Quickly–Using a Consultant
We have seen some software organizations move to Agile quickly by bringing in an outside consultant to coach their team into doing it correctly. This can be a fast approach and works if your organization has the money to pay the consultants. It seems, however, that you’ll need someone in-house that knows the Agile and Lean principles to help sustain the transition after the consultant leaves. Simply following a recipe will not work as well as understanding why it works.
Slowly–Do It Yourself (DIY)
The other way is slower, has less risk, is potentially less expensive, and may help you gain more expertise before you face the challenges of scaling while trying to apply this approach with the larger organization. First, read books and blogs from others currently trying it. Next try it on a micro scale, testing pieces of the approach, and encounter some problems where the practice for the exemplar organization does not work for you. Then, study the principles behind both Agile and Lean to better understand why what you have done didn’t always work as expected.
Then try again on a small scale. Use the PDCA cycle to improve. Learn more and study more about solutions to problem you experience. Adjust and try again. For example, initially we were unsure about how to interface Agile with Earned Value Management, and it took a while to test those ideas.
Then go for a larger rollout of Lean-Agile. First, with a single team of strong staff. This may be bumpy initially, but then it gets better. Continue to improve, striving to perfect some of the adaptations you made to the software version of Agile for use in learning experience development.
Next, expand Lean-Agile to multiple teams doing multiple projects. Continue to refine as you learn what sounds better on paper than works in a live development team.
Even internally, we have seen the need to have an in-house Lean-Agile Coach stay involved to help the teams apply the approach consistently.
Whichever way you choose, work hard at it and use the PDCA cycle to test each hypothesis against the experimental evidence. Then adjust and try again.