Project Management

Please login or join to subscribe to this thread

Estimation Padding and Velocity in Scrum

linkedin twitter facebook  
avatar
Mohit Joshi Germantown, Tn, United States
Hello,

I know one of the foundation of Scrum is trust and transparency. The Scrum master and PO must trust the development team in deciding what work can be done in the upcoming Sprint and how they will do it. The team is Self-organizing and shouldn't be directed by others.

Now, I assume a common expectation from the stakeholders (or Product Owners acting as an representative) is to finish everything the development team committed to in a Sprint. The team realistically may not be able to complete all items but this expectation could lead to development teams under committing the work. They may try to select enough items that they don't get in trouble for being lazy but not so much that they risk not finishing it all. On the flip-side, the development maybe tempted to pad estimates to boost productivity.

If there is an under commitment, wouldn't it impact cost? For example: The value that could have been delivered in say 4 Sprints, might take 6 (2 additional Sprints). And if there is padding, it can show an inflated velocity (I think it would make more sense to consider the value delivered instead).

Has anyone experienced such issues with estimation in real life? And, how do you ensure the velocity is improved without using it as a measure to gauge development team's performance?

Would love to hear and learn from your experience & suggestions.

Thank you.
Sort By:
< 1 2 >
avatar
David Portas London, United Kingdom
Velocity is primarily a tool for helping the team plan what they can achieve in a sprint and for identifying blockers. Real value needs to be measured by results and it is the PO's job to maximise the value of the work done by ensuring the team are always working on the business's priorities. The PO can use sprint reviews and other communication channels to show evidence that value is being delivered through product improvements.

It's true that motivation and productivity can be a problem for any workforce, regardless of the delivery frameworks used. Scrum arguably can help motivate teams because it encourages customer engagement, collaboration and getting regular feedback on the results. Team motivation and commitment is ultimately a function of the culture, working conditions, rewards and may other factors.
...
1 reply by Mohit Joshi
Aug 14, 2020 10:41 AM
Mohit Joshi
...
Thank you David for your perspective.
avatar
Kiron Bondale Retired | Mentor| Retired Welland, Ontario, Canada
Mohit -

Padding and/or taking on less work than could be reasonably completed at a sustainable pace happens when team members have been punished in the past for missing (unrealistic) deadlines.

Trust and respect are two key tenets of successful delivery approaches. Senior leaders need to trust that team members will be trying their best and will have a legitimate sense of urgency about completing their work so team members then will be encouraged to do their best.

Remember that a sprint forecast is just that - a forecast. There could be any number of reasons why less work is completed than was planned. Information radiators such as a burn-down chart provide an opportunity for the team to self-assess their progress and for stakeholders to feel comfortable that progress IS being made. If the burn-down chart stays relatively flat till late in a sprint, the team should be challenging themselves to discover ways of delivering in more of a flow rather than a batch manner.

Kiron
...
1 reply by Mohit Joshi
Aug 14, 2020 10:44 AM
Mohit Joshi
...
Thanks Kiron. Yes, I understand trust & respect should be there.

What I wanted to know is how well it usually works in real world.
Have you experienced padding or under commitment by development teams? And having known that, how did you address that to improve the situation without making the development team accountable for it?
avatar
Aaron Porter
Community Champion
IT Director| Blade HQ Payson, UT, United States
Keep in mind that the team shouldn't stop working just because they've finished all the work planned for the sprint. You should have a healthy, prioritized, product backlog. What's the next item on the list? Can it be finished before the end of the sprint?

Keep track of your actual velocity over two or three sprints and then, during a retrospective, discuss with the team whether to increase their planned velocity. Help them make the decision, whether, or not, to change how much work they can commit to completing in a sprint.
...
1 reply by Mohit Joshi
Aug 14, 2020 10:52 AM
Mohit Joshi
...
Thanks Aaron. Seems like constantly completing all forecast backlog items over multiple Sprints is also not a good sign...maybe the team is under committing then.
avatar
Mohit Joshi Germantown, Tn, United States
Aug 14, 2020 1:55 AM
Replying to David Portas
...
Velocity is primarily a tool for helping the team plan what they can achieve in a sprint and for identifying blockers. Real value needs to be measured by results and it is the PO's job to maximise the value of the work done by ensuring the team are always working on the business's priorities. The PO can use sprint reviews and other communication channels to show evidence that value is being delivered through product improvements.

It's true that motivation and productivity can be a problem for any workforce, regardless of the delivery frameworks used. Scrum arguably can help motivate teams because it encourages customer engagement, collaboration and getting regular feedback on the results. Team motivation and commitment is ultimately a function of the culture, working conditions, rewards and may other factors.
Thank you David for your perspective.
avatar
Mohit Joshi Germantown, Tn, United States
Aug 14, 2020 8:31 AM
Replying to Kiron Bondale
...
Mohit -

Padding and/or taking on less work than could be reasonably completed at a sustainable pace happens when team members have been punished in the past for missing (unrealistic) deadlines.

Trust and respect are two key tenets of successful delivery approaches. Senior leaders need to trust that team members will be trying their best and will have a legitimate sense of urgency about completing their work so team members then will be encouraged to do their best.

Remember that a sprint forecast is just that - a forecast. There could be any number of reasons why less work is completed than was planned. Information radiators such as a burn-down chart provide an opportunity for the team to self-assess their progress and for stakeholders to feel comfortable that progress IS being made. If the burn-down chart stays relatively flat till late in a sprint, the team should be challenging themselves to discover ways of delivering in more of a flow rather than a batch manner.

