Visier CEO John Schwarz: decentralized BI is not the answer

Profile picture for user jreed By Jon Reed October 9, 2015
Visio CEO John Schwarz isn't just critical of "big BI." He's also got issues with decentralized BI. In part two of our series, we find out why.


In my first piece with Visier CEO John Schwarz, he didn't hold back on the problems with "big BI" solutions, and why he left such companies to co-found Visier. But he had equally critical things to say about the rise of "small BI."

In this piece, we'll get into his critique of the decentralized BI solutions that have thrived in the wake of legacy BI solutions. Schwarz also provided a detailed looked at workforce analytics in action, and how one of his customers is addressing a common talent issue: the problem of attrition, and recruiting a diverse talent pool.

What's the problem with decentralized BI?

Schwarz sees the centralized/decentralized BI trends as cyclical:

The hype cycle around analytics is that we go through these phases where we oscillate between centralized and decentralized approaches to solving problems. We typically start with a decentralized approach because it's simpler, it speaks to a specific end user, and it solves that user's problem.

We tend to idealize decentralized BI solutions as "empowering users," but Schwarz advises caution:

The user buys a decentralized BI solution, they evangelize and propagate the solution throughout the enterprise, and then pretty soon everybody has that solution. The problem with it is that everybody has a solution, but they don't talk to each other. They cannot be managed, they cannot be correlated and it's a mess.

Up until the last few years, the pendulum was in favor of centralized BI:

Then the pendulum swings and people start to aggregate and consolidate and centralize and manage, but the managed world gets to be unwieldy, it gets to be too heavy, right? People can't get individual needs solved, so we start, again, decentralizing. Well, we're at this phase where we've had Tableau and QlikTech and Domo and other players out there preaching the decentralized solution, which is great for the individual user, because it's flexible and nimble.

Actually, that sounds pretty good to me Mr. Schwarz. So what's the issue? He adds:

It sucks at the enterprise level because it's impossible to manage, you cannot correlate, you cannot get consistency, efficiency goes to hell and you end up in a mess where people argue about whose data is what. We're one of the companies that's coming in to put some structure around the rather crazy, unwashed, bubbling mess of analytics that are being done either on spreadsheets, or on some of these simpler analytic tools.

Democratizing BI is the (clear throat) "cool vendor" way to go. But the way Schwarz was talking, I figured he'd feel different. So does Visier want to get their analytics in the hands of all HR users? Short answer: no. Schwarz:

We aim at about ten percent of the customer population to be the right user base. In essence, everybody who's a manager of people in any organization is the right target for us. Not every customer goes there day one, but ultimately we hope that they all will.

At last, a vendor that isn't trying to "democratize dashboards." Schwarz:

The domain that we're solving now is about people. The data that we have is pretty sensitive; you don't want it to be kind of completely available to everybody. Our customers, which range in size from two thousand employees at the low end to ninety thousand employees at the high end, identify the qualified, authorized management folk to understand what's happening in their organizations.

John Schwarz

Use case number one: employee attrition leads to diversity setback

It's always helpful to hear vendors say exactly what they do, and what they don't. But I wanted a specific use case. Schwarz started with an example of employee attrition. Attrition is always a problem if you're leaking talent, but it gets worse when your diversity goals are undermined. Schwarz provided an actual example that he extrapolates here:

Let's say you're a company that is trying to improve the diversity of its employees; a pretty important objective these days, and in many industries, demanded by regulatory rules. So you set certain recruiting objective to hire a set number of diverse candidates... Then, after you've spent all this money, at the end of a year, your results are no different than they had been. Despite the fact that you recruited with preference for diverse candidates, and you have promoted a preference for diverse candidates, you lost them through attrition. So what do you do?

Schwarz says their solution has helped customers like this one:

We can tell you not only who's leaving the business - which most applications actually do track - but we can tell you why people are leaving the business, and when they are leaving the business relative to their career or relative to their new job.

