Why you suck at estimating – a lesson in psychology(blog.muonlab.com)
blog.muonlab.com
Why you suck at estimating – a lesson in psychology
http://blog.muonlab.com/2012/04/12/why-you-suck-at-estimating-a-lesson-in-psychology/
8 comments
Yeah you might be getting burned by these, but even if you are very careful to avoid them, estimating software development in particular is pretty much impossible for anything beyond trivial. The fractal complexity of software means there are often unknown unknowns waiting inside various libraries and architectures that you plan to use. I realize other domains have similar problems (large construction projects come to mind), but to paraphrase Dijkstra, no human endeavour spans a greater scale in orders of magnitude than programming, and it's all just a heap of text with minimal transcription cost!
Given the nature of software, I think that the only sane process is an agile one, where the estimate and scope is under continual refinement as the project develops and becomes better understood. Sometimes business requirements dictate otherwise, but that will always lead to a suboptimal solution.
I think that may be the key reason why I prefer startups to other forms of programming work: in a startup you can not afford to do anything for ceremony or to CYA, the focus on optimizing the product must be relentless.
Given the nature of software, I think that the only sane process is an agile one, where the estimate and scope is under continual refinement as the project develops and becomes better understood. Sometimes business requirements dictate otherwise, but that will always lead to a suboptimal solution.
I think that may be the key reason why I prefer startups to other forms of programming work: in a startup you can not afford to do anything for ceremony or to CYA, the focus on optimizing the product must be relentless.
Agree 100%. For those who haven't done Agile estimation before, I strongly recommend Martin's 'Agile Estimation and Planning' book - quick to read and lots of good stuff, in particular:
* Write your requirements as 'User Stories' - discrete units of functionality that are as independent as possible and can be developed and tested as a unit. This gives you maximum feedback on how far along you are against your original scope.
* Estimate complexity, not time. For the reasons this guy gives, people suck at estimating how long something takes. However they are better at estimating how complex something is, particularly relative to other similar things. This means you can estimate each story using 'story points' (an arbitrary measure of complexity, relative to other stories), and then be reasonably confident that two stories rated the same are going to take about as long to write. There is some variation here, of course, but averaged out over all your stories you can get quite accurate results, and usually increasingly-accurate as you go along.
* The people making the estimates should be the ones that are going to be doing the work! Seems obvious, but many places have one person (often a manager or team lead) estimate a task, then have the developers build it, and then the manager gets upset that the developers are 'lazy' (i.e. his estimates were rubbish). As the author says, use planning poker with your development team. This can seem like an Agile kool-aid gimmick to those who haven't used it, but it is a very effective way of getting everyone's input. The best developers are often those less likely to volunteer information in a group setting, and this is a great way of getting their opinions without putting them 'on the spot'.
* Write your requirements as 'User Stories' - discrete units of functionality that are as independent as possible and can be developed and tested as a unit. This gives you maximum feedback on how far along you are against your original scope.
* Estimate complexity, not time. For the reasons this guy gives, people suck at estimating how long something takes. However they are better at estimating how complex something is, particularly relative to other similar things. This means you can estimate each story using 'story points' (an arbitrary measure of complexity, relative to other stories), and then be reasonably confident that two stories rated the same are going to take about as long to write. There is some variation here, of course, but averaged out over all your stories you can get quite accurate results, and usually increasingly-accurate as you go along.
* The people making the estimates should be the ones that are going to be doing the work! Seems obvious, but many places have one person (often a manager or team lead) estimate a task, then have the developers build it, and then the manager gets upset that the developers are 'lazy' (i.e. his estimates were rubbish). As the author says, use planning poker with your development team. This can seem like an Agile kool-aid gimmick to those who haven't used it, but it is a very effective way of getting everyone's input. The best developers are often those less likely to volunteer information in a group setting, and this is a great way of getting their opinions without putting them 'on the spot'.
(Really) useful resources on estimating:
- http://www.mountaingoatsoftware.com/presentations-estimating...
- http://www.mountaingoatsoftware.com/books/1-agile-estimating...
(I'm not affiliated with Mike Cohn but took his estimating and planning course and must say it's worth the money).
- http://www.mountaingoatsoftware.com/presentations-estimating...
- http://www.mountaingoatsoftware.com/books/1-agile-estimating...
(I'm not affiliated with Mike Cohn but took his estimating and planning course and must say it's worth the money).
I often find that over-enthusiasm for a task can often be a huge factor in getting estimates wrong: over-enthusiasm will lead to the compound effect of underestimating past experiences, over-confidence, and, when your interest wanes, you end up dragging your feet and spending a lot more time doing the gritty, uninteresting parts of the task.
The other half of the problem is that management (especially nontechnical "PMP" types) suck at understanding the realities of software development and obsess too much on deadlines and "accountability" when they would do better to focus on removing obstacles for the team or, god-forbid, perhaps jump in and do some of the actual work.
Nooooo, please don't let them get involved in the actual work!
EDIT: Maybe I should politely rephrase that: If they're anything like my previous non-technical managers, I'd rather they make the most of their talents elsewhere.
Moving back on topic, the article is worth reading just for the mention of Planning Poker alone. Another moment of enlightenment for me on HN.
EDIT: Maybe I should politely rephrase that: If they're anything like my previous non-technical managers, I'd rather they make the most of their talents elsewhere.
Moving back on topic, the article is worth reading just for the mention of Planning Poker alone. Another moment of enlightenment for me on HN.
Wondering whether this article should credit Daniel Kahneman's "Thinking, Fast and Slow" as these ideas seem to come fairly directly from that book.
http://www.amazon.co.uk/Thinking-Fast-Slow-Daniel-Kahneman/d...
http://www.amazon.co.uk/Thinking-Fast-Slow-Daniel-Kahneman/d...
Scrum attempts to solve many of these problems, rather than just labelling them:
http://www.writemoretests.com/2012/02/estimating-like-an-adu...
http://www.writemoretests.com/2012/02/estimating-like-an-adu...
Take the 'psychology' out and insert 'software development'. We're actually really quite good at estimating things in a psychological sense, it just stands out when we mis-estimate.
We (as a industry) tried it with UML and we found out that it is more work than actually writing the code and it was still useless because there isn't a possibility to check for errors.
Because of this, "the developers" and the "managers" need to reach an agreement on a more emotional level where they can trust each other and fill in the gaps with their expertise and experience.
Just keep on trying to document the requirements more detailed, estimating more factually will be a waist of time and it will lead to bureaucracy and mediocre software.