DevOps West 2017 - Process & Metrics
Monday, June 5
Agile Risk Management
Software componentization has made software more unpredictable because unforeseen conditions can cause components to interact in ways we hadn’t imagined. Greater complexity, increased user expectations, and our desire to use agile with ever increasing velocity require that we actively manage uncertainties and risks. Classic risk management identifies risks and prioritizes them to determine impact to the project, but how does that differ in an agile project? Agile is designed to handle uncertainty in requirements as new features are requested and priorities shift. What about the...
Tuesday, June 6
Thinking Inside the Box – Root Cause Analysis with The Six Boxes
Do you want to improve business and user value delivery, quality, efficiency, and productivity of your software engineering team? OK, who doesn’t? Poor productivity problems, quality issues, failing to meet commitments, and general team inefficiencies are unfortunately still commonplace. And what is at the root of most problems? James Waletzky says the answer is those highly imperfect creatures—humans. So how do we go about fixing the problems? First, we must discover the root causes, not just the symptoms, and those are not always obvious. In this hands-on tutorial, James focuses on a...
Lean/Agile Data Driven Decisions Demystified
For many agile practitioners, software metrics beyond a burndown chart are little understood or, perhaps, very scary because poor metrics can be worse than no metrics. In this enlightening session, Larry Maccherone explores how you and your organization can use metrics to bring management and lean/agile teams closer rather than becoming a wedge that drives them into conflict. Larry covers the entire lifecycle of the metrics process—from metric selection to reporting data—in compelling ways. Join Larry to gain an understanding of a wide range of concepts including common (101-level) metrics...
Statistics in Big Data Analysis: Beyond Counting
Do you have data—lots and lots of really big data? Do you know what it’s telling you? Maybe your organization is stuck in dashboard mode, counting how many of this or that. And since you have big data, perhaps you even calculate an average every once in awhile. Therein lies the problem: Organizations don’t advance their use of big data. They just spend more time and money processing more data so they can count it just a little better. Ken Johnston helps you go beyond counting and into statistical relevance. This statistics primer is designed for those with no prior experience or those...
Wednesday, June 7
Why Agile Works—and How (Not) to Screw It Up!
Agile practices can be the easy part of agile. Getting people into the agile mindset is often a greater challenge. Do you have a team member who doesn’t quite support agile or someone who’s playing along but not really committed? One step toward obtaining real commitment is a better understanding of why agile works, why it’s different, and when it is the right approach. In this fast moving session, Perry Reinert provides a fun look at some of the theory that gets to the core of why agile works. Yes, we really can use the words fun and theory in the same sentence! Combining parts of the...
Thursday, June 8
Big Data: The Magic to Attain New Heights
There is magic in Big Data. There are also goblins, fairies, and rainbow-colored unicorns. The problem is to take all that data and turn it into magical insights that help make your software products better and help the business grow. Ken Johnston has been employing data for decision-making for years in service operations, cloud development, A/B testing, and recently with business intelligence. Ken shares examples of building big data infrastructures, using telemetry to predict and react to service outages, deploying A/B experiments to drive post sales monetization, mining text data for...
The #NoEstimates Movement
When will you deliver that feature? How much will this project cost? What can I have right now? All of these are reasonable questions that both management and customers want answered. The problem is that when developers try to answer these questions, often more harm than good results. Estimates turn into commitments, dollars are committed based on misinformation, and all parties involved end up feeling cheated and confused. Ryan Ripley explains that the #NoEstimates movement is a critical look at how estimates are used in the software industry. Are estimates needed at all? Is getting...
Metrics to Assess Risk in DevOps
Preview
As software development becomes more value-focused, the need for a fluid production process emerges. That process is DevOps. However, when the number of release cycles rises, so does the risk of disruptive code entering the system and eroding the value that development creates. Traditional risk assessment techniques create a false sense of security. Risk is not a simple “go or no-go” decision; it is an input to an informed decision that requires extensive risk analysis. As the velocity increases and the focus on user acceptance and functional validation increases, the...
Include Automated Testing in Your Definition of Done
Even though most teams appreciate the benefits of test automation, it is commonly viewed as too time-consuming to be included as part of an agile sprint. This results in automation being done in isolation, typically months after the user story has been completed. This can lead to several problems including automation team members being disengaged and missing key aspects of the requirements, as well as teams going through a period where new features are being introduced but no regression testing is occurring. Angie Jones provides agile-friendly approaches to test automation that allow teams...
Setting and Measuring Individual Performance in Agile Teams
PreviewWhen software development teams work in waterfall environments, traditional performance management programs can help encourage personal development and innovation. However, Tina Rusnak says that when organizations move to agile, measuring performance takes on a new form that often causes confusion. If you're an agile developer, the days of receiving a high performance rating for writing the most code are over. Why? Agile is all about creating value, not writing code. If you're an agile tester, the days of receiving a high rating for finding the most bugs are over. Why? Agile...