We can tell you which managers have a greater propensity for losing people, which locations have greater propensity for losing people, what are the characteristics that describe the people who are leaving, and which characteristics are common, so that we can start growing some inferences from the kind of people that you're losing and why. Then we apply that knowledge about the people who have left the company, and we use that knowledge of the characteristics of people who have left and apply it against the people who are still in the company.

Fine - understanding why valued employees are leaving is useful. But what about taking action going forward? Here's what Visier did for this customer:

We gave them a ranked view of the people who are at risk of leaving tomorrow, because they happen to share the same characteristics with the people who have already left. This is now actionable information, because going back to my diversity recruiting example, if you have certain managers who lose diversity candidates because they are racially insensitive or gender insensitive or have some bias, you can find who these managers are and deal with them.

Use case number two: losing an aging workforce in oil and gas

An oil and gas customer was facing a different predicament: an aging workforce that was taking their know-how out the door as they retired. Schwarz:

The oil and gas industry has an interesting problem - and for that matter so does the energy industry in general. They have been in a fairly depressed economic cycle for a while. That depressed economic cycle has limited their ability to hire, and so they've had very limited hiring; typically something in the vicinity of two or three percent. What they recently realized is that the Baby Boom generation that has all the skills necessary to drive a nuclear reactor or to understand a seismic test are going to retire within the next five years.

The obvious answer is beefing up skills training, no? Ahh, but there's a problem: it takes ten or more years to teach younger workers to be effective in these roles. For the customer in question, that means there is a five to ten-year gap looming:

They knew they were going to lose people to retirement, and they didn't have enough people in the pipeline to backfill those folks. How do you deal with that? Well, you have to devise very specific programs to offer retiring employees the opportunity to stay beyond the normal retirement age.

But surely this customer was already aware of this problem? Indeed they were, but their spreadsheet-based approach to coping with it was failing:

They spent about three years trying to build a set of spreadsheet analyses that would help them to identify the key employees, filter out others and then try to build policies. But using spreadsheets for this is a nightmare. The problem with spreadsheets is you really can only test two variables at a time; it's a two-dimensional solution.

The other problem with spreadsheets is that anybody who gets one can change any of the assumptions, and you never know whether you are looking at consistent data. Having spent three years breaking their back with spreadsheets, they finally concluded it was simply was not doable. They needed a more sophisticated analysis and planning tool.

Visier took a different approach:

We've built very automated and productive scripts or processes that allow us to collect data from any source the customer may have, whether it's their HCM application or their payroll application or their performance management application or learning management application or whatever applicant tracking application, and we ingest the data from these disparate systems and populate our schema with these data streams. We do that remotely, the customer doesn't have to come on our site and we don't have to come on their site.

And no more employee-tracking spreadsheets? Schwarz: "No more spreadsheets." And a result: a more successful retirement postponement initiative. While we didn't have time to get into the numbers on that one, Visier has done their diligence with detailed, documented case studies and the numbers in cost savings, etc. are cited.

Final thoughts

Schwarz's frank analysis of the problems of BI and HR are worth a listen. How much of a dent Visier makes in these workforce issues remains to be seen. But with around sixty large enterprise customers and counting, they are certainly in the game.

For now, Schwarz seems to be enjoying the ride, happy that he made the leap from corporate exec to startup-with-gravitas (and heavy lifting). He's doing his best to embed his career lessons into Visier's own culture:

I don't want employees at Visier; I want partners and co-owners. I want people to feel that they own the company, that they are a part of something that is really special... We hire a lot of fairly young, fresh graduates. To inoculate this kind of culture in somebody who is looking for a job is an interesting challenge. They don't necessarily understand what we're asking for at first. We want them to feel like they are making a difference every day. They are asked to make decisions every day, not just be dependent on being told what to do. When they get it, it's wonderful to see.

Image credits: Head shot of John Schwarz provided by Visier. Feature image: mädchen mit leiter schreibt formeln an die tafel © contrastwerkstatt -

Disclosure: Diginomica has no financial ties to Visier. Workday, SAP and Oracle are diginomica premier partners.