Agile - To be or not to be?

Want to work with our team?  Check out http://jobs.dxc.technology! 

This has been my focus for a weeks now, so thought I'd share a bit about what our team is doing, because it's exciting!  In India, I'm looking for three automation engineers, a data scientist, and a machine learning specialist, and there are many other leaders looking to fill other roles as well, with focus on artificial intelligence, catalog services, and more!

Who are we?

Prior to 2016, "Waterfall" was the method for managing projects.  We would develop and release many artifacts by going through a series of phases that typically took months to go through.  We would identify requirements for all the components, then prepare a component design document for each component we wanted to develop, then develop each component, then conduct designer testing for each component, then conduct integration and acceptance testing for all components as a single set, then finally release.  A release took us many months!

In about 2015 though, we started to explore a new way of working...  "Agile" methodologies.  Some of us (including myself) studied Scaled Agile Framework (SAFe) 4.0 in 2016, and became certified as Scaled Agilists.  Our team evolved our approach.  Our team is now a Scrum team, within a few Scrum of Scrums teams, where I am "Product Owner" (a.k.a. internally as "Engineering Dev Lead") for some "element" solutions intended to support delivery of DXC offerings of services, including but not limited to:

• Workplace Devices Services (WDS)
• Virtual Desktop and Applications (VDA)
• Mobile Enterprise Services (MES) 

Now, instead of taking many months to release updates to our "elements", we have evolved to release about once each quarter, and we're still working to become even more agile!

Our Current Approach

Now, about once per quarter, the team I lead decides the key scope of focus for our element, typically based on need to support a new version of key vendor products, and then we log a backlog of stories to take care of the various phases necessary to evergreen and add components in our solution, with concentration on supporting installation, configuration and operation of these products integrated with other products.  We also continuously add stories in our backlog to describe changes we feel we should make to improve our existing collateral.

Our team includes experts who focus on "capabilities", such as ensuring a COTS product can be installed and configured with best practices in mind or ensuring a new service can be delivered through integration of various components together, and we also have experts who focus on developing components to differentiate DXC from our competitors.

Once every two weeks, we choose a set of stories from our backlog, plan the changes needed in our components by documenting "sub-tasks" in each story and estimating time we'll require to get each task done, aiming to provide each engineer with a set of sub-tasks that can be completed in the next two week period, and then we set to work and review at end of the two weeks.  We repeat this procedure 8-10 times, and that ultimately leads us to deliver a Minimum Viable Product (MVP) release of documents and other artifacts to produce an element release.

Now, instead of taking many, many months, we typically take about about a quarter to produce an updated release of our solution related to our key vendor products of focus.

Our Challenge

With each commercial-off-the-shelf (COTS) product from a vendor, our small team of engineers now needs to maintain about 60 DXC-developed automation components (about 50,000 lines of code) to install, configure and integrate about 40 components from our vendor partners.  We also maintain over 100 documents capturing details of design, test records, and best practice recommendations.

In some cases, our documents and components are clearly valued for delivery of offerings, while in other cases we have started to identify we should retire them (some have been around for 15 years).  We want to focus on developing content with highest value.  We want to bring in the future, as we are learning about data science, machine learning, and more!  Sometimes, it can even be a bit scary to think about - watch "Black Mirror" on Netflix to get some ideas about why I say that ;-)

Our key challenge now is that this is a LOT of artifacts for such a small team to manage, so right now we are working on DOUBLING our team size!  Exciting!

What's Next?

Recently, our company (DXC Technology) released a white paper called "Creating the Agile Organization":
http://assets1.dxc.technology/innovation/downloads/DXC_Creating_the_Agile_Organization.pdf

The paper highlights that Agile is typically great for managing development of custom components, but weaker as a methodology for managing collateral related to COTS components.  This describes well what I feel is the latest "aha" moment for our team... the next phase in our history of loyalty?! 

We are now working on a shift, where we will focus more on the DXC-developed components in our solution that are intended to differentiate DXC from our competitors.  We are considering breaking up our solution to manage a bit differently.  We need to continue to use Agile to manage changes in the DXC-developed components that are considered of high value - we want to shift to release changes to those components separately much faster - maybe even every two weeks!  We are exploring changing our way of working..  maybe we could use Kanban to manage evergreen of our support for the COTS components?  We have lots of options - the challenge is finding time to figure it out while we keep it all going and interview people, too!

Time will tell where we go next.

Are you the right person to help us?  If you want to be a part of this and have the right skills, please consider applying!  Check out http://jobs.dxc.technology!