FYK: Kanban Metric Essentials

FYK: Kanban Metric Essentials - Throughput

Рет қаралды 5,786

Күн бұрын

Пікірлер: 15

@Kanifolya 2 жыл бұрын

Thank you for focusing on Kanban metrics. So many resources out there focus on Scrum while Kanban is often overlooked. I would love to hear more detailed examples/case studies about teams you worked with where WIP was set, throughput calculated, and cycle time analyzed and more importantly how having those metrics in place helped you forecast.

@OskarPiano 2 жыл бұрын

The explanation of getting to stable system first and that metrics are meaningful after that is the most valuable thought, remark in those videos so far. It really allows (among other things) for easier buying in team(s) into certain practices for their benefit.

@thomasj7506 2 жыл бұрын

Every time Todd and Dan are together, expect some kind of riffing. On a serious note, I gained so much value in less than 10m as per usual. 🙏🏽

@OskarPiano 2 жыл бұрын

Your explanations are great. They bring a lot of clarity and perspective. I noticed that I particularly miss putting the throughput in context of a CFD chart.

@lolmastergames1891 2 жыл бұрын

Can you guys post a video with step by step guide demostrating how to work out throughput using CFD in Jira please? Actionable agile tool helps but working in regulated companies where getting and approving the tool is a nightmare. I followed instructions on Dan's book Actionable agile metrics for predictabilty and worked out thhroughput however, unsure on the data extracted especially getting to know the team. BTW Ryan and Todd keep up with great work-thanks!

@Chasethe1le 6 ай бұрын

Given that many estimation techniques are viewed as guesses at best based on lagging indicators due to the inherent complexity and unpredictable nature of software development are we fooling ourselves into being convinced other metrics will make is "predictable". I know we aim to optimize predictability but i don't think we should ever fall into the trap of thinking any metric will make any estimate so reliable that we are predictable. To me that's an important distinction because in SW we learn by doing and the fog clears as we do the work. In reference to throughput specifically I do think it has value on its own even in a vacuum. About 8 years ago Allen Holub actually found that projections between story points and just counting stories (sprint throughput) resulted in similar timeframes. Now I'm not a fan of story points but if we can get the same forecasting value out of a simpler process I am all for that. Question is how much further do you take that in the name of predictability when predictability is simply limited in complex work?

@AD-nr5dt 3 жыл бұрын

The more I listen to you guys talk about kanban the more I feel like it's just a more specific implementation of scrum. What's the major differentiator? Is it really a big deal? If so, when would you take one over the other?

@anlamhcmc Жыл бұрын

I have a problem with the throughput metric. Majorly it's related to the size of each PBI. Given that we have 3 developers in a team, and we measure the throughput in a weekly basis. - Week 1: the throughput of the team is 15 (each developer get 1 work item done per day) - Week 2: the throughput of the team is 9 since the complexity of the BPI were increased (each developer just could finish 3 work items this week) - Week 3: the throughput of the team is even lower, it's 6 (each developer just could finish 2 work items this week) - Week 4: the throughput of team got back to nearly 15, which is 12. So how can we make throughput metric reliable in this situation? Thanks everybody

@AgileforHumans Жыл бұрын

I think the key word you used is reliable. Throughput just is what it is I don't think it's worth considering it reliable. It will have variance. As in cycle time, throughput will vary even with our best attempts at right sizing. It's worth the team taking a look at a cycle time scatter plot every now and then to see if there are means to stabilize process a little bit. Now the good news! If your doing forecasting for multiple items, the variance in throughput is considered when doing a monte carlo simulation. It's why we love probabilistic forecasts.

@karenn1526 3 жыл бұрын

😂😂😂 💙 Kanabanga! I echo this question guys. When do we choose Kanban over Scrum and vice versa. I have encountered this dilemma recently with a a team and my observation there is that they could not adopt Scrum optimally because they are not a Product Team. They are a Shared Services Team. They have a multitude of scopes with specific team members taking care of specific tasks ranging from data pipeline development, maintenance, bug fixing in content structures, creation of schemas etc. Stakeholders are mainly internal, and even if they wanted to they could never organize themselves around 1 sprint goal. We agreed Kanban was a better choice. Your thoughts on what is an effective criteria that helps choose?

@deedance3507 2 жыл бұрын

When calculating Throughput does the complexity of each work item does not play any part ?

@AgileforHumans 2 жыл бұрын

It does not. Perhaps this video on right sizing will help: kzfaq.info/get/bejne/lbCeaZqEmbyZmmg.html

@deedance3507 2 жыл бұрын

@@AgileforHumans Thank you so much! Requesting you to please make a video focusing a bit more in detail about the stable system(and other assumptions) as mentioned in Little's law.