Data analytics has emerged as one of the most important business and technology differentiators for organizations, giving them the power to draw keen insights about virtually any aspect of their operations and thereby gain an edge on the competition.
Research firm Gartner earlier this year predicted that 2017 would be the year data and analytics go mainstream, creating value both inside and outside organizations that have prepared for the shift. Approaches to data analytics are becoming more holistic and encompassing the entire business, the firm says.
Among the key trends emerging, according to Gartner: Analytics will drive modern business operations, not simply reflect their performance; enterprises will create end-to-end architectures allowing for data management and analytics from the core to the edge of the organization; and executives will make data and analytics part of the business strategy, enabling data and analytics professionals to assume new roles and create business growth.
And companies are investing huge amounts of money on analytics tools. International Data Corp. in a March 2017 report forecast that worldwide revenues for big data and business analytics will reach $150.8 billion this year, an increase of 12 percent over 2016, which the firm estimates will continue through 2020, when revenues will be more than $210 billion.
[ Keep up to date with the 10 hottest data analytics trends today (and 5 going cold). | Bolster your career with our guide to the big data certifications that will pay off. | Get the latest on data analytics by signing up for our CIO newsletter. ]
And yet with all this emphasis on data analytics, many organizations are falling into traps that jeopardize or squander the true value of analytics. Here are seven sure-fire ways to fail at analytics, according to IT leaders and industry experts.
1. Jump in without knowing what you’re looking for
Without knowing what specific trends or signals to examine in your data, how can you expect to draw any true value from it?
“The biggest problem in the analysis process is having no idea what you are looking for in the data,” says Tom Davenport, a senior advisor at Deloitte Analytics and author of the book Competing on Analytics: The New Science of Winning.
“This idea behind data mining that you could have the system find out what's interesting in the data has led many companies astray over decades,” Davenport says. “Even with machine learning, it's helpful to know what you're looking for in terms of relationships in the data.”
Weather.com puts an emphasis on finding “people who know how to query our data and tell a complete and accurate story of what the data is trying to say,” says Todd Eaton, quality assurance manager at the weather site.
“The right people are passionate about using data to answer questions and then are willing to constantly question their findings to make sure the data is not just fitting a narrative but can explain what we are seeing and helping to predict where we are going,” Eaton says. “It is important that everyone knows what we are trying to find with the data and our overall goals, and to collect consistent measurements and data.”
A sure recipe for failure is lacking focus when launching an analytics effort. “Data teams will be most successful when they are focused on a prioritized set of outcomes,” says Christina Clark, chief data officer at multinational conglomerate GE. “Often teams will fail because they are expected to address too many business demands at once, ultimately being stretched too thin and not making meaningful impact to maintain interest or funding.”
2. Build (and maintain) your own infrastructure
There might be a strong temptation to build and maintain your own big data infrastructure. But that could jeopardize the mission of your analytics efforts.
“This generally wastes a lot of data scientist time on tasks other than actually developing better analytics,” says Oliver Tavakoli, CTO at cyber security company Vectra.
“We knew we wanted a lot of data to base our analytics on,” Tavakoli says. “We started by doing what everyone tells you to do: We bought a bunch of servers with lots of disk capacity, we put them in our co-location facility, we created our own Hadoop cluster on top with Apache Spark and had our data scientists write Scala code to interact with the cluster.”
The cluster would break, sometimes due to hardware failures, more often due to software failures. The software packages would get out of date and sometimes hours would go by with the cluster being unavailable.
“We finally had enough and decided to outsource this part of the problem,” Tavakoli says. Vectra went with an outside provider and has since spent little time “on the nuts-and-bolts issues, and almost all of our time has been dedicated to feeding data into the system and analyzing the data in it,” he says.
3. Be a data divider, not a data unifier
Enterprises have long struggled with the problem of “data silos” that prevent different departments from sharing information in ways that could benefit the organization overall. The same challenge applies to analytics.
A good best practice is to unify disparate data, says Jeffry Nimeroff, CIO at Zeta Global, a customer lifecycle management marketing company.
