ALM best practices for advanced DevOps organizations
A comprehensive collection of articles, videos and more, hand-picked by our editors
BizOps. Or is it BizDevOps? And DPM -- digital performance management. They are the trendy new buzzwords. Think of it as IT's DevOps welded onto a tight connection with the business mission and its processes and operations. Tammy Everts, who has devoted two decades to creating the best possible user experiences (UX), talked about BizOps and its relation to performance in an exclusive interview with SearchCloudApplications. Currently senior researcher and evangelist at SOASTA, she explores web performance, UX and business metrics. Her latest book, Time Is Money: The Business Value of Web Performance, was published in June 2016.
What is BizOps, or BizDevOps, and how does is compare to DevOps?
Tammy Everts: BizOps connects IT with the rest of the business. On the business side, people want magic numbers -- unicorn metrics -- to measure the success of the business. It's tempting to say the goal of BizOps is to give you these unicorn metrics, but it's not quite that.
Isn't part of the challenge simply that BizOps and DevOps are evolving strategies, making them difficult to define?
Everts: It's like a decision support mechanism. We think historically of how a business has been guided by extroverts with big ideas who shoot from the hip and think with their gut. BizOps is the opposite of that because that approach hasn't been proven successful in several financial meltdowns.
You're saying that intuition is not a prudent approach. Is data the missing link?
Everts: Now that companies have access to so much data from every aspect of their business, you need to harness that and use it to create a decision support mechanism that lets you optimize day-to-day operations and think about, what are the important strategic questions that a business needs to consider? BizOps helps you make decisions that are driven by data, not by the people who talk the loudest in a boardroom. BizOps is an overarching umbrella that includes analytics.
Who does best with this data-driven decision strategy?
Everts: The people who do well in BizOps are IT people who get business metrics and business people who get IT.
Another term that comes into this mix is digital performance management. What role does that play?
Everts: We position SOASTA as more of a digital performance management [DPM] provider than application performance management [APM]. Our DPM platform brings together APM, analytics and site metrics like revenue, conversions, bounce rate, session length, page views and more. To bring this together, we have developed a data science set of tools. Doing advanced analytics separates DPM from APM.
What does digital performance management look at to find performance problems?
Everts: We've developed some advanced algorithms. One is for conversion impact scoring and the other for activity impact scoring.
Can you define those terms?
Everts: Conversion is what happens when a person on your site or in your app becomes a user or buyer. For a SaaS [software as a service] vendor, it's when a person signs up. For a retailer, when a person buys something. But this applies to anyone, not just retailers. The Conversion Impact Score ranks page groups by their propensity to negatively impact conversions due to slow load times. The Activity Impact Score is related; it uses your own user data to answer the question 'how much impact does the performance of this page have on session length?'
What does this mean in practical terms for mobile or cloud application developers?
Everts: We can look at all of your site traffic, at the render times for all the pages visited, and see how this correlates to conversions or session lengths. Historically, if you're a developer and wanted to make pages faster, you tended to operate from your gut, starting with the homepage, then landing pages, then category pages. Alternatively, you could use traditional APM tools to know which specific pages were the slowest. The problem with this is you may not see any improvement because the slowest pages may not be visited often enough to make a difference.
How does the developer's work translate to the core business?
Everts: By looking at all of the user data, we can, with a high degree of accuracy, tell you that if you make your pages 100 milliseconds faster, what the likely outcome will be in terms of revenue, shopping cart size or other metrics. On the machine learning side, we can take all of your past data and figure out the acceptable tolerances for performance and the impact on the business. You can then use this to set parameters for monitoring, say, a marketing campaign.
Does that make digital performance management a technology or a strategy?
Everts: DPM is a strategy. BizOps to me is more of a philosophy you adopt in creating a DPM strategy. It's a trifecta of people, tools and processes. The companies that are most successful have broken down the walls between the business side and IT. When we talk about DPM or BizOps, it's not enough to have a tool or processes in place if you don't also have a culture in place.
Where can this go wrong?
Everts: DPM and BizOps are essentially situational awareness; an indicator of what's happening right now. One of the problems we're trying to fix is when you have only a partial glimpse of your data and [developers] make fixes based on this incomplete understanding.
An example for developers?
Everts: If you know your site is too slow, but have only a narrow insight into how it's performing based on which pages are slowest, you'll likely fix those first -- and it's a lot of work. Making pages perform better is challenging because conditions are always changing. To go through all of that work only to find that it doesn't have any impact on the business can be really demoralizing. This makes it harder to invest on improving performance in the future. As a developer, you have to know how what you work on will help the business. You have to be able to tie performance improvements to results in the business.
Check out Tammy Evert's blog.
Rely on data to drive cloud management
Social media metrics can drive business decisions
Multicloud complicates application performance
Cisco acquires AppDynamics to broaden its performance management portfolio