Business intelligence has continued to grow as a priority for businesses as data analytics tools continue to improve, but several BI challenges trip up organizations as they try to gain insight and maximize the advantages afforded by BI.
As part of a recent study conducted by Mike Leone, senior analyst, and Bill Lundell, director of syndicated research at Enterprise Strategy Group, 392 surveyed IT and business professionals across large midmarket (500 to 999 employees) and enterprise-class (1,000 or more employees) organizations in the United States and Canada reported the challenges organizations face in expanding their BI use.
Lack of BI training
As organizations grapple with the best ways to utilize BI to enhance operations, the largest roadblock is skill gaps or lack of training among the employee base. With 42% of the 392 respondents supporting the issue, it was by far the most prohibitive aspect of BI integration. However, only 28% of the companies in the survey reported taking steps to address those training issues.
"Collaboration is the key to data-centric success -- I think that there's awesome technology tools and services made available to everybody, but at the end of the day there's only so much that can be delivered through that technology and then it falls on the people," Leone said.
David Menninger, SVP and research director at Ventana Research, agreed that collaboration is necessary to understand and overcome BI obstacles.
"It's a way to share that knowledge," Menninger said. "Collaboration tools could see three or four people all used this analysis and liked it."
Leone identified several factors behind the training issues based on his experiences and interactions with companies during his time at ESG. One is the generation gap: Leone has found companies can struggle with a divide between younger and older employees when adapting and incorporating new technology into the workplace.
Some of it has to do with the ability to embrace new tools, analytics, technologies that may be unfamiliar to certain employees.
"It's a comfort zone … the ability to kind of openly explore and try something new. Partially, I think it's a generational thing," Leone said. "You have the folks five years away from retirement and their mentality is 'I don't need to do that. I'm out of here soon, hire somebody else to focus on that.'"
Menninger said while he's seen training differences manifest from younger to older generation, there's a different angle to consider as well.
"I don't think training is the issue," he said. "We're trying to train people to use BI rather than embedding BI into their daily job. They shouldn't have to go use a BI tool; applications should make things simple."
Another contributing factor has been the lack of guidance. Organizations want to be more BI-centric, but that will require help from members of the organization who already have expertise within the field.
"There's aspects of 'hey we have data rockstars in our business today who already know how to leverage these tools,'" Leone said. "But if these technologies don't enable collaboration, don't enable sharing of some of those insights and ways to get from point A to point B so generalists can start faster, that's a big deal."
The solution to improving the use of BI in organization, in Menninger's eyes, lies in natural language processing and embedded analytics: Make the tool do the brunt of the work to lessen the burden on employees who are less familiar with it.
David MenningerSVP and research director at Ventana Research
"It's much more accessible than teaching people to use a BI tool," Menninger said. "I think those two things are much more key to have an impact in the usage of BI than training."
"[For example], we didn't train more people to be switchboard operators. We removed the need for switchboard operators."
Menninger noted training will continue to be important until the tools have the capabilities to make life easier for the user.
"We don't have tools that are as easy to use as we need them to be," he said.
Learning how to use a new database or program is daunting for many, especially those outside their comfort zone. There's also risk involved: A mistake made by one person could impact the rest of the employees with access to that set of data.
One way to both alleviate fears of employees who are scared to make any mistakes and protect against risk of major issues is to create a sandbox environment, where employees can learn on their own without fear their mistakes will impact their co-workers and company.
"If you can provide a sandbox-type of environment where you know interacting with data, playing with data isn't going to impact other folks in the organization, that's valuable," Leone said. " The last thing you want to do is expose a number of generalists to a new data platform, say 'ready, set, go' and one of them does something wrong and ruins it for everybody."
BI cost hurdles
Cost was a distant second to skill gaps and lack of training, with 31% of respondents reporting it prevented employees from utilizing BI and analytics platforms. Expanding BI across an organization certainly opens up new security and governance issues. Exposing many people to possibly new data and technology increases an organization's surface of attack and its ability to deliver the right level of compliance. Security and governance were the top cost factors impacting end-user usage or adoption, cited by 38% of respondents.
"They really need to evaluate cost implications of exposing entire organizations to platforms, business intelligence, analytics, whatever it may be," Leone said.
While training end users is important, the second highest cost factor is training the support staff -- which is the IT team in 67% of surveyed cases -- with 36% of respondents choosing it as a top three factor.
"We want to make sure those that are supposed to support the end users are able to," Leone said. "I think the scale component is really interesting when it comes to cost."
Leone said the areas IT must support for BI may not fit under traditional IT skill sets. This presents a new opportunity to expand educational knowledge.
One of the other strategies companies have pursued to fill the gaps in knowledge and support is outsourcing BI needs to a third party. This brings immediate expertise to support employees learning new systems. However, it can leave a lack of knowledge and support behind if the companies part ways and internal measures haven't been taken to create an internal support infrastructure.
Utilizing AI, embedded analytics and natural language processing can make the transition for employees much easier as well.
"The idea of setting people up for success as you're asking them to do something outside of their core job responsibility is very valuable," Leone said. " It's easy to get lost in the excitement and the potential business benefit of leveraging data and leveraging it more frequently and getting insight faster, but empathy goes a long way when leveraging some of this technology and exposing some of these folks to it."
Investing in BI tools also requires an organization to make sure it selects the tool that fits.
"Most offer built-in insights or augmented intelligence. Many are just suggesting things -- so suggesting which reports or analysis to look at," Menninger said. "Your organization may be more amenable to NLP. Other organizations may be more data literate and what they need is something to do more traditional analysis."
Tackling the accessibility issues, whether training or increased automation, and balancing costs will unlock efficient and effective BI use. Vincent Granville, founder and self-publisher at datashaping.com, worked on a project while at LookSmart that ran BI queries directly against the Oracle database with a client, which required little user expertise and investment to the business.
"Even though in the end they were using my Perl scripts to access the database -- entering their own SQL queries generated by Toad -- as a text file input to my Perl code, the training was minimum, the costs essentially zero," he said. "They only needed to know a few basic UNIX commands, and nothing at all about Perl programming -- I designed it as a black box for them -- and indeed, not even about SQL coding. In the end, it saved a lot of time to these employees."