17. Lets break that down
Chaos Report 1995
For the same combined challenged and impaired projects, over one-third also
experienced time overruns of 200 to 300%. The average overrun is 222% of
the original time estimate. For large companies, the average is 230%; for
medium companies, the average is 202%; and for small companies, the
average is 239%.
Time Overruns % of responses
Under 20% 13.9%
21 – 50% 18.3%
51 – 100% 20.0%
101 – 200% 35.5%
201 – 400% 11.2%
Over 400% 1.1%
~68% of
projects 51%
or more late!
18. Let’s continue to break that down
Chaos Report 2009
Average cost overrun: 45%
Average time overrun: 63%
Chaos Report 2011
Average time overrun: 63%
20. Source Gartner survey of project failure, 2012
Failure, means total
failure, not just late
21. Of the large systems that are comple
ted, about 66% experience schedule
delays & cost overrun
Source: “Project Management Tools and Software Failures and Successes” by Capers Jones
Crosstalk, the Journal of Defense Software Engineering
23. Agile project: 40% failed of
challenged
Source: Project success survey by Scott Ambler
24. 17 percent of large IT projects go so
badly that they can threaten the
very existence of the company
Source : McKinsey & Company in conjunction with the University of Oxford
Type of survey : Study on large scale IT Projects
Date : 2012
25. More data coming, sign-up at
NoEstimatesBook.com to receive
this report when available
27. In 1986, Profs. S.D. Conte, H.E. Dunsmoir, and
V.Y. Shen proposed that a good estimation
approach should provide estimates that are
within 25% of the actual results 75% of the time
--Steve McConnel, Software Estimation: Demystifying the Black Art
36. In 1986, Profs. S.D. Conte, H.E. Dunsmoir, and
V.Y. Shen proposed that a good estimation
approach should provide estimates that are
within 25% of the actual results 75% of the time
--Steve McConnel, Software Estimation: Demystifying the Black Art
49. After just 3 sprints
# of Stories predictive powerStory Points predictive power
The true output:
349,5 SPs
completed
The predicted
output: 418 SPs
completed
+20%
The true output:
228 Stories
The predicted
output: 220
Stories
-4%!
50. After just 5 sprints
# of Stories predictive powerStory Points predictive power
The true output:
349,5 SPs
completed
The predicted
output: 396 SPs
completed
+13%
The true output:
228 Stories
The predicted
output: 220
Stories
-4%!
51. Q: Which ”metric” is more
accurate when compared to
what actually happened in the
project?
61. All dates within 3 weeks of
each other in a 38 to 42 week
project (a margin of ~8%)
62. Data used with permission
from Bill Hanlon at Microsoft
”At that point, I stopped
thinking that estimating
was important.”
Bill Hanlon:
http://bit.ly/BHanlon
63. In 1986, Profs. S.D. Conte, H.E. Dunsmoir, and
V.Y. Shen proposed that a good estimation
approach should provide estimates that are
within 25% of the actual results 75% of the time
--Steve McConnel, Software Estimation: Demystifying the Black Art
In this presentation you have seen examples that far outperform what
estimation specialists consider a ”good estimation”. In common they have
one way to look at work: #NoEstimates
64. #NoEstimates testimonial
The deviation between estimated and
actual velocity would have been
approximately 15% lower if we would have
used #NoEstimates.
We have analyzed data from 50 Sprints…
…at no time the story point based
estimation was better than #NoEstimates.
65. Principle #9
Don’t bet your company on poor
track record methods, use methods
with a proven track record
aka: Hope is a bad management
strategy
67. The 7 Step Journey To NoEstimates
1. Start using Story Points
2. Stop estimating tasks
3. Limit the calendar duration for
Stories and Features
4. Reduce the number of
allowed estimates (say, 1,2,3
and 5 only)
5. Track your data
6. Use your data
7. Simply count the number of
stories
If you are bored with those endless estimation meetings
You are bored and demotivated by the long, low-value meetings that end in endless discussions about the size of features that are not yet fully defined.
