Built-in Customer Experience Analytics Boost ROI

Everybody wants to see into the future until they run a cost-benefit analysis on a crystal ball.

That’s been the experience of Parag Kulkarni, Nutanix’s vice president for SaaS engineering. Kulkarni leads teams with three critical responsibilities: creating SaaS apps for customers, developing software for internal company use, and using analytics data to generate predictive insights.

Nutanix apps and machines generate immense volumes of useful data. Kulkarni said customers understand that data is valuable. They don’t just hand it over — first, they have to understand why sharing system data with Nutanix will bring true value.

“The data from our machines can figure out, for example, when a customer’s going to run out of capacity,” Kulkarni said.

He said that kind of predictive insight can help customers avoid data system downtime by alerting them before it’s time to add more hardware or allocate more resources to the cloud.

Kulkarni said there are two types of analytics: those you see within the product that can provide insights and those created based on the machine data sent back to Nutanix. His team is responsible for the second type. Before the analysis kicks into gear, customers must first agree to share their product usage data.

“It seems like we should be able to just ask customers to send us data, but it’s not always that easy,” Kulkarni said. “You really have to show the customer the value they will get in return.”

One of the biggest predictive-analytics payoffs is in customer support. Kulkarni said Nutanix can correlate data from thousands of customers to identify problems that crop up when specific features are running. That’s when predictive analytics starts giving our customers a glimpse into the future.

“Most of our current efforts are aimed at delivering a true ‘invisible infrastructure’ experience to our customers,” said David Sangster, Chief Operating Officer at Nutanix.

“Our vision is to create a ‘No-Ops’ environment, where software and software-defined hardware are provisioned dynamically, and all the mundane core administration tasks such as patching, backup, database management and security management are all handled autonomously and automatically using AI to scan for anomalies and deep learning to update rules in real time.”

Sangster said Nutanix is well on its way to making this a reality.

“If we know what could happen, then we can be truly predictive and proactive about support, instead of being purely reactive to support requests,” Kulkarni said.

Customers are starting to appreciate the value of sharing their data. When Kulkarni came to Nutanix in 2013, about 10% of customers sent data back, and that data came from a much smaller volume of products.

“That number is closer to 60% today on a much larger installed base,” he said. “And I think there is still a huge scope for improvement.”

A Silicon Valley Veteran Gets a Rocket Ship Ride

Kulkarni came to Silicon Valley two decades ago, after getting his start writing code for British Telecom in India. He worked at Juniper Networks for 10 years until he caught the bug to try something new at Nutanix.

“I didn't actually know exactly what I was going to do here,” he recalled. “I just joined on a whim to see how things would go.”

Back then, the company had only a few hundred employees. And Kulkarni had an authentic startup experience.

“I've been pretty fortunate in the sense that because I joined so early, I ended up doing a lot of things that I would not have normally done at other places.”

In the early days, he would get a call on a Friday asking his team to throw something together by Monday. And they made it happen.

“We could work something out that might not necessarily be possible at a very large company.”

That can-do spirit helped propel Nutanix to an IPO in 2016. He quickly saw the company grow from hundreds to over 5,000 employees.

“We’re hustling and getting things done — it’s been a rocket ship for the last six years.”

Charting a Path Toward Predictive Analytics

A group of video monitors across all the support centers around the world provide real-time visualizations of data from customers using Nutanix software to run and manage their modernized data centers.

Predictive analytics are built into products, and Kulkarni’s team transforms the data sent back from the field into real-time charts and graphs that make it easy to identify customer trends and system health indicators.

“We built the data visualization initially for our support organization,” said Kulkarni, describing Watchtower, a Nutanix application that runs on monitors at all support centers. “But after talking to customers, we realized there was so much value in sharing this information with them,” Kulkarni said.

When customers saw the power of these analytics applications, Kulkarni said, they grew more willing to share data with Nutanix, which has transformed customer experiences. These tools have the potential to help anticipate how new software updates will work and what could cause a system to fail.

“One of the big things we are trying to do is come up with a way to predict how a specific engineering release will perform in the field before we release it,” Kulkarni said. “Because if you can predict that the release is going to be bad based on certain criteria, you really don't want to release it at that time, right? You can spend more time hardening it.”

Another thorny challenge is figuring out when customer-support scenarios are about to heat up.

“If we can predict that a customer issue is becoming an escalation, we can take proactive measures to prevent that from happening,” Kulkarni said.

Tough Work Lies Ahead

Kulkarni hears a lot of companies talking about how they’re using artificial intelligence (AI) and machine learning (ML). “But when you dig deep, very few of them actually use ML/AI to do predictive analytics because it's a very, very tough problem. It takes a lot of research. You really don't know how it's going to work.”

Today, analytics are about using software to generate charts that help people visualize trends. Even real-time data becomes backward-looking with each tick of the clock.

“The journey has been from reactive to proactive to predictive,” Kulkarni said. “In reaction mode, you study the past and hope to avoid future missteps. In proactive mode, you use data to anticipate problems in the hope of preventing them.”

Predictive mode changes everything — especially in the customer experience realm.

“With predictive analytics, we have all of this data,” Kulkarni said. “It gives us insights that let us truly do things that can address future trends now.”

Tom Mangan is a contributing writer. He is a veteran B2B technology writer and editor, specializing in cloud computing and digital transformation. Contact him on his website or LinkedIn.

© 2019 Nutanix, Inc. All rights reserved. For additional legal information, please go here.