Kiron
Thanks Kiron. Yes, I understand trust & respect should be there.

What I wanted to know is how well it usually works in real world.
Have you experienced padding or under commitment by development teams? And having known that, how did you address that to improve the situation without making the development team accountable for it?
...
2 replies by Kiron Bondale and Maria Lekha Johnson
Aug 14, 2020 1:22 PM
Kiron Bondale
...
I've usually encountered the latter - the teams I've worked with tend to be over-confident in their ability to deliver and chronically forecast more work than they can complete in a fixed amount of time.

Kiron
Aug 14, 2020 3:48 PM
Maria Lekha Johnson
...
Generally the team under-estimates. I have similar experience like Kiron. Under-estimation is a general occurrence. But given that, estimation is just that - an estimation - which is subject to change and we have reliable justifications and revised ETAs, we can manage.
avatar
Mohit Joshi Germantown, Tn, United States
Aug 14, 2020 10:35 AM
Replying to Aaron Porter
...
Keep in mind that the team shouldn't stop working just because they've finished all the work planned for the sprint. You should have a healthy, prioritized, product backlog. What's the next item on the list? Can it be finished before the end of the sprint?

Keep track of your actual velocity over two or three sprints and then, during a retrospective, discuss with the team whether to increase their planned velocity. Help them make the decision, whether, or not, to change how much work they can commit to completing in a sprint.
Thanks Aaron. Seems like constantly completing all forecast backlog items over multiple Sprints is also not a good sign...maybe the team is under committing then.
avatar
Kiron Bondale Retired | Mentor| Retired Welland, Ontario, Canada
Aug 14, 2020 10:44 AM
Replying to Mohit Joshi
...
Thanks Kiron. Yes, I understand trust & respect should be there.

What I wanted to know is how well it usually works in real world.
Have you experienced padding or under commitment by development teams? And having known that, how did you address that to improve the situation without making the development team accountable for it?
I've usually encountered the latter - the teams I've worked with tend to be over-confident in their ability to deliver and chronically forecast more work than they can complete in a fixed amount of time.

Kiron
avatar
Sergio Luis Conte Helping to create solutions for everyone| Worldwide based Organizations Buenos Aires, Argentina
There is a big mistake on the statement itself. Padding does not exists in estimations. If you people think that is valid then they are DOA (dead on arrive). So, forget to talk about estimation if you consider padding as an option.
...
1 reply by Mohit Joshi
Aug 14, 2020 2:10 PM
Mohit Joshi
...
Thanks Sergio for your frank feedback. Appreciate it.

I agree it may not be ideal but I don't see why padding shouldn't be consider when it can have an impact on the ROI? If the estimates are padded and the team can give 'n' no. of reasons to justify it, the no. of iterations to deliver value can increase, resulting in cost increase.

I may be incorrect in my understanding.
avatar
Mohit Joshi Germantown, Tn, United States
Aug 14, 2020 1:47 PM
Replying to Sergio Luis Conte
...
There is a big mistake on the statement itself. Padding does not exists in estimations. If you people think that is valid then they are DOA (dead on arrive). So, forget to talk about estimation if you consider padding as an option.
Thanks Sergio for your frank feedback. Appreciate it.

I agree it may not be ideal but I don't see why padding shouldn't be consider when it can have an impact on the ROI? If the estimates are padded and the team can give 'n' no. of reasons to justify it, the no. of iterations to deliver value can increase, resulting in cost increase.

I may be incorrect in my understanding.
...
3 replies by David Portas, Maria Lekha Johnson, and Sergio Luis Conte
Aug 14, 2020 3:51 PM
Maria Lekha Johnson
...
Regarding padding, there was a very active thread:Padding: Is this really a "Professional Option" ?

IMO, instead of calling it as "padding" or "buffer", we can call it as "contingency". However, an estimate is only really an estimate and not final number, at least in a technical project. So we should be open to changes in estimations.
Aug 14, 2020 3:53 PM
David Portas
...
Hi Mohit, estimates do not increase cost. Inefficiency, lack of motivation and laziness will cost you more but estimates hurt no-one - unless you are somehow misusing them or believing them to be something other than a forecast. I think that is Sergio's point.
Aug 14, 2020 5:04 PM
Sergio Luis Conte
...
With all my due respect and with the aim to help, Mohit my recommendation, just in case you did not that, take a closer look to estimation theory. But at the end, if people like to use padding, go ahead just my recommendation is being aware about the impacts and risk this type of strategy will make into the whole initiative.
avatar
Maria Lekha Johnson Paris, France
Aug 14, 2020 10:44 AM
Replying to Mohit Joshi
...
Thanks Kiron. Yes, I understand trust & respect should be there.

What I wanted to know is how well it usually works in real world.
Have you experienced padding or under commitment by development teams? And having known that, how did you address that to improve the situation without making the development team accountable for it?
Generally the team under-estimates. I have similar experience like Kiron. Under-estimation is a general occurrence. But given that, estimation is just that - an estimation - which is subject to change and we have reliable justifications and revised ETAs, we can manage.
< 1 2 >

Please login or join to reply

Content ID:
ADVERTISEMENTS

"I would have made a good Pope. "

- Richard M. Nixon

ADVERTISEMENT

Sponsors