IT Challenges, Not Tech Features, Should Define an Analytics Solution
When seeking an analytics solution for IT, confusing messages from vendors, analysts, and opinionated peers can cause business leaders to contract buyer's paralysis. Hype sweeps leaders towards certain features, many of which are presented as an end unto itself. Yet, these features can only provide direct benefit when they are implemented as part of a comprehensive set of solutions designed to alleviate a specific business challenge.
Artificial intelligence (AI) provides a perfect example. A survey conducted in 2017 found that 80% of enterprises reported investing in AI. Yet, a January 2019 survey from Gartner found that only 37% of organizations have implemented AI in a meaningful form.
Many of the enterprise leaders who act reluctantly toward AI likely have an inability to envision how this technology will directly benefit their organization. A review of CIO attitudes towards AI found that many who are slow to implement the technology cite "clear technology risks, as there are no clear winners and losers across tools."
The central problem is this: analytics features, such as AI, aren't a specific "thing" to be bought and used. Instead, they are a means to an end — i.e., they exist as a part of a workflow designed to help you succeed in your goals.
As such, those in the market for an IT analytics solution should first define the scope of their business problems. Then, they can determine the specific analytics capabilities that will allow them to make progress on their goals.
Defining business problems and identifying the analytics capabilities needed to address them reveals a list of criteria for a best-fit solution. IT leaders can then make apples-to-apples comparisons between solutions with specific use cases in mind.
Putting Your IT Challenges and Goals Into Words
Organizations want to avoid being too general with their solutions criteria.
For instance, they may hear a vague observation like: "Organizations that use data are 19 times more likely to be profitable than those that don't," which comes from McKinsey research.
But what does it mean for an organization to "use data"? Installing a data collection and analytics system doesn't magically make you more profitable at the snap of your fingers. Instead, it makes your operations profitable when it can alleviate specific pain points or add to productivity.
To wade through vague feature promises within analytics unique selling points (USPs) while avoiding harmful assumptions, always start with the problem you want to solve and then work backwards towards a potential solution.
Define the problem in specific terms. Consider the people most affected by it and the exact effect on business processes. A highly specific viewpoint allows you to then determine the most appropriate metrics that give shape to the problem.
Here are a few examples of somewhat vague IT business challenges and what metrics can allow you to closely monitor specific effects:
- "We have too many business service disruptions"
- Metrics: mean time between incidents, incident volume for mission critical applications
- "IT services are too costly"
- Metrics: Mean time to resolution (MTTR), # of assignments to higher tier incident groups
- "We need to monitor SLA compliance more accurately"
- Metrics: incident volume within sensitive services or key stakeholder user groups
Each individual metric monitored is important in isolation, but by creating a unified view of all IT metrics across all systems of record IT leaders can discover the insights needed to facilitate nearly all other goals. In this way, individual capability criteria — i.e. the ability to calculate certain metrics across silos with high accuracy — assembles to form an idea of what a more complete solution would look like.
Uncovering Analytics Capabilities, Not Just Feature Sets
Defining business challenges and asking how to get to the bottom of them will reveal which capabilities are most important. You can then seek out a solution — whether custom built or off-the-rack — that meets the criteria.
In other words: Considering how a solution helps you solve a business challenge allows you to weigh your solutions options objectively, break through misleading claims, and avoid assumptions that lead to an apples-to-oranges struggle.
To demystify what capabilities they need without falling into the trap of assumptions, IT leaders should start with their specific business challenges and the capability that can solve them. Examples include:
- Problem: "We're getting inconsistent signals from siloed views"
- Solution capability: Pre-built data adapters for systems of record, automated data normalization for apples-to-apples analyses, global analysis of all available IT data
- Problem: "We need to be able to measure performance in ways that reveal our biggest opportunities for improvement"
- Solution capability: Custom expressive KPIs depicted on an interactive dashboard visualization that allows drill-down and exploration
- Problem: "We need to know which critical applications are at imminent risk of major incident"
- Solution capability: AI/ML algorithms that monitor signals from throughout your IT landscape and point to potential risk mitigation actions.
It can also help IT leaders to consider how analytics capabilities will answer key questions, like:
- "What happened?"
- "Why did it happen?"
- "What can happen if I do ___?"
The ability to report on certain metrics accurately can help with the first question, while a capability like root cause analysis (RCA) can help with question #2. Question #3 requires predictive analytics, which combines machine learning (ML) with AI in order to project possibilities and score their respective performance outcomes.
Discovering What You Need Out of a System of Intelligence
Regardless of what capabilities they want out of an IT analytics solution, leaders should focus on the available options that allow them to get a full, accurate, 360° view into their data across all systems of record in order to accurately monitor the right KPIs.
AI and ML allows an organization to make a deeper level of analysis by uncovering cause-and-effect relationships and the unconsidered metrics that reveal them. AI/ML capabilities can also automate certain tasks, like scoring a change-related risk's predicted impact. Another AI/ML use is that it can give structure to unstructured data, such as through natural language processing (NLP) coupled with topic clustering.
The solution needs to be transformative and self-sustaining. It should empower users to expand the scope of IT activities that are run and optimized using data-driven methods, and, in the process, create a dynamic and agile culture that uses metrics to drive continuous improvement.
Ultimately, procurement of the right solution is not about going out and buying something specific like, "an analytics AI," but more about understanding how something like AI can apply to your specific business needs. You want to be able to ask tough questions, and then you can get specific answers that point you down the path towards business success.