A few years back, I worked at a global retailer in Tech Operations management supporting business programs, engineering, and operational needs. The team was unique in that we were not the traditional internal-facing enterprise IT, but a division breaking through traditional IT barriers to quickly deliver service to end consumers. Our responsibilities spanned the application lifecycle, and every decision we made could have massive repercussions, cause delays in delivery, and potentially result in missed revenue opportunities.
As a core team within Operations, we defined the Application Services team’s primary objective to ensure the quality, reliability, availability, and security of off-the-shelf applications and processes. At any given moment, our team operated more than 35 applications across functions as varied as agile development, service delivery, content and workflows, and operations intelligence.
To ensure we worked as a unit, I first needed to understand who were the members of our team, what were their personal goals, and what did they see as each of their priorities? Secondly, to better understand our priorities to the business, we needed to align as individuals and as a team around key ideas:
- Adopt the corporate mission
- Understand our core business model and matrix
- Walk the end-consumer and internal customer journey and understand the application of our technology across channels
- Align the Operations model and strategy to core business model and strategic initiatives
- Develop our team mission and outline our primary objective
Looking Back at Our Team Challenges
There was no shortage of challenges in this highly matrixed organization; thus, the curious could identify underlying opportunities yet struggle to articulate the value to the business. Data silos made it difficult for us to capture metrics such as pain, ROI, risk, impact, and level of effort. Measuring these factors effectively would have enabled us to better understand demand, prioritize work efforts, assess effectiveness, and align resources. IT leaders face this challenge daily, and continually find themselves falling back on tribal knowledge and prioritizing efforts based on perceived urgency.
IT analytics address this data and metrics challenge directly by extracting and modeling core IT data sets in a centralized data warehouse, applying metrics, and exposing visuals enabling the past to tell us about the future.
Here are a few examples of the IT Operations challenges our team faced.
Application Services had a pivotal role to prepare and stabilize applications for product launches, build releases, and retail seasons. Communication and transparency were essential for us to anticipate demand, system constraints, and known issues, as well as for scheduling resources. Every day we had to calculate the optimal balance of resources among project allocations, event-based needs, and day-to-day operations. Escalated events were particularly tricky, requiring advanced planning to ensure sufficient staffing and effective resource coordination.
We regularly encountered resource allocation issues for specific projects or operations, which led to some resources being overloaded while others missed out on training opportunities. Analytics would have enabled us to plan daily demand, better understand our allocation of resources, and pinpoint which projects we were funneling too many resources toward.
Team leaders needed the ability to understand resource allocation for projects and KLO, capture how unplanned events caused delays, identify gaps and concerns, and report the status of projects, application health, and employee satisfaction. We used Excel to monitor these metrics, but faced challenges around stale data, missing information, and accurately importing data from several sources.
We knew the data revealed inefficiencies such as project slippage and system behavior/health, yet we struggled to pinpoint underlying causes. We needed the ability to visualize core trends and outliers instead of relying on the tribal knowledge of a common few.
Analytics would have enabled us to more effectively measure areas such as risk scores for change and release. We could have created effective scoring based on incidents caused by change, resource over- and under-allocation, health, and even opportunities for improved training or use of knowledge articles.
Non-production environments were complex yet critical for development and testing. However, our automation tools could only take us so far and depended on our vendor’s ability to deliver. Inconsistencies in configurations and delivery time impacted our development teams, projects, and release schedules. Our cost-planning efforts also had to include buffers for unplanned costs around vendor escalation charges, software licenses, non-production environments, and time overruns.
The accessibility of cloud environments proved valuable in reducing time to provision, maintenance, and resource usage. However, ensuring our off-the-shelf tools were able to deliver to cloud environments was a complex process. Analytics would have helped us to more consistently plan for the de-provisioning of these environments, ultimately impacting our costs and provisioning processes in a positive way.
The Key to Solving Challenges
In dealing with these operations challenges, I knew the data would provide valuable insight. Yet I also knew that I needed methods for interpreting that data to help me answer common questions such as:
- How are my resources allocated against planned, unplanned, and KLO?
- What blockers are impacting results and why?
- What KPIs for my organization are trending up/down, and what is the impact of trend?
- What’s the timing of incoming requests and incidents for planning?
- What is the work-around time for known problems?
- What are the measurable results of core foundational projects such as planned maintenance?
- How can we be more transparent and accurate in release and change risk scoring?
I found Numerify to be on the forefront of answering these types of IT questions, largely because of Numerify’s focus on fundamentals: accessibility, performance, applied industry expertise, pre-built models, and an easy-to-use interface.
[Photo courtesy of Unsplash.]