The Business Intelligence of Putting BI in the Cloud

    Slide Show

    Five Tips to Improve Your Data Analysis

    If you’re working in one of those companies where the wonks do all the analytics and the idea of delivering insight from those analytics to business users is kind of foreign, you’d do well to spend an afternoon chatting with Brad Peters. His head is in the cloud, where that notion isn’t foreign at all. In fact, it’s the whole point.

    Peters is CEO of Birst, a San Francisco-based SaaS provider of what it calls “enterprise-caliber BI born in the cloud.” Peters and his colleague Paul Staelin cofounded Birst in 2005 after leaving CRM software provider Siebel Systems, where they spawned the company’s analytics business. They left Siebel just a few months before it was acquired by Oracle, because, as Peters put it in a recent interview, “the writing was on the wall.”

    In that interview, I pointed out that all of the big BI vendors are pushing their cloud technology. So I asked Peters what difference it makes that Birst’s BI was “born in the cloud.” He said it makes “a ton” of difference:

    My first experience at Siebel, before doing analytics, was driving a product line called, which was Siebel trying to do cloud. My job was to take Siebel Enterprise CRM and put it in a cloud application. Let me tell you, the thing was not designed for cloud. It was horrible. We tried to do two things with it. We tried to host it, to just run Siebel in a hosted data center, a la MicroStrategy or Business Objects. This ASP model didn’t work. Running Siebel hosted in someone else’s data center, where you had to pull all the same levers that you’d have in your own data center, doesn’t buy you anything—you still have bodies doing all the same stuff. There’s no reengineering of the application. There’s no economic transformation that’s occurring. You’re just shifting where the servers are running. So that’s not a huge value-add. That didn’t work for operational applications, so I don’t think it’s going to work for BI. It’s already killed companies like Oco [which was purchased by Deloitte in 2011], and a bunch of folks who kind of ran with that model. It has not proven successful so far. Business Objects On Demand hasn’t worked for a while; MicroStrategy is making a go at it, but it’s more of a managed service—you’re using it as if it’s in your own datacenter, and it’s just as complex. It’s fine, if you want the pain in someone else’s data center, where you have even less control. But it’s still every bit as much pain as you had before. The hosting thing is not the value.

    Peters explained that the value lies in how the application is engineered:

    All of those on-premises applications were fundamentally engineered originally as desktop tools, then they got ported to a server as an afterthought. They were not designed for multi-tenant, cloud-based applications, with automatic upgrades. Try upgrading Business Objects. It’ll kill you. With Birst, we built all of that application management infrastructure into our architecture from Day One. It’s just a very, very different model. The funny thing is, failed for multiple reasons—not just because the software was really hard to implement in the cloud, but because we lost two-thirds of every release cycle just keeping up with the base releases. So that killed our productivity. Beyond that, our entire sales structure in the organization was geared around perpetual licenses and large deals. Our sales guys couldn’t sell cloud software. They couldn’t sell subscription models, which means they didn’t really understand how to maintain a customer after they bought the software. So all your reps at MicroStrategy, all your reps at Business Objects, care less about the customer after they buy that. One of the main values of cloud is the vendor stays with you after you buy the software.

    Call me a SaaS novice, but I was surprised to learn, as I was preparing for the interview, that Peters had said Birst makes no money in the first year with customers. So I asked him to explain how that’s a viable option. He said if you only have them for a year, it’s not:

    Look, we’re investing in our customers, and they’re investing in us. BI is, to some extent, a shared investment—it’s not a finished good. This is one of the things people don’t understand about analytics—every customer’s implementation is different. You’re essentially creating an application around a business use case each time, so there’s shared investment there. It creates aligned incentives in the process. It’s true, for most SaaS companies, that there is a payback period of time, and it’s generally around a year, so we’re not that different from any other SaaS company in that sense. So as a SaaS company, you’re fundamentally trying to keep that customer satisfied post-sale for more than a year. If you do that in BI, where a lot of projects fail at a disproportionately high rate, then you can deliver more value than any of the on-premises guys can deliver, because you’re actually thinking about these customers a year afterwards. We’re highly incentivized to make sure our customers are successful afterwards. We’ve got a lot of product innovations, but we’ve also got a lot of fundamental business-model innovations—we’re really trying to rethink how analytics are delivered, because historically, the vendor is not very well aligned with the customer.

    I asked Peters who within a company tends to be the biggest champion of what Birst is doing. Not surprisingly, he said it depends.

    In different segments, we tend to have different types of champions. For mid-size organizations that don’t have large, preexisting BI and data warehousing investments, we see a broad spectrum. There’s IT, who doesn’t want to build a highly- specialized BI practice, because they get to have essentially enterprise-caliber capability without having to build a team of eight to 10 people to support it. So for mid-size organizations that don’t have that huge emphasis and specialty, it’s a huge win. In large companies, CIOs have very little incentive to drive business value. They get a ton of headache and pain when business users try to do more, so they really want to contain their projects to very few things. What that means in my industry is that the vast majority of business users in big companies are completely unserved by those CIOs in central organizations. So we tend to deal with the departmental business leaders who need access to their business information. They’re turning to cloud as a way to take some control of their own destiny, to be able to have that business insight they need. The CIOs generally go along with it, because they’re not going to get beaten up. The business leader is taking responsibility for it, so the CIO is not going to get beaten up if something goes wrong.

    So whether an IT department tends to fight the democratization of BI, or is inclined to embrace it and facilitate it, again, depends:

    The bigger the company, and the more established that company is in terms of IT—any company that quotes to you how many developers they have internally is going to fight democratization of BI. On the other end of the spectrum, you’ve got companies that are newer, younger, nimble, growing fast, that have leaner, meaner IT staffs—midsize organizations, technology companies—they actually like it. Some of our best champions are in IT, but they work in companies where IT doesn’t exist just to feed IT.

    The bottom line, Peters said, it that Birst is out to help people understand that Big Data and analytics are just as important for the average business user as they are for the data scientists:

    I think there’s a much bigger opportunity out there, and that’s where there’s been a lot of confusion. Right now, there’s a lot of emphasis on these highