Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.
March 9, 2010
7 Min Read
For more on business intelligence, see Intelligent Enterprise.
Devices like the iPhone will slash mobile BI query speeds -- as long as the underlying BI architectures don't get in the way. So says MicroStrategy Chief Operating Officer Sanju Bansal, a 20-year company veteran who has more than a few bold things to say about performance, analytics, dashboards and the competition.
InformationWeek: MicroStrategy highlighted mobile BI during its January users' conference, but we haven't noticed much uptake for related products. Do you know something we don't?
Sanju Bansal: I agree that the current state BI mobility is a yawner. When you look at the current tools, including our own, they tend to present simple graphs and grids. But after studying what's happening with smart phones, it's clear that Apple, with the iPhone, is picking up momentum and market share pretty aggressively. Apple has shipped about 70 million iPhone and iTouch devices, and they will probably ship another 50 million units this year. Within two years, there will be more than 200 million devices in the market that run that operating system and browser stack.
What makes these smart phone devices interesting but that is not well understood is that the impedance to queries is dramatically lower. Not only do they know where you are because they have built-in GPS capabilities, they also have built-in cameras and other sensors. If you're a manager in a retail store, for example, the device already knows where you are and you can take a picture of the barcode on any item with the camera; with barcode lookup services, the device knows what the item is within about half a second. Before, if you wanted to look up the sales or inventory of a product, you had to look up and enter the product name or code and you had to enter your location.
The challenge is going to be performance, because people don't want to wait much longer than about two seconds for query results. The dirty secret in BI is that performance is still a problem. In most corporate environments, query response times are anywhere from 20 to 45 seconds. That's obviously way too long for a mobile device.
InformationWeek: So what is MicroStrategy doing about the problem?
Sanju Bansal: When we launch our new mobile offering for smart phones in the second half of this year, you'll see deep integration with native sensors. Second, you'll see tight integration with other software on iPhones and other devices, like mapping. We're also working hard on the BI performance part of the equation. Our goal is sub-five-second and, in many cases, one- to two-second response time.
InformationWeek: Much of the talk about query performance now centers around appliances, parallel processing, in-database processing and the like. Is this really a data warehousing problem?
Sanju Bansal: The topics are related, but the BI vendors in general are behind because they all said, "it's a database problem." The truth is, there's a good way to model data in a database... but the way business users want to see the data is far different than that. Even if the query time against the database is zero, there is still a potentially massive amount or transformational time that goes through the BI tool set -- not to mention rendering and display time. The problem is you can't push that metadata-understanding and transformational work back into the database. So the BI tools have to get better.
InformationWeek: Better in what way?
Sanju Bansal: The issue with BI architectures is that they have been client-server architectures that have been evolved to the Web. They are loose architectures that are not as tight as database kernels or operating system architectures. The energy and effort devoted to efficiency has been somewhat low I think there will be a realization that all the BI vendors need to rearchitect using the same techniques that the database and operating-system coders use to gain efficiency. We'll see that become a big battleground over the next few years.
InformationWeek: How far along is MicroStrategy in addressing architecture?
Sanju Bansal: We just announced MicroStrategy 9 version 2, and I'd say 80% of the effort in the release was performance related rather than feature related. In talking to Gartner and the other analysts, they tell us we're the only BI vendor talking about performance right now. Either we're missing something, or we're the only ones you can talk about it because we decided to focus on performance two years ago. But forget what I say. For customers, performance is the number-one topic. You see that in industry surveys year after year.
InformationWeek: In a recent InformationWeek Webinar, Joseph Openshaw of discount retailer Meijer was raving about your analytics capabilities. Many think of MicroStrategy as a BI vendor with strong ROLAP, dashboarding and scalability stories. Why don't we hear more about MicroStrategy analytics?
Sanju Bansal: With the exception of SAS and [IBM] SPSS, we've long had the best analytics story among our BI brethren. Even compared to SPSS and SAS, we do a lot of work they don't do in terms of set analytics and set comparisons, which can be used to compare one set of stores to another set or to compare target markets by demographics such as age, income or race.
The quick story is that MicroStrategy customers have had -- and we ship embedded in the platform -- a few hundred analytical functions, including statistical analyses, time-series analyses, and set analytics functions. They are just part of the platform and we probably under-market them. Most customers do take advantage of the set analytics features and love them. They don't often take advantage of the advanced predictive analytics features we have, like correlations, projections and so on.
InformationWeek: Why aren't customers using these capabiliites?
Sanju Bansal: A lot of customers are still struggling with the "rear-view-mirror" questions across hundreds of areas of their businesses. They haven't had had the time or the guts to trust the data for more than simple, algebraic predictions. They might not use statistical techniques for forecasting sales or product outages, but if the sales velocity for a product is 100 items per week and there are 200 items in store, they'll know that they have a two-week supply. They are applying simple algebra rather than advanced statistics. I think most of our customers find simpler approaches to be adequate for 90 percent of their predictive needs. We've had more advanced capabilities for at least five years, but we have to evangelize to get customers to use the free features they already own.
InformationWeek: Other than performance, are there any other hot-button topics for 2010?
Sanju Bansal: The other issue we're hearing about from customers is dashboards. Frankly, I'm surprised at the lack of energy that [IBM] Cognos and SAP BusinessObjects have put into dashboarding given that it's the hot topic. This has created a lot of opportunities for QlikTech, which has done a nice job of exploiting the lack of dashboarding capability out there. A lot of people think of QlikTech as an in-memory BI vendor because that was its marketing claim. But the reason customers we talk to usually site for buying their software is that nobody else had offered good dashboarding. People want dashboards because they are sexy, they are easy to understand and people get them quickly.
InformationWeek: So how well does MicroStrategy stack up on that score?
Sanju Bansal: We're just as good or slightly behind QlikTech, but everybody else is a bit asleep on that front right now. Others are probably working on it, because I can't imagine that they are not hearing about it from their customers.
When I think about where we're going to make money in 2010, the story is dashboards. But what users care about most right now is performance. Where will we draw interest in 2011 and beyond? I'm becoming a huge believer that it will be mobility.
People are going to wake up over the next 24 months and realize that there are 200 million iPhones and maybe 100 million Google phones that are essentially portable computers in people's pockets. That's pretty much what happened with the Web in 1996 and 1997. By 1998, we found that all our customers were demanding Web-based business intelligence even though there was no demand the previous two years. It's a cultural shift that is happening all around us, but lots of people just take it for granted.
About the Author(s)
Executive Editor, Enterprise Apps
Doug Henschen is Executive Editor of InformationWeek, where he covers the intersection of enterprise applications with information management, business intelligence, big data and analytics. He previously served as editor in chief of Intelligent Enterprise, editor in chief of Transform Magazine, and Executive Editor at DM News. He has covered IT and data-driven marketing for more than 15 years.
You May Also Like