Skip to main content
Exploring what’s next in tech – Insights, information, and ideas for today’s IT and business leaders

How the cloud is changing data analytics

Analytics are seeping into more functional areas of enterprises, often without IT's involvement and with mixed results.

Once upon a time, understanding and communicating business metrics was the job of the IT department or, more specifically, the IT function responsible for what we used to call “data processing.” But today, analytics is spreading out across organizations, fueled by business units and departments that want something better than what IT is providing, faster than IT can provide it. As a result, analytics are seeping into more functional areas of enterprises, often without IT’s involvement and with mixed results.

That has a downstream effect on who is managing which data, and under whose control. And it's happening primarily in the cloud. Most on-premises analytics vendors are adding software-as-a-service versions of their products, while the number of native SaaS analytics options continue to grow. Both vendor types are targeting business units and departments (in addition to or instead of IT) because the line-of-business teams are hungry for analytics and have their own IT budgets.

Why cloud-based analytics are so attractive

The scalability and elasticity of the cloud, combined with its computing and storage capabilities, enable organizations to work with larger data sets from which they can gain insights that were previously difficult or impossible to unearth. With cloud analytics, users can combine internal data in new ways, mix internal data with third-party data, and get predictive views of success levers, such as customer behavior and supply chain impacts as opposed to historical views only.

It isn’t all rosy, however. While SaaS-based analytics capabilities are attractive, there are some very real issues enterprise buyers should consider, whether they’re adding cloud analytics to an existing mix of on-premises solutions or migrating from on-premises solutions to cloud alternatives.

It’s also important to realize that different analytics options serve different roles within the organization, including data scientists, data analysts, business analysts, and business users. Sometimes, the attractiveness of one type of option is offset by limitations that hadn’t been considered.

For example, IT industry association CompTIA learned firsthand that some cloud analytics platforms do a great job of uploading data in its current format and displaying the data, but the software is not necessarily designed to do calculations. “If you have raw data and you want to show the mean, the median, and the range, you may want to do regressions analysis that not all analytics platforms are able to do,” says Tim Herbert, senior vice president, research and market intelligence, at CompTIA. “You may have to perform the calculations using whatever you’re using on-premises, Excel or database analytics, and upload it to the cloud.”

You're adding to on-prem solutions

Many organizations have on-premises analytics solutions and are adding SaaS analytics solutions to the mix. The cloud offerings likely need to be populated with data, more data than is practical to move over a network connection.

“You can’t do that over the wire. So you have to do it by another means, such as a box the vendor ships you, like AWS Snowball or AWS Snowmobile, which is a truck they drive to the customer facility,” says Mike Gualtieri, vice president and principal analyst at Forrester Research. “Once you do that, it’s not a big deal to move data back and forth because it’s drips and drabs compared to the original move.”

The city of Los Angeles uses Amazon Web Services (AWS) for cybersecurity analytics. According to CIO Ted Ross, Los Angeles is the biggest security target on the West Coast: It has the second busiest airport in the U.S., the largest port in the Western Hemisphere, and 4 million residents. With a high-profile police department—think of the number of TV shows and news stories in which it’s been featured—that makes it top of mind for hackers, who know any hacking success gains more notoriety.

“We’re ingesting 240 million records every 24 hours across 37 different departments,” Ross says. “It’s the proverbial haystack in which we have to find the needles that represent breaches. The cloud provides an effective mechanism at a reasonable cost for us to perform large amounts of data analysis, whether it’s cybersecurity or otherwise.”

We’re ingesting 240 million records every 24 hours across 37 different departments. It’s the proverbial haystack out of which we have to find the needles that represent breaches.

Ted RossCIO of the city of Los Angeles

Organizations with significant investments in on-premises hardware and software have to decide which analytics processes to migrate to the cloud, at what pace, and for what reason. 

“It’s a workload conversation first because it depends on what applications have been developed and when. Some of those applications may not be cloud-ready,” says David Rubal, chief technologist for data and analytics at DLT Solutions, a government-focused value-added reseller. “Technology is advancing so fast that there have been systems, applications, and databases that were developed years ago and [their creators] have gone out of business. So there’s an island in the IT environment [that requires] a separate migration path.”

When migrating from a traditional data warehouse environment to the cloud, unanticipated latency issues can arise. Latency can adversely affect application performance and therefore user experience, the timeliness of analytics, and even the accuracy of time-sensitive insights.

