RESEARCH & RESOURCES

Q&A: Centralized Governed Architecture Needed for Today's BI

As BI expands its footprint across the enterprise, centralized management and governance are key.

As business intelligence tools evolve from departmental to more mature enterprise-wide deployments, a tool that allows centralized, governed management while still meeting user needs is essential. "BI today has evolved from a nice-to-have to a must-have," according to Omri Kohl, co-founder and CEO of Pyramid Analytics. "The big challenge that organizations are facing today is how to consolidate the silos or departmental solutions into one standardized offering?"

BI This Week: The BI market is clearly evolving. Can you talk about your perspective on that evolution? What are you seeing in terms of users and in how BI is perceived and used in the market?

Omri Kohl: Our vision in launching the company hasn't changed -- actually, the market has [moved toward] our approach. We see productivity tools merging into the organization, to the point that companies like Tableau and [TIBCO] Spotfire and [Qlik] QlikView have a solid presence within companies. We continue to see the need for enterprise and corporate applications to consolidate and standardize the way their users access data.

How do you see the market in terms of types of BI users?

[Here at Pyramid,] we differentiate users as follows: we have the analyst -- the person in the sales department, or engineering department, or marketing department who is creating reports for colleagues or managers. That person is looking for a tool that allows more sophisticated analytic capabilities.

Then there is the consumer -- the user who is not interested in doing calculations but who also doesn't want to simply get static reports. That user wants to get something to slice and dice, and then ask the next question but they don't want to do that calculation themselves.

Finally, we have the BI developer -- someone who works in a business intelligence department. That's definitely a trend we see developing very rapidly in the enterprise market in larger organizations. We see BI departments and excellence centers, all with people who develop advanced analytics, business logic, and models -- all of which is being consumed by the analyst.

Last but not least, we have a typical IT department that oversees the entire organization. The challenge for those IT workers is how to make sure that the data remains secure, that our tools are performing, that our users get the performance they need, that the system is stable, and that data integrity is ensured.

How does that mapping of types of users translate into user needs?

At the end of the day, we see a map of users with different needs and different demands. We all talk about self-service, we all talk about one version of the truth, data consumption, and ease of use, but for different users, these phrases mean different things. This is where an enterprise BI platform comes into play, and this is what has happened in the last few years. There is clearly a need for a BI platform that gives all those different users what they need. It needs to do that without taking away the innovation from the end user, but also without undermining the IT investment in the data warehouse -- the security infrastructure, the way they would like users to access data, the way data is distributed inside the organizations. The pendulum actually has moved to consolidate and standardize the way users are touching data.

This is what Gartner calls "governed data discovery." We keep innovation in the hands of the users and give them the tools to do data discovery -- whether you are the most sophisticated person in the organization and need to do advanced analytics or you are a BI developer who wants to build models and business logic or you are simply a data consumer and you want an interactive dashboard to see the performance of your department. At the end, IT is going to assure that everyone has a tool that works and is stable and secure.

On top of all of that, you need to satisfy users who have been educated about what BI can be, [who expect tools] that have advanced features that don't require coding, and don't require developers to provide those capabilities.

All this needs to be wizard-driven. It needs to be a click away versus a developer away. It needs to have various publishing tools included -- it should have a way to publish to a dashboard, to a website, to e-mail, even alerts to a phone. All of that is BI.

What we are seeing today is a market that is getting very mature. The demands get greater and greater. The need suddenly became more structured and more organized within each organization. Many companies now actually have a road map for BI and a center for BI excellence. You can compare BI to more mission-critical technologies like ERP now. I think BI is on its way to becoming mission-critical, if it isn't already there.

I think BI today has evolved from a nice-to-have to a must-have. The big challenge that organizations are facing today is, how do I consolidate the silos or departmental solutions into one standardized offering?

You mentioned delivering data to phones. How are mobile devices changing things in the BI market?

It's an interesting question. When we introduced our mobile offering, we definitely studied the market to determine the right offering. It was simply a checkmark back then. You had to have mobile. In the last two or three years, however, you can see the trend of bring-your-own-device growing. People are coming to work with iPads, iPhones, Surface tablets, and other devices. Both internally and externally, it has become a huge trend, and not just for BI. People are becoming more connected. Maybe you don't even show up for a meeting any more with your laptop, because you have everything you need on your iPhone.