“Every data silo creates a barrier between interconnections that can yield value,” Nimeroff says. “For example, think about a rich user profile either connected or disconnected from website activity data. The more data that can be interconnected the better, as those interconnections are where predictive power can he found.”
This doesn't mean having to move all the data from their originating systems into a monolith, Nimeroff says. “Instead, we use one of the modern integration technologies to provide a unified view of the data while it rests in its current systems,” he says.
4. Eschew good data hygiene
If the data you’re analyzing is not accurate, up to date, well organized, etc., the value of the analytics can drop drastically.
"Garbage in, garbage out is a problem that is magnified by the volume and scope of raw business data,” Nimeroff says. “The best [data analytics] teams want quality to permeate. As such, building processes and leveraging technology that enforce quality standards is a winning combination.”
On the process side, ensuring repeatability of processes and then auditability of the results is important, Nimeroff says. On the technology side, deploying data quality tools including profiling, metadata management, cleansing, sourcing, and so on, help ensure better quality data, he says.
Organizations need to use tools to “clean out debris — incomplete and broken data — and massage data from different sources to make it compatible and comprehensible and to make it as easy as possible to analyze,” Tavakoli says. “Make the data as self-describing as possible so all members of the team understand the meaning of the various bits of data.”
High quality data “is the key fuel for generating useful insights,” says TP Miglani, CEO at Incedo, a technology services firm. “You need to build data warehouses and data lakes to bring in the structured and unstructured pieces of data together. Successful [organizations] make sure they improve quality of the data with cleaning, computing missing values, [and] labeling it accurately.”
Good data hygiene also means keeping data as current as possible. The data needs to be fresh and the “data universe” constantly expanding for companies to draw value from analytics, Nimeroff says.
“Data freshness requires having an understanding of the timeliness of your current data acquisition processes,” Nimeroff says. “Obviously, the more real-time a system is, the better the freshness. Freshness can also be supported by using third-party services to augment your existing technology and processes.”
5. Forgo executive sponsorship of analytics initiatives
As with any other type of major IT project, not having the blessings of senior executive leadership on data analytics projects can be a detriment to success.
“The objective of analytics teams is to generate insights by connecting the data with a company’s tactical and strategic decisions,” Miglani says. “One example of failure would be if a data science team did great data analysis, developed accurate predictive models, but the results were not implemented because it required changes in organization and culture.”
Building a data management foundation takes sustained effort, often over multiple years, Clark says. “Some of the work a data and analytics team needs to drive will not have obvious immediate results, which may be out of alignment with business partner expectations. This requires strong leadership buy-in and efforts to educate business partners to enable a more data-driven future.”
6. Ignore middle- and lower-level managers
Analytics performed in a vacuum by data scientists and other experts without solid input from the business managers who are closest to the need for analytics will likely not be as successful.
“Without the active involvement of mid- to lower-level managers, the information delivered by the analytics team often fails to actually help the management team do their job better each day,” says David Giannetto, COO at Astea International, a provider of service management software.
“The information will be directional, point out larger process flaws or areas that can be improved, but management will get to that someday — when they have time,” Giannetto says. “And most managers never have extra time. It is only when the team is comprised of people who actually know the business and the information the business actually needs access to each day that the information delivered becomes tangible enough to positively impact the business.”
If analytics tells users where a real problem is — where they are likely to fail — in enough time for them to prevent it, they will use this information each day, and the initiative will be successful, Giannetto says.
7. Lack the culture and skills to support good data analytics
This is a common problem for organizations, in large part because skills such as data science are so hard to come by. But if data literacy is not central to a company’s culture, the chances of failure with analytics is greater.
“For folks who are not familiar with analytics, data science is perceived as some sort of magical way of solving problems,” Miglani says. “The concept of prediction and self-learning is very hard for people to grasp. It will be hard to convince your business partners to make decisions on opaque algorithms. You will need to educate them first.”
And organizations continue to struggle to find data scientists and other professionals with analytics skills. “One of the best ways to develop this capability is to groom this talent, instead of scouting out superstars outside your organization,” Miglani says. “Many projects fail or get delayed because [companies] are not able to hire analytics folks on time, or lose them to high attrition.”
Related data analytics articles: