You don’t need IT to glean answers from data.
We’re working in the Analytics 3.0 era, after all, where data driven insights are delivered via self-service tools that no one even needs to know that you’re using. You can grab personal editions of consumer-friendly tools like Tableau, DataHero and SAP Lumira from the web and get busy creating beautiful, engaging visualizations in short order.
Let the overworked IT department whittle down their backlogs — you don’t need their help in getting your job done.
This is a good thing, right?
Maybe not so much, argued Pyramid Analytics CEO Omri Kohl. He called the aforementioned setup “Me BI.”
Enough About You, What About Me?
“Me BI served its purpose of bringing productivity to individual users,” said Kohl. But it has its limitations, he explained. “It is a more individualistic approach to BI and analytics that leads to disjointed, inconsistent, siloed data.”
Wayne Eckerson, a consultant and author of books like "Making Data Governance Work: Tales from the Trenches" and "Secrets of Analytical Leader Insights from Information Insiders," may very well agree with Kohl’s postulate. In a blog post about Tableau he wrote:
“Like a wild vine, if left unchecked, Tableau will strangle your BI program, creating ungoverned spreadmarts that increase your support costs, undermine data consistency and waste your staff’s time reconciling reports.”
Needless to say, Tableau would more than likely counter this criticism.
Though we didn’t point the Gartner leader specifically to Eckerson’s blog, here’s what Tableau VP of Marketing, Ellie Fields told us:
“We believe strongly in putting the right tools in the right hands at an organization; sometimes, it’s the entire company that deploys Tableau at once, while other times it may be one small department within the big picture,” she said.
Fields pointed us to two different scenarios.
“One of our customers, Deloitte, gives Tableau to each new analyst they hire, nearly 10,000 people across the company. Other customers initially deploy the software in one department where it can create immediate value, and then it grows from there,” she explained.
The latter is a “land and expand approach” that Eckerson takes exception to. “It can create silos,” he said. “When people pull data from different sources they get different answers,” he noted. Not only that, but it seems that it can raise costs as well. Buying 2,000 desktop copies may very well be more expensive than IT making a deal for the enterprise.
This doesn’t surprise Jayne Landry, global vice president and general manager, Business Intelligence at SAP. SAP’s Analytics team worked closely with customers while they built Lumira, SAP’s self-service data visualization solution, and heard about the pains they were experiencing.
“Customers shared with us that other data discovery solutions in the market delivered the agility that business users were looking for but often at the expense of trust. In some cases organizations had lost customers as a result of decisions made on inaccurate information, in other cases data discovery was 'running rogue' and leading to confidential information going public. Many organizations were becoming increasingly concerned at the costs associated with maintaining these systems. One customer commented that the total cost of ownership of their data discovery deployment 'was insane,'" she said.
SAP, which stands out as leader for “completeness of vision” in Gartner's BI and Analytics Platforms MQ wouldn’t be talking about this if they hadn’t built a solution that eliminates such problems. Landry said that SAP delivers “the agility business users are looking for but together with trust, security and scale for the organization.” More specifically, Lumira was designed to tailor the data discovery experience to users with different skills levels.
Business Analysts, for example, can easily access any data and share visualizations on-premises, to the cloud or even to a public website as required. On the other hand, businesses may provide a broader set of users with a more controlled and tailored access to trusted "approved" data sources and possibly a "sandbox" where the work can be reviewed before it is promoted more broadly.
Kohl would likely make similar claims. He sees problems with “desktop tools that typically fly under the radar or are disconnected from platform infrastructure,” adding that they invite risks of access and security breaches. "They can also limit performance, and ability to share not just analytics content but also business logic, data models, and sophisticated queries,” he said.
Do Cloud Services Change the Game?
But DataHero, a (relatively) newer analytics startup founded in 2011, approaches BI a little differently. Chris Neumann, the company’s Chief Product Officer and founder explained that DataHero is built on the security of the company’s partners which include cloud service providers like Salesforce, Marketo, Box, Stripe and at least 16 others.
“Permission is given by IT. DataHero builds on top of IT,” said Neumann.
And when it comes to the “land and expand” approach used by vendors like Tableau, Neumann said that DataHero is “all in.”
At the end of the day, it’s not a question of whether self-service analytics are in or out — everyone agrees that they’re here to stay. Where the differences among analytics vendors seem to lie is how issues like governance, cost containment and control are achieved; whether initiatives belong to end users or IT; the advantages and disadvantages of enterprise-wide vs. individual and small teams solutions; and what is gained and lost in each.
Eckerson said that a tugs-of-war of sorts is in play at the moment. Businesses struggle between the need for speed and the need for standards, as well as the need for agility and the need for governance.
At the moment end users seem to be calling the shots, but it’s unlikely to stay that way forever. Vendors are working hard to address Corporate BI’s and IT’s concerns without tying anyone’s hands.
If they fail, corporate CEOs and CFOs will have no choice other than to get involved.
Title image from Sidney Lumet's Network