigm, in other terms what it tries to solve knowing some brutal facts.

No Estimates Are … Estimates

Before explaining the rationality behind the paradigm, we first want to evacuate a very mundane mistake. You will have estimates … but not before you start! We, at Lato Sensu Management, believe you do not need detailed estimates to start with a new initiative, a new program, a new project, … What you need is a vision and to know what you're ready to invest. Waiting for estimates is procrastination!

Instead of “No-Estimates” we would be better off speaking about Effortless Estimates, deferred to the last responsible moment.

What Are We Aiming At? What Do We Try To Solve?

The problem at hand is four-fold. First, nothing gets started before we get detailed estimates, which is nothing else but trying to cover a financial risk. As stated before, waiting for estimates should be considered as procrastination, an excessive delay that has no value because software development is no predictive activity and we're all terrible estimators (see after). Second, as soon as we require detailed estimates we are de facto interrupting work that was already started, even though it has been considered important and top priority before (this triggers the idea of writing a sarcastic novel about how this actually goes in companies). Thirdly, detailed estimates are a misconception, a false reasoning because software development is no predictive process; it is an adaptive process instead. You will NOT know what it will cost before you start building it. Last but not least, we are all terrible estimators, who are permanently 3 to 4 times off-mark, no matter how seriously we take the job.

So, what are we after? What we are striving for is to find a way not to delay valuable initiatives by unnecessary estimates we know, no matter how seriously they were conducted and consolidated, do not bear the sense of accuracy we expect. Trusting such estimates is simply naivety.

On top of this first goal, no less important, we are trying to make Management aware of the hidden costs of work interruptions such as … detailed estimates.

Takeaways

Have a vision
Start initiatives that implement the vision
Know how much you're ready to invest
Accept that you don't know (estimates will come after) and still be ready to go, a bit in the wild
Do not interrupt work you have already started and prioritized. If you do, then accept that (1) people will be late in the current projects (because you have interrupted them), and (2) book the time spent on new detailed estimates on current projects (important and non controllable bias) (3) you incur task switching costs, which are only making things worse for point 2 (4) timesheets become even more useless (because of points 2 and 3) (5) people (including all levels of Management) enter a doom loop because of all of the above, which ultimately trigger even more detailed estimates requests

In Case You Want More …

Back in March 2014, we have written a post about the L(i)VID Formula, a way — not the only way — to quickly prioritize initiatives based on 4 simple parameters, one of the parameters being L, which stands for Labor, a placeholder for estimates/budget. More on The subject with Agile: the L(i)VID formula.