Continuous Measurement and Feedback are Critical to Improving Software Delivery
Overview: It’s clear from DOES London that companies are making tremendous progress on their DevOps journeys, but they need to be measuring their development processes so they can continuously improve software delivery quality, velocity and efficiency. There hasn’t been an effective tool to enable that – until now.
In this Age of Software, unicorns like Uber, Snapchat and Instagram share the limelight, but not every company can be a unicorn. That said, doing business like you’ve always done it is tantamount to taking the express elevator into irrelevance. Large established companies must understand it’s not enough to be big; you’ve got to be continuously creating digital experiences that delight your customers–and continuously improving through measurement. Agile software development enables teams to continuously turn worthy ideas into delightful digital experiences that positively impact the bottom line. And the DevOps Enterprise Summit (DOES) is THE place to learn, share and collaborate on how to practice Agile software development.
I recently had the pleasure of attending DOES London hosted by Gene Kim. I heard great stories from organizations just starting their DevOps journeys to companies like Adidas using Agile methodologies to grow their digital revenue stream from $2 billion to $4 billion. There was even an impressive deep-dive of a high-nine functional programming architecture that supports warehousing and inventory management at Walmart.
Every story echoed what Chris O’Malley and Joe Aho from Compuware made clear in their Fireside Chat: Across all stages of successful business evolution, if your organization’s engineers are allowed to work on new and interesting customer-facing innovations, employees will be more engaged, customer satisfaction will increase and so will your company’s cash flow. Amazing innovation and thought leadership is emerging from teams who are allowed to experiment and challenge the status quo.
Impressive Progress, But…
While organizations have made really impressive progress with their Agile transformations, many are in search of an elusive creature they’ve yet to catch: accurate measurement of the Agile, creative, and development processes. Huge kudos to the companies that have evolved over the last 10 years, but how did you measure your practice as it evolved? How did you get there and how did your people evolve as new enablement tools, technologies and platforms were introduced? Thousands of decisions are made every day, but which ones yield increased engagement, customer sat and cash flow? The DOES presentations, while compelling, left these questions unanswered.
As I parse through my notes, a silver lining does emerge: Agile and persistent pursuit of the better process is still the only clear way forward. We all trust Agile, but because of its dynamic properties and its perpetual state of change, there needs to be a highly adaptive measurement solution to help companies stay the course and remain directionally correct.
My Perspective—and Questions—as a Solutions Architect
As an architect, I’m always looking to improve the overall health, agility and culture of my organization, but how do I know if the decisions I make regarding technology or process have an influence on customer satisfaction, employee engagement and cash flow? To do this I need to understand the current state of my organization and how we tick technically and culturally.
This was the genesis and the force behind evolving Compuware’s VIP program into what it is today: zAdviser, a dynamic and evolving framework built to lend extraordinary insight into mainframe software development practices. Today, zAdviser understands nearly 800 different user behaviors through a common data model within the value stream that connects ideation to feature delivery.
With the help of zAdviser, I have begun to delve into how mainframe software is built and where the bottlenecks are so we can continuously improve the quality, velocity and efficiency of our software development and delivery. For example, we saw that by “shifting left” and adding automated testing early and often in our development cycle, we reduced bugs and decreased meant time to resolution (MTTR) from 29 days to 20 over the past year. By reducing bugs, we’ve been able to spend more time innovating on behalf of our customers.
Beyond Data Collection
The ideation and experimentation process is pivotal to my success as a solutions architect and ultimately drives higher customer satisfaction. To build high-quality products that our customers enjoy each and every day, I need the ability to leverage a solution that provides insights into our customers’ creative and development processes. With zAdviser, I have that and so much more. It acts as a medium to help communicate what is working and what isn’t. As Gene pointed out, in the Age of Software, we need to be able to leverage the scientific method to build better software.
For most development organizations hypothesis validation is one of the hardest metrics to come by–except for those that have zAdviser. With access to zAdviser, you can tap into a live stream of how new and legacy product features are being utilized around the world. Not only does zAdviser allow me to quantify and support the decisions I have made so that the C-suite continues to invest in my “radical” new ways of thinking, but it effectively decreases my frequency of failure so that our customers are consistently wonderfully delighted with new product functionality.
On-boarding New Developers
zAdviser has powerful implications for hiring and retaining next-gen developers. More complex hypotheses are forming around how next-gen mainframe developers with object-oriented backgrounds are approaching the Assembler and COBOL language constructs. This helps us shape the language on-boarding and training process to effectively increase our ability to consistently build better software, faster.
The world relies on mainframes to power some of the most critical services we use today and we are leveraging zAdviser to help understand a very important and critical value stream in keeping the platform alive and healthy. Everything we learn from our usage or our customers’ usage of zAdviser will potentially help shape the future of the next phase of mainframe software development.
My thanks goes out to Gene for bringing us all together to celebrate our successes as well as our failures so that we can grow together.