A lean journey is a never ending story. At the Lean IT Summit 2017, David Bogaerts and Leendert Kalfsbeek explained where thaey are now: how they involve everyone in continuous improvement and unlock the improvement talent and power of all their people.
They also explained how they help their people keep a daily focus on the most important goals and challenges they know they must overcome in order to create the best banking experience for their customers.
Discover the 3rd episode of ING Lean IT journey, the previous ones and more Lean IT stories on www.lean-it-summit.com!
3. Where do we work?
3
Omni Channel IT
~ 450 engineers
4. 4
What do we want to share today?
1. Our agile / lean journey so far
2. Our current challenge (how it works
in theory)
3. How it works in practice
4. Simple, but not easy (What we’ve
learned)
5. Summarizing (what we wanted to tell
you)
5. Our agile / lean journey so far
5
Some results
Numberof
deploysperweek
Leadtime
Number of agile teams
2010 3 pilot teams
2011 ~80 teams
2013 ~150 teams
2016 ~300 teams
Leadtime
7. 7
What do we want to share today?
1. Our agile / lean journey so far
2. Our current challenge (how it works
in theory)
3. How it works in practice
4. Simple, but not easy (What we’ve
learned)
5. Summarizing (what we wanted to tell
you)
17. …and one more thing
17
We ALL must try and become
exceptional improvers
18. 18
What do we want to share today?
1. Our agile / lean journey so far
2. Our current challenge (how it works
in theory)
3. How it works in practice
4. Simple, but not easy (What we’ve
learned)
5. Summarizing (what we wanted to tell
you)
21. Improving availability
21
Why did we start with this storyboard:
1. Nowadays customers that want to do banking all expect a full online service, anytime,
anywhere.
2. Within the bank we invested a lot in new technologies and did several improvement
projects (not kata’s) to improve the availability.
3. However we saw that the actual downtime was not decreasing but
increasing to on average 500 min per month.
Challenge
22. New technology, old processes
22
We investigated the causes and learned a lot about our current result:
1. 73% of all downtime was caused by our own changes in the systems. Only 27% was
caused by incidents.
2. Of these changes +/- 40% was with technology, designed to be always
available.
3. There were multiple teams responsible for these systems and doing changes
independent from the others.
4. The procedures used to change the systems required downtime for
all changes.
Current result
23. With these experiments we solved the obstacles and realized our target result:
1. Define strict norms for allowed downtime and make agreements to not accept
any exception. This reduced the downtime from 500 min to max 150
min per month.
2. Extend change procedure with specific scenario for new technology. This
reduced the number of changes that require downtime with 40%.
From more than 500 min to max 150
min
23
Target result
24. Accelerate the engineering teams
24
Why did we start with this storyboard:
1. Every quarter the teams deploy around 2.000 applications to production.
2. The majority of all teams deploy multiple times each sprint, some daily.
3. However, deploying a new application to production (89 each quarter) takes still
more than 3 months.
Challenge
25. Its a long journey to build a new app
25
We asked the users and learned a lot about our current result:
1. For all new applications an engineer had to take 114 different steps.
2. Nobody knew why some of these steps were done and if they were all still needed.
3. With these steps it took a team by average 100 days to bring a new application live in
production.
Current result
26. With these experiments we solved the obstacles and realized our target result:
1. Review all individual steps to simplify the engineering journey. This reduced
the number of steps from 114 to 30.
2. Automate the remaining steps that did not fit in 2 weeks. This reduced the lead time
from 42 to 21 days.
3. Make clear agreements with parties outside our tribe (risk, infrastructure etc) for the
steps that could not be automated. This reduced the lead time from 21 to 10 days.
4. Improving the whole: One roadmap for all products to realize the target condition
instead of independent roadmaps per product.
From 100 days to 2 weeks
26
Target result
27. 27
What do we want to share today?
1. Our agile / lean journey so far
2. Our current challenge (how it works
in theory)
3. How it works in practice
4. Simple, but not easy (What we’ve
learned)
5. Summarizing (what we wanted to tell
you)
28. We were improving in the dark
28
We lacked data and we did not truly understand the problems
29. We didn’t truly understand our direction
29
Cascading context is not the same as repeating the company strategy
30. 30
Help! We are going too slow
Find time to analyse problems that must be solved today
31. 31
Help! We are going too fast
And are loosing people on the road
32. 32
What do we want to share today?
1. Our agile / lean journey so far
2. Our current challenge (how it works
in theory)
3. How it works in practice
4. Simple, but not easy (What we’ve
learned)
5. Summarizing (what we wanted to tell
you)