Open Side Menu Go to the Top
Register
KanBan KanBan

01-26-2018 , 01:21 PM
Does anyone have experience using a kanban board / system in their software development? We're trying to implement it but are getting push back from the business side of things and I'm not sure how to address their questions/concerns.
KanBan Quote
01-27-2018 , 01:37 AM
I've used it for the last 4-5 years. Works great for small teams, not sure how it would be for huge teams.

What questions/concerns are they having?

What do you use now?
KanBan Quote
01-27-2018 , 11:07 AM
Current system: A bunch of tasks are selected to be worked on and expected by a certain date which leads to overtime when the date approaches. Also leads to lesser quality in my opinion as there's a strong pressure to hit the date so people take shortcuts in development or QA.

Concerns: They don't like the idea of work in process limits because they think less work will get done. We don't have any full time QA people so some of the developers were trying to help out with the QA step in the process but that was taking away from development which was frowned upon. Also, they can't say the exact day something will be delivered.
KanBan Quote
01-27-2018 , 01:01 PM
It allows you to easily break up and prioritize tasks
KanBan Quote
01-27-2018 , 03:39 PM
are you using software to organize the tasks?
KanBan Quote
01-28-2018 , 10:03 AM
We're trying to use the Kanban board that is a built in feature of JIRA (our issue tracking software).
KanBan Quote
01-28-2018 , 01:20 PM
These feel like situations where it should be easy to just try it and see how it goes.

It doesn’t even need to be a full team test. Pick a small group of people (ideally working on related things) and have them use kanban for the period of time that everyone else is doing it the old way with the deadline. Then see how it went at the end.
KanBan Quote
01-28-2018 , 02:57 PM
The notion that you can determine which features will be ready by a given date is pure madness, and I am surprised that people still feel this way in 2018. It's the kind of thinking that led to books such as "The Mythical Man Month", a book that was written before I was born, and, spoiler alert, I am ****ing old.
KanBan Quote
01-28-2018 , 07:41 PM
Quote:
Originally Posted by RustyBrooks
The notion that you can determine which features will be ready by a given date is pure madness, and I am surprised that people still feel this way in 2018. It's the kind of thinking that led to books such as "The Mythical Man Month", a book that was written before I was born, and, spoiler alert, I am ****ing old.
its insane. my old team just let us put whatever down for todo and tasks. I would put this:

dev work
dev test
unit tests
build and elevate
sync with business

my new team wants me to break down every single task I do and estimate a time. its like, how the hell am I supposed to know all of the dev work that needs to be done before I have even started on it. and how the eff and I supposed to know how long it is gonna take.

oh but we need to know how long the project is gonna last. and we need time estimates for our "burn down" chart. pure idiocy.

listen, do you want me to try to do the work and fix the bug or create the feature or do you want me to spend hours estimating the fix?

its very frustrating bc my old team and manager were really cool about just getting it done and doing it in our own way. new team leads are just so unreal stereotypical corporate micromanagers that its like a culture shock. as cliche as it sounds, they are just like the guy in season 1 of halt and catch fire who came in to run the coders. well, not quite that bad, but thats what they are striving for.

/rant
KanBan Quote
01-29-2018 , 12:11 AM
Well, at least now I know someone else watched Halt and Catch Fire. I can not ****ing get anyone to watch it, not even my tech friends. I lived in Richardson for years, where most of the first season or 2 takes place, albeit like 15 years after.
KanBan Quote
01-29-2018 , 12:34 AM
I had a boss who really kind of got it, once.

We were working on a big new feature. He wanted to know when it would be done. I didn't want to commit to a date, there were a lot of unknowns. He said "will be it done this time next year?" I said yeah, for sure. "6 months from now" I said yeah, almost certainly. He said "3 months?" and I kind of hemmed and hawed and he said "ok, 6 months"

He understood at a deep level that delivering a high quality feature in a suboptimal time was better than delivering a suboptimal feature in an optimal time. We finished well ahead of schedule

(and we've been sitting on it for like, god, 3 or 4 months because marketing wants to announce it along with some other features)
KanBan Quote
01-29-2018 , 01:22 PM
Quote:
Originally Posted by RustyBrooks
The notion that you can determine which features will be ready by a given date is pure madness
I feel this way too, but not sure how to bridge the gap. I understand the business side feeling uneasy about an 'unknown' delivery date especially when customers are demanding to know when x feature will be released.

Thanks for the feedback everyone. I'm mostly just venting about our process as I don't have any real power to change it.
KanBan Quote
02-01-2018 , 07:00 PM
At my last place we'd write up all the jira's, breaking them down into things that would take 2 days max, then I'd take a guess at how long they should take using t-shirt sizing.

Our system had 4 sizes
xs 1-2 hours = .2 points
s 1-4 hours = .5 points
m 1 day = 1 point
L 1-2 days = 2 points

I'd assign them to the developers and then I'd have the developers look at my guesses and raise or lower them.

Then I'd total up the total amount of work (1 point = 1 man day) divide it by the number of developers I had and give my boss a guess of when everything would be done, and let him know there's a day or two on top of that for testing.

Every day at stand up we'd show how many points were done how many were left and the projected date of completion.

Everything would sit on the kanban board and it was easy to see where we were at.
KanBan Quote
02-01-2018 , 07:43 PM
Quote:
Originally Posted by blacklab

Every day at stand up we'd show how many points were done how many were left and the projected date of completion.
I guess this is the sticking point. They don't want the completion date to move.
KanBan Quote
02-01-2018 , 09:29 PM
Quote:
Originally Posted by Biggle10
I guess this is the sticking point. They don't want the completion date to move.
Does the completion date move now? Or is it just that people work overtime to make it work?
KanBan Quote
02-01-2018 , 10:36 PM
Quote:
Originally Posted by Biggle10
I guess this is the sticking point. They don't want the completion date to move.
if they don't want it to move ask them what they'd like taken out of the release
KanBan Quote
02-02-2018 , 05:33 PM
Read the Phoenix Project to get some good sources for thinking about WIP and Queue theory. It's up to the team members to stop cutting quality corners to make a data, particularly some of the fake dates that pop up in software.
KanBan Quote
02-02-2018 , 08:01 PM
Completion dates don't move now and quality is usually what gets sacrificed as scope is usually not reduced. The QA step is sometimes skipped completely to get everything out on time. Like I said, I'm mostly just venting and don't have any power to change anything.
KanBan Quote
02-10-2018 , 09:50 PM
We use Kanban boards and value demand boards through JIRA for a fairly large enterprise SSD project as part of SAFe (Scaled Agile Framework). We moved to this fairly recently, but it seems pretty good so far (although I can't imagine how much money we are paying these consultants to help implement it!)

I understand the huge reluctance to committing to dates, but in some projects it's just a necessity. If we don't make certain dates to get clients engineering samples, qualification samples, etc. then the whole project might as well get cancelled.

Last edited by KatoKrazy; 02-10-2018 at 09:57 PM.
KanBan Quote

      
m