Self service BI, dead or alive?

A brief stir erupted several Friday mornings ago when Yellowfin CEO Glen Rabie declared to the Boulder BI Brain Trust that self service business intelligence is “dead.” It was one of those statements that makes you sit up and listen — one that leads you all the way to a surprising observation that now drives Yellowfin strategy.

If there had been a death, a body would surely have shown up, at least reports of missing solutions. We would have had to contact the family, starting with Tableau, Spotfire, and who knows who else. It’s a big, busy clan.

But this industry will never produce anything as cold and hard as a dead body. There was no body, and there was no death. And self service BI is no more dead than self-service email, phoning, dressing, driving, or cooking.

I followed up with Glen, whom I knew from the recent Pacific Northwest BI Summit, to find out what’s going on.

He knows self service BI still lives. What he wants to kill is the widespread assumption in the BI industry that self service is what everybody wants.

My rendition of the industry’s dream goes like this: If we could only show those holdout knuckleheads the beauty, the insight, the oh-my-god-I-didn’t-know-that amazement of data analysis, then simply everyone with a brain would sign on.

“Most people who use analytics want to be consumers of a service that is provided to them,” he write to me in email from his office in Melbourne, Australia. “Building it yourself is not convenient.” Google Analytics, for example, is not hugely insightful, but it costs little.

“If we start with the assumption that most people never want to write a report, and never want to do self service analysis,” he writes, “it will force us to innovate and think about alternative ways to deliver BI.”

What about the roaring success of self-service tools?

They’re for data analysts, he says, who’re willing to handle the tricky backend part of data preparation.

“To do BI really well is a special skill,” he said in a call a few days ago. “There are whole books written on what charts to use. You’re just a generalist prior to knowing that.”

Yellowfin discovered this, he said, after trying to serve the self-service market. “We really did believe Yellowfin was for everyone. We thought everyone wanted to write reports,” he said. But they found too little demand for that.

“In reality, customers didn’t want to do data.” Yellowfin began an 18-month shift to “productionized mass distribution of pre-defined analysis.” They’ve broadened the collaboration features, made a more engaging user interface, and built in device independence. In November, they say they’ll introduce storyboards.

We’ll see, but he might be right about the average user’s preference, at least among today’s workforce.

Next questions: How will the specialists be organized within the enterprise? What skills will they have? How will organizations prevent the slothful response that still drives much of the demand for self-service? If Glen’s right, these questions — already the most interesting in BI today — will be even more important.


  1. Ad Stam says:

    Self service BI has never lived. It was from the beginning a waste of time and effort. IT overestimated from the beginning of data warehousing the willingness of users to do their own “programming”. I once (and that was even before the turn of the century) heard a very senior partner of a very large and well reputated consultancy firm say “Oh let the users just learn to use SQL and all their problems are over”. Only 10-20% of users are willing to spend time & energy in more than just thinking about the information they (or someone in their vicinity) need. And that was already the case in the 70ths and has never changed. IT should concentrate on improving the speed and quality of the information logistic value chain between source and screen. The business should invest in improving Master Data Management and Data Quality. That is the only way forward to speed up and improve information delivery to users.

  2. Can see both sides, would disagree that 10-20% of users are willing to use SQL, at most 1%. Whilst most folks are only interested in consuming content, changing behaviour of users to sit up in the chair and peek and poke a report. Some more capable users can do this heuristically without an application, or with a quick calculation manually or in Excel. Changing behaviour to take this to the next level with tools to be able to insightfully answer questions is for me why self serve BI not not yet dead or really gotten started. Watch this space!

  3. Qwalytics: Self-Service Business Intelligence and Natural Language Reporting says:

    Very nice article, I used some of your useful concepts to write my own article: Self-Service Business Intelligence and Natural Language Reporting. I’d love to get your feedback. Don’t you think that Natural Language Business Intelligence is the next step of Business Intelligence’s evolution?

Leave a Reply

Your email address will not be published. Required fields are marked *