Published 22. Aug. 2019
Preventing negative business impact with predictive IT insights
IT management has become an integral part of businesses. Today, organizations cannot afford to be reactive and need to be more proactive when dealing with the impact of disruptions and IT management.
When IT is unavailable or performs poorly, this also applies to your business. The impact of disruptions is constantly increasing, making reputational and financial damage inevitable. Although we like to talk about innovation, digitalization, and perfect customer experience, these are only feasible when IT can proactively manage its complex landscape. How to ensure your IT is always on, performs at its best and positively impacts your business? And how to take the next step towards predictive insights?
As it used to be: just reactive
Five years ago, Cloud was not adopted yet, DevOps was something for start-ups, and the rates of change in your landscape were still manageable. With technical infrastructure monitoring, IT organizations used to check whether the lights were still on. In case of incidents, IT responded by looking in all the silos involved trying to find out which element was to blame. This often led to vendors blaming each other, unhappy users, and an impact on business processes. It quickly became clear reactive monitoring was just not enough to ensure a stable landscape and to keep users happy.
Getting a grip: becoming proactive
In the past years, many are beginning to be aware that monitoring from an end-user perspective helps to create insights into the actual quality of your IT landscape. After all, the end-user is the one using the service an IT organization is providing. It is important to realize that an end-user makes use of a complete chain of IT elements, not just one element. That is why it is important to measure end-user actions in the context of this complete chain and not only measure individual technical performance. By creating insight into the end-user experience, early fixes become possible. For example, a user action takes more and more time: a sign something might be wrong. IT organizations can act on this timely, by looking at the connected IT components, fix any problem before it significantly impacts business processes.
Being proactive: when is good, good enough?
Being able to manage your IT proactively causes new (luxury?) problems: because when is good, good enough? And to where should budgets be allocated? In order to see which components in the landscape should be improved, KPIs are needed. These KPIs must be comparable for every application and all end-user experiences. The Apdex (Application Performance Index) can help. The Apdex is an open industry standard formula to index performance measurements. It is calculated based on performance expectations and targets set by business and IT together, resulting in a value of zero to one for every user-action, application, and end-to-end chain. Now all departments have one factual truth, a number to judge the performance on. This immediately shows which parts in the IT chain negatively impact the digital experience the most and helps IT organizations to prioritize investments.
Maturing IT: linking to the business
Proactive IT management using the Apdex offers the possibility to closer link IT to business goals. For example: if the conversion ratio is a business goal, the IT systems linked to the e-commerce chain should perform optimally. Improvements in the Apdex rates of this e-commerce chain might directly impact conversion ratio. By correlating factually measured business goals to IT performance and stability, we can determine the impact of IT investments. In short: with Apdex we know where we need to improve in the chain, and by adding business KPIs we can measure the results of these improvements.
The next step: predictive insights
All the (monitoring) data collected should become relevant information that helps achieve business goals and prevents incidents from happening. Data can be correlated by hand, but there are a lot more ways all that data is interconnected. The first step to mine IT operations data and business data is to implement event management: monitor which events occur in the IT landscape and process these in an event management engine. When all infrastructure configurations, end-user experiences, application metrics, and business drivers are also added to this event management engine, automatic analysis, correlations and clear visualizations can be made. This will help create one single point of truth which the complete organization can use to make informed decisions.
To take that next step towards predictive insights, a machine learning algorithm is needed. With this, patterns in the extensive data set can be found. Experience shows that 60% of IT incidents have occurred before. This means when the algorithm has learned such a pattern and the same pattern starts to occur again, it can send an alert. Voila, predictive insights! Now every recurring incident – the 60% – can be predicted 20 to 30 minutes before user experience or revenue is impacted.
Having the ambition to improve the quality of IT and have IT organizations positively impacting business processes, Ymor and Splunk together provide the tools and data needed to create these predictive insights!