alphaspirit - Fotolia

Self-service BI and analytics tools face evolving set of challenges

Self-service BI tools have become the go-to standard for business intelligence software, but users are pushing vendors to rethink how they address enterprise needs and advanced analytics.

When the IT staff at Ivy Tech Community College started evaluating self-service BI and analytics tools about six...

years ago, they developed a long list of must-have features. The ability to produce top-quality graphics was not on the list.

"We needed more than pretty pictures," said Lige Hensley, CTO at the Indianapolis-based community college.

Hensley and his team eventually settled on software from Pentaho. While he acknowledged the data visualizations produced by the software are not as attractive as those of other tools, the data preparation and governance features made it a good fit for the school's diverse needs and broad user base.

Self-service has become the de facto standard for business intelligence, and a number of lightweight, easy-to-use tools have come to dominate the space. But as more and more enterprises look to put these tools at the center of their data strategies, some are finding that the leading players don't have all the features they need for enterprise-wide deployments.

"Tableau is good on the graphical perspective, but it didn't cover the needs we had," Hensley said. "If you need data access, that's not the tool you go with."

At the time Ivy Tech brought in Pentaho, Hensley and his team were looking to stitch together 40 different data systems that stretched across all school departments, including the registrar office, financial aid and fund raising. They wanted users in these departments to be able to put together reports for themselves.

Reports had to offer drilldown ability so users could interrogate the data, and any data taken out of a report had to be auditable. The Pentaho software attaches metadata to data as it moves around, so anyone looking at information in a report knows where it came from and how it got into the report they're viewing.

Users clamor for advanced analytics

Enterprises need to consider carefully what they hope to get out of self-service BI and analytics tools, and make sure the one they're implementing fits their needs, said Peter Maynard, senior vice president of data and analytics at credit bureau Equifax, based in Atlanta.

He said self-service BI and analytics tools like Tableau and Qlik are handy because just about anyone can use them. But they don't offer the more advanced analytics features that many businesses are looking for today, which limits them somewhat in terms of their applicability throughout the enterprise. To get to predictive and prescriptive analytics, organizations still need more highly trained people operating complex tools.

"Those [self-service] tools, in my mind, are more for a business degree," Maynard said. "The higher-performance stuff is graduate degree."

Problems like those are affecting the bottom lines of self-service BI vendors. Tableau Software Inc. did beat its revenue and earnings forecasts for last year's fourth quarter, after missing financial targets in previous quarters. But the company only forecast 6% revenue growth and break-even results for 2017 as a whole. Its stock price, which dropped precipitously early last year, initially spiked upward after the Q4 numbers were released, but it has dropped off again since peaking in early February.

For its part, Qlik, which went public in 2010, was sold to a private equity firm in August of last year. Some analysts saw the move as a sign that the company was having a tough time standing out in an increasingly fractured, competitive market for self-service BI and analytics tools.

View from the vendors

The vendors say they're on top of the issues. Dan Jewett, Tableau's vice president of product management, acknowledged that the company's former strategy of signing up business units for purchases, rather than starting with organizations' IT shops -- an approach sometimes known as land and expand -- earned Tableau few friends in enterprise IT circles.

But he said that the vendor has been moving away from that strategy over the past few years, and that its new CEO, Adam Selipsky -- hired last September after a stint as a vice president at Amazon Web Services -- brings an increased focus on the enterprise. "Land and expand got us into business units five or six years ago, and IT didn't really like us, but we're all-in on the enterprise now," Jewett said.

Bringing in Selipsky to replace co-founder Christian Chabot as CEO is one of a series of management changes made by Tableau. At the same time, the company promoted former vice president of product Andrew Beers to chief development officer in place of another co-founder and made former vice president of product management Francois Ajenstat its chief product officer. And last month, it hired former Oracle executive Dan Miller to run its worldwide sales, services and support organization.

On the advanced analytics front, Jewett said Tableau has integrations with statistical analytics languages R and Python, and it will continue adding functionality as users demand it.

James Fisher, Qlik's vice president of global product marketing, said his company will soon also add R and Python integration. Qlik's processing engine also supports API connections to a number of other advanced tools for things like natural language processing and advanced search.

But there's still work to do

Donald Farmer, a former Qlik and Microsoft executive who now heads consultancy TreeHive Strategy, said vendors like Tableau and Qlik are being challenged on multiple fronts. In addition to grappling with growing user interest in advanced analytics capabilities and dealing with governance issues, the shift toward lower-cost cloud BI systems and stronger competition from Microsoft are hurdles self-service BI and analytics tools need to clear. Also, while Qlik has positioned itself to compete for large enterprise deals, Tableau trails behind in meeting user demands for higher-level BI features, according to Farmer.

"Tableau hasn't really cracked that nut yet," he said. "It's still primarily a data visualization tool that front-ends someone else's analytics engine. They've got to get up the enterprise stack." The need to do so explains why the vendor brought in an executive from Oracle to manage its sales and service operations, Farmer added.

"Tableau is under pressure to add all this functionality and be more than just a fancy Excel replacement for data visualization," agreed Wayne Eckerson, founder and principal consultant at Eckerson Group. For example, Tableau's software still has shortcomings in areas such as governance, usage auditing and security management -- "basic functionality that an enterprise BI tool needs to have," Eckerson said.

As a result, end users in different departments running their own Tableau systems often create separate data extracts that can't be managed centrally, Eckerson said. "If Tableau spreads organically, as it does in most organizations, it's hard to keep data governed, secure and consistent. It can create a mess if you aren't careful."

And Tableau's fourth-quarter showing wasn't enough to convince Farmer that the company has put its problems to rest. "It's not clear that they've turned the corner yet," he said.

Craig Stedman, executive editor for SearchBusinessAnalytics, contributed to this article.

Next Steps

Focus on user is a boon to self-service BI adoption

Users urged to be wary when rolling out self-service analytics tools

Self-service tools compensate for shortage of experienced data pros

Dig Deeper on Self-service and collaborative business intelligence