#NoEstimates summary of podcasts
#NoEstimates summary of podcasts
A Panel Discussion on #NoEstimates:
Sometimes you have to start with an extreme to get the
discussion
1. If it is under 3 months, not worth to detail it, just do it well and quickly.
2. Can I afford it (high level)? This is budget.
3. Deal with surprises later, they are coming for sure.
4. Counting stories better than estimating them.
5. Acceptance criteria as an alternative to estimating for sizing and complication discussion
6. Don’t mindlessly apply estimates. What is the need for it? High level product cost. Will estimation help? If yes, how precise can we be?
7. Don’t do it until you have to.
8. Your plans are not reality, don’t believe them.
9. Late is against value not against the plan. Are we late to deliver value to customers on time or to meet regulatory demand? Or are we late compared to our estimates?
10. How can we do it differently? If there is an estimate, teams won’t challenge themselves, they might fall back on their estimate and just work to try to meet it.
11. Estimates at worst creates an illusion of control.
12. Is there a better way to serve you without estimates?
13. Predictability from stability, not from predictions.
14. We can’t be better at estimates, there are problems we cannot fix (not trying), then produce inaccurate estimates.
15. Publishing it creates a commitment, which is seriously wrong.
16. Our interactions using estimates incorrectly is the problem.
17. Moving from “when we can get the value?” To “how can we deliver value faster?”
18. Focusing on the needs of the estimates takes you to the people instead of the technicalities. That is where the need and progress is.
• Planning can be seen as sharing goals, where sometimes people believe they finished the work when they have planned it.
• The illusion of estimates is comforting because it reduces our perception of risk and unpredictability but that is very dangerous.
Lessons from #NoEstimates Critics
“Estimates are natural, ubiquitous, useful, and
unavoidable in practical life and in business. Estimates are an important part
of the process of collaboratively setting reasonable targets, goals, and
commitments within an organization. The process of estimating, in and of
itself, has by-products and benefits. Given that a rational estimating process
is an integral part of making decisions in the presence of uncertainty, it is
hard to understand why anyone would state that a desirable goal is to push
forward into limiting estimates; down to zero where possible.” –Peter
Kretzman http://www.peterkretzman.com/2014/09/24/the-case-against-noestimates-part-1-introduction-and-common-sense/
• Help make decisions.
• Estimates means no planning, too extreme.
• Continuous deployment and work is highly visible=> planning horizon is short. Won’t need estimating. An alternative is estimate vs. actuals.
• How can we improve estimates? Results in smaller and less estimates and risk.
• Estimating shouldn’t be used as a stick to punish people and use commitments to box people in.
• Work (experiment) should shift the estimate (hypothesis) and keep improving.
• Estimating vs. goals. Negotiating estimates, estimates vs. ambition. Use “how much can we get done by ….?” Instead of “can we deliver this by …?”
• Planning to fail, estimation abuse: Negotiate -> cut corners with quality or push -> Aggregate a bunch of estimates. If one fails, all breaks down.
• Spending too much time on estimating and improving it, is it worth it? {turns to statistical and becomes better with predictability}
• Three types of software development:
- Product: lean start up, less estimating, fast
- Service: no way you can estimate that, Kanban style, rapid turnaround, requests in- answers out.
- Contracting: Product as a series of projects make some sense for estimating. Can be broken down to types.
• Low tolerance for variance then estimates are needed.
• Think of changing the cycles of work instead of forcing estimation on ambiguous and complex kind of work.
• The point is to ask questions, no estimate is just a provocation to do so. Do not stop questioning to keep improving and tailor what’s best for you at that specific time.
• Most abuse of estimates happens on execution level.
#NoEstimates with Vasco Duarte:
http://noestimatesbook.com/ by
the speaker “Vasco Duarte”
• Using data to prove the lack of need for estimation
• Number of stories is a better approach for predictability
• Story: 1 idea in 2 weeks was implemented in an innovation sprint that add a feature that didn’t go through the governance and added .5 million to revenue.
• #NoEstimates= making concrete business decisions on data (such as number of stories) not on guesses.
• If we want to be ethical as an IT industry to do what we can to scale to bring higher quality of deliverables and value.
• Estimates become targets especially when it is an X-theory mindset.
• All improvement wave is a response to bad management. If all managers where perfect, no need for improvement.
• Agile is a Developer centric methodology.
• Focus on Value. Value comes first.
• Agile doesn’t live in the iron triangle (Scope, Duration, Cost) that is needed to be known upfront.
• Nice story about a CEO who is going to clients and telling them about no estimates and offering to collaborate (34:20). Not all clients would accept that, but see who might accept it.
• Cost the estimation process, how much is estimation costing… Data that are billed. This can be used to show that this is not value oriented and can get customers to the side of less estimates.
• If you cannot do it reliably, do not do it. If you use it for abuse, do not do it.
• How would we define value? Translating strategy->metrics, if there are no way to measure, can’t learn and improve.
{If you do not trust the team, then you have bigger problems (theory X).}
Comments
Post a Comment