Also, endeavor to understand the comparative TCO and ROI for on-premises and cloud analytics. To optimize the respective workloads and investments, do your best to judge what each is best suited to. 

“If you’re putting something into the cloud, you’re not locked in. You need to be able to move workloads up or pull them down, so you can start off in the cloud and move on-prem if that makes sense, or start on-prem and move into the cloud,” says Ross. "A wise organization is always evaluating and always getting the best possible benefit.”

[Further reading: Top 31+ data and analytics conferences to attend in 2017Top 10 big data and analytics predictions for 2017]

You're migrating to cloud solutions

Some organizations will migrate to the SaaS version of an on-premises business intelligence or analytics offering or to an entirely different solution or set of solutions. Either way, there may be more factors impacting the decision than simply product capabilities, features, and functions. Depending on the circumstances, technical issues such as data mapping may come up, particularly when moving from one vendor’s product to another vendor’s product.

“You have to plan ahead because you have to look at how data in your on-premises setup might map to whatever your setup might be in the cloud. I think anytime you do migration or export, you need to think carefully and make sure that you don’t have any problems in terms of data hygiene," says Daren Orzechowski, a partner in the Sourcing & Technology Transactions practice at international law firm White & Case. Maybe fields don’t map the same way or things aren’t coming out in the format they were in the old environment versus the new environment.”

Leaders must also consider the bi-directional portability of data because terms and conditions among vendors can vary. Uploading data to a public cloud service is usually not an issue. Getting the data out can be a challenge, however. Buyers should familiarize themselves with the terms and conditions of service, including data portability, before pressing the “I agree” button.

Cost is another issue you need to understand in detail up front. Cloud services are generally assumed to be less costly than on-prem solutions, but if you don’t understand the pricing structure, you may be in for a surprise. Some users have found out the hard way that as usage scales, so does the cost.

“I’ve heard of [companies] getting huge bills from AWS because people had an account and had free range to do what they wanted to do with data, but it can be very expensive if you have a consistent workload,” says Forrester’s Gualtieri. For example, he says, "if you're running portfolio risks every day, it may be cheaper to run them on-prem."

Don’t forget security

Some organizations are still so skeptical about cloud security that they avoid the cloud entirely. Others consider cloud solutions inherently secure. Generally speaking, enterprises have become less concerned about security given the massive investments Microsoft, Amazon, and others have made to protect their respective infrastructures. Yet as always, cloud security is only as strong as its weakest link.

“My security staff needs to understand how to perform security like we did with on-prem, and now that we’ve moved onto the Internet, we need to approach it that way,” says the city of Los Angeles' Ross. “You need to be able to know how to manage an application and data that aren’t residing within your on-premises architecture.” 

Some analytics vendors have built products on AWS, Azure, and Google Cloud Platform to take advantage of the scalability, elasticity, and security. However, even if a vendor’s security mechanisms meet your needs, your internal security practices may nevertheless open the door to breaches.

“In some cases, the situation may be exacerbated because you have more parties accessing cloud data than on-premises data, so you have to ensure you have your administrative rights set properly,” says CompTIA’s Herbert. “Usually, there will be options that allow users to get dashboard data and perform some basic manipulation, but they don’t have access to the dataset. They can’t export it or change it.”

Backups can also be an issue when the data sent to the cloud hasn’t been updated regularly. Although most businesses back up their data, many are not as disciplined as they should be when it comes to updating data assets in the cloud.

Bottom line for cloud analytics

Cloud analytics provide businesses with scalable, flexible, and often compute-intensive options that can be used to create competitive advantage. However, the ease and speed with which these products can be adopted may overshadow the fact that the implementation is more complicated than first apparent. For example, more lines of business are adopting their own cloud analytics solutions, but they may not consider who has access to the information or whether the vendor's end-user license agreement violates their company's security or privacy policies. 

Most have policies designed to govern the behavior of their employees, but those policies aren’t always enforced and the details can get lost in the complexity that has become the modern business environment. As always, a partnership between business and IT is wise when it comes to making technology decisions.

Analytics in the cloud: Lessons for leaders

  • Don’t expect analytics nirvana. Many cloud analytics platforms have frustrating feature limitations.
  • It’s common to use both on-premises analytics and SaaS applications. Choosing which to migrate requires answering hard questions about the appropriate analytics processes to migrate to the cloud, as well as when and why.
  • Consider how sharing data in the cloud may add new security vulnerabilities.

Related links:

How server memory advances real-time analytics

Increasing Product Quality and Customer Satisfaction with Big Data Analytics

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.