What happened that made the adoption very rapid is that you could suddenly become the smartest person in the meeting room. In a meeting, if the boss asked for our previous year's results for product A in territory B, all you had to do was open your BI platform on your iPhone and give the answer instantly.

Mobile has actually proved to have the ability to, first of all, make the people who use it the smartest people in the room, which drove a lot of adoption. Second, looking externally at people in the field -- and this goes back to my earlier comment about a BI platform versus a productivity tool -- when you have a secured BI platform, and you can control what data different tools can access, you actually feel very comfortable in letting people who are on the road [access data]. When it's a BI platform and it's part of the overall BI strategy, then why not allow access? It's secure, it's stable, and it performs.

When you make data more easily available to users, governance can become more of an issue for IT. How do you balance that -- data governance versus ease of access?

Governance seems to be a word that people avoid -- it sounds like policing what users are doing. However, in the context of BI, governance is actually a very positive aspect. If we take the word "governance" and replace it with "BI platform," we can see that we are actually serving users at the end point, making sure that what they need to do is accessible.

If you work outside your organization's platform, one of the challenges is accessing your organization's data. Essentially, you're bypassing the investment the organization made in building the model, building the warehouse, and collecting information for years or even decades. That data is highly valuable. When you look at real estate, or health care -- that data is a huge part of data-driven companies in those industries. When you are not using it in a centralized manner, you lose much of your ability to understand what is going on in your organization.

When you talk about governance, you want to be talking about an integrated end-to-end platform. It's actually a tool that serves every user, as we discussed earlier. It should be easy to access your organization's data from any device. For that to happen in a secured, architected way, it needs to be managed by a centralized IT department.

So centralized management is important. What else?

Time to market is also critical. If you want to do across-the-board analysis for various users on various subjects, you need to have the ability to access all of your organization's data -- not just what you've created on your desktop, or in your spreadsheet, not just something that you've collected from the Web. You want to easily and immediately consolidate different data sources that reside on-premises, and publish them for consumers. That's something that is possible only in a centralize, governed environment.

Here's another thing that's important: collaboration. How do you collaborate with your colleagues? How do you publish information to them? How do you share information? What if you go to a brainstorming session and everyone brings their own numbers? When you ask for productivity tools, what if everyone comes with their own tools? If everyone uses different sets of data, how do we even agree on a number?

At the end of the day, collaboration and consolidation is critical. That's also being done only in a governed, centralized deployment.

When we talk about delivering a governed BI solution, how is that done in terms of architecture and execution?

When you talk about execution of a governed BI solution, you're talking about a multi-tenant, private architecture that allows every department and every user to have their own department and access to the data.

For example, you also want to manage load balancing so you can make sure that different queries in big organizations can be managed. You want to make sure that when one person hits the "Run" button to perform a massive query, he doesn't [affect] performance for everyone else. You're talking about load balancing, large query engines, and tools that make the administration of the platform easy and accessible. You need an administration console that governs loads, security, profiles, and access to data. You also want to make sure that you can on-board new users very easily. You want to make sure that the next user is a click away, versus a deployment away. ... Managing and administrating users is key.

If you need to deploy two users or 20 users, it's easy. However, when you need to deploy 20,000 users or sometimes even more, you cannot go one computer at a time. You need a centralized tool. ... With a centralized tool, you can share business logic. Someone can create a calculation ... and many users can access it. We want to make sure that everyone can use it, and in a centralized environment that can happen.

How is Pyramid Analytics positioned in today's market?

Everything we've talked about today is what Pyramid is all about -- a very big, robust, governed data discovery tool. Our vision all along was to find a way to allow organizations, particularly enterprises, to incorporate and to standardize their BI tools. To do that, you not only need a strong back-end tool that supports IT, you also need to be very compelling to a variety of end users with different needs and objectives. We had to build a tool that would both address the needs of IT, but on the other end, would be innovative, easy-to-use, driven by self-service, and code-free for users. At the end of the day, really, a successful BI platform today needs lots of substance that is compelling to everyone.

TDWI Membership

Get immediate access to training discounts, video library, BI Teams, Skills, Budget Report, and more

Individual, Student, & Team memberships available.