The need for Engineering Ops
Companies have vaulted five years forward in consumer and business digital adoption in a matter of around eight weeks: McKinsey — March 2020
Recent macro events have resulted in a large number of businesses finally accelerating their digital transformation. This transformation had already started a few years back with the migration to the cloud, but the past few months have underscored the point that business patterns are fundamentally going to change, and we are entering a whole new world. As an example, retail companies now have to think about drive-through services. This requires those companies to adopt a whole new set of processes and software. Real estate companies are investing in VR enabled applications. This will translate into more cloud projects and data storage. Digital adoption and, more importantly, the speed of software delivery will be an unfair advantage in the coming years. In a customer experience-driven economy, here are few areas that matter:
Data-Informed Decisions
It is vital to get insights into all aspects of software delivery across various data and tools to make data-informed decisions. An example is the time it takes to understand the need to refactor a service based on customer support cases. In an ideal world, this feedback loop would be small. However, in most organizations, this latency is large and doesn’t get the attention it deserves, until customers start escalating and churning. To improve this latency, an engineering organization must correlate data from tools like ZenDesk, JIRA, PagerDuty, GitHub, and more. Every data point from sprint analytics to technical debt analysis must become readily available for consumption.
In most cases, reporting responsibility falls on overburdened program managers and team managers. Some companies are lucky enough to have an internal tools team to build custom reports for this problem. But, most organizations are unable to get data across tools without spending significant time and resources.
Lack of quick, actionable insights generally means data-informed decisions are not easy to make. A lot of decisions are made subjectively and on gut feel. Is the quality getting better with every release? I think so. Is our security posture improving? I guess. Should we probably refactor the authentication service? A lot of “I think” or “I guess” is what we hear very often.
Technology management, like any other management discipline, is part art and part science. I don’t think that will change or should change. But zeroing in on a few metrics and KPIs that matter, and then corralling the team towards those metrics could be hugely beneficial. Having the right data and insights enables the entire team of engineers and leaders to ask the right questions and make data-informed decisions.
Process Automation
Process engineering is a skill that becomes very important, once your team and product scales. Every agile organization needs to evaluate its processes to locate and perfect the ones that work and discard the ones that do not. This must be repeated ad infinitum because of the ever-evolving environment. What processes worked when you were a ten-person team won’t work when you are a hundred person organization. What worked when you were all co-located won’t work when everyone is remote.
When a new process is introduced, changed, or removed, it becomes essential to measure along the way. This is true for re-organizations and structural changes as well. For example, if your leadership decides to merge the QA and Dev role into one role or choose to eliminate developer on-call rotations, how do you measure the impact of these decisions?
The quicker one can operationalize this feedback loop, the more ammunition they have to thrive in the customer experience first economy.
Automating processes play a significant role in unlocking speed and efficiency. Automating the process of code-to-deployment, bug triaging, external and internal communication, runbooks, or security reviews are just some examples of processes that need to be fully or partially automated.
If you ask a Sales Ops leader for the next year’s revenue numbers or a Marketing Ops leader for the efficacy of an email campaign, you are likely to get quantitative data. There is a lot of automation for sales and marketing. But engineering organizations lack this automation and visibility not because of capability but because of inadequate time and resources.
We at Levelops are trying to solve these problems of visibility and automation and are building an industry-first Engineering Ops platform to drive velocity, quality, and security outcomes.
Contact me at nishant@levelops.io to schedule a demo to learn more.