If this is how it feels to present your estimates to your stakeholders
If you want to focus on making customers happy, instead of producing even more software
In 2014, JR Central reported that the Shinkansen's average delay from schedule per train was 54 seconds. This includes delays due to uncontrollable causes, such as natural disasters.[15] The record, in 1997, was 18 seconds.
Their trick: Build a system that performs, don’t try to extract performance from an old system that was not built to perform.
In a waterfall project no one is in control at the end (show bug curve at the end).
Much overtime will be needed to get this product out, and that will be on your backs and on the backs of the product development group (be it software or other complex product). This typically leads to a pattern that some have called “Death March” (concentration camp picture), and that’s how it feels. It feels like we are marching to a concentration camp (rant about lives destroyed by this approach).
You have a responsibility to avoid this pattern in your place of work, and you can do it. Here’s how.
And this is only one of the reasons why Estimates don’t work. There are others.
Accidental Complication
My friend JB Rainsberger talks about Accidental and Essential complication. In a short video available online he explains
We use relative estimates (story points), but…
When was the last time that you worked on a perfect code base, where the cost of the accidental complication (how messy the code base was) was either zero or the same multiple of g(e) as when you implemented feature A?
He concludes that often the cost of accidental complication – dealing with organizational and technical issues specific to that code, and that organization – dominates (yes, dominates!) the cost of a feature. So if Accidental complication dominates the cost of a feature, and it cannot be predicted, what does that say about Estimates?
Here’s another reason why estimates fail….
Some people also take queuing very seriously…. (wait for reaction).
I bet you do to. Have you fixed a bug that had been waiting for months in JIRA?
So queues of work affect greatly the estimated time to complete. But wait time in queues are also unpredictable!
One question comes from understanding that what we are trying to do here is learning to live without trying to impose control on the development system. The fact is that what makes sense to do will change as up learn more. It does not help to set a target outside a system's capability.
In other words, if a system is stable and reliably churning out a certain output, it is of no use to set higher (or lower) targets for that system.
Conversely, if the system is not stable, setting a target is useless because you cannot predict reliably enough the output of the system.
Let me give you an example. Remember Office Space the movie that came out in 1999 (what a fitting year) and starred Jenifer Aniston, Ron Livingston and others?
Chaos report, 2004
Source: http://www.csus.edu/indiv/v/velianitis/161/ChaosReport.pdf
eWork and eBusiness in Architecture, Engineering and Construction: ECPPM 2012
Sources:
http:// people.senecac.on.ca/david.bath/BCS590/Powerpoints/05ch.ppt
eWork and eBusiness in Architecture, Engineering and Construction: ECPPM 2012
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
What was the most successful project in this chart?
Success is not linked to meeting your schedule
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
You come to the office and your boss catches you sitting down. ”You’re late Peter!” ”Oh sorry Mr Lumbergh. Traffic was really bad this morning, yeah”
”Sure, Peter. I’ll let this one slide. But say I would need you to deliver 10 stories in the next sprint. Do you think you can do that?”
Did you see what he did just there?
Milton knows best how to react to this kind of situations.
Right, now back to the story.
What Mr. Lumbergh did was Estimate Bargaining. Here’s what it means in practice….
This an example of what I call Estimate Bargaining, only one of the 4 Estimate Dysfunctions I describe in the NoEstimates book
One question comes from understanding that what we are trying to do here is learning to live without trying to impose control on the development system. The fact is that what makes sense to do will change as up learn more. It does not help to set a target outside a system's capability.
In other words, if a system is stable and reliably churning out a certain output, it is of no use to set higher (or lower) targets for that system.
Conversely, if the system is not stable, setting a target is useless because you cannot predict reliably enough the output of the system.
Let me give you an example. Remember Office Space the movie that came out in 1999 (what a fitting year) and starred Jenifer Aniston, Ron Livingston and others?
Let’s take a look at one example of how we can use #NoEstimates in a real project and what were the results...
Here are the questions that I started with...
Here are the questions that I started with...
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Agile is alive and kicking. Take #NoEstimates and experiment! Be Agile!
Just last week – as if prepared on purpose for Oredev…