Predictive analytics may be the future, but first we need to solve the data problem

“Your voice is breaking.” “We lost you there for a while.

How many times have we all heard or said these things? What about the white circle of endless caching? We’ve all experienced endless glitches, downtime, and broken apps that affect us more than we’d like to admit.

For enterprises, the move to the cloud and reliance on SaaS applications has made the Internet the corporate backbone. The internet is a digital supply chain that gives users a great digital experience. But there is no certainty that your application and its many components distributed across multiple cloud environments actually perform on an equal footing.

So for today’s organizations that want to be more proactive and automated in the way they operate and manage their environments: is delivering predictable digital experiences in an unpredictable online environment even in the cards? I would argue that the answer is yes, but only if you solve the data problem first.

The data problem that is the Internet

IT operations are an overwhelming place these days. In today’s connected world, where every business, application and device rely on a digital connection every hour of every day, providing the highest quality digital experience is crucial. But with applications running in the cloud and accessed from multiple remote endpoints, the number of new blind spots has created huge challenges for anyone tasked with troubleshooting flawed user experiences. This complexity causes the network model to be plagued by reactive troubleshooting, and the user experience is regularly degraded.

Network experts tell us that responding to disruptions and adapting to new business needs are two of the top network challenges. For these companies, the quest for predictive intelligence is to be able to move from reactivity to prevention, thereby pinpointing issues before they impact the user experience. Predicting and taking control of what is happening in the cloud has now become the backbone of the corporate network.

Predictive Intelligence: Unlocking Performance Growth and Opportunity

However, predictive intelligence promises real productivity gains. For organizations with hybrid employees, the benefits can be significant. Predictively identifying a single service impacting an outage and fixing it—for example, by switching providers and paths carrying application traffic during peak periods—can save hours of downtime or reduced productivity for a single employee. Multiplied by the employee base, this number quickly becomes significant.

The same applies to meeting consumer demand. In an era of exponential choice, proactively preventing any disruption is key to delivering the always-on digital experience shoppers need and demand. In fact, expectations for digital experiences have skyrocketed.

Unlocking productivity gains and opportunities to increase brand value is a true turnaround in predictive analytics.

Determining the size of the data challenge in predictive analytics

Troubleshooting is a largely reactive endeavor based on analysis and informed decision-making to improve the situation or identify potential root causes of an active incident.

Figuring out what’s going on or what went wrong satisfies an immediate need, but it doesn’t help in any way to avoid a cycle of users leaving a lagging app or an unavailable cloud service.

This is the promise of the predictive internet: the ability to leverage a rich dataset and visualization to analyze historical patterns in a complex network of proprietary and third-party networks to predict service outages or degradations and take remedial action before users are impacted.

Predictive intelligence at this level is both a data problem and a scale problem. Solving these problems is the key to making it a real reality.

It takes a huge amount of data to predict the onset of performance degradation or deterioration with high accuracy. While the amount of data needed to train a model has been around for some time, the data has often not been as clean as it should be. This resulted in a flow-on effect in the statistical models. Without good data, the models simply couldn’t generate detailed assessments and actionable recommendations.

Thanks to advanced modeling technology, supported by high-quality data collected from the client’s extensive network, predictive intelligence is at your fingertips.

The guiding hand

So what does predictive intelligence look like today? It starts with visibility and ends with trust. Data-driven visibility that provides insights into cloud and web environments that the organization does not own—but that have become part of the corporate network and thus critical as a mechanism for delivering digital experiences—is critical. It is equally important to supplement this visibility with the data you have from an analytics model that learns from past behavior and predicts future events.

Third, and perhaps most important, is to recommend what actions to take based on the data and insights of continuous measurement and evaluation of performance. Giving up control of your IT infrastructure is an impossible task without first building trust. Recommendations build trust. Trust that the data is correct and trust that the recommended action will produce the intended result.

Predictive intelligence should be taken as a guideline that helps companies see and measure performance across all networks that impacts the user experience, predicts issues based on historical data, and influences decision making.

Mohit Lad is the co-founder and CEO of the company Thousands of Cisco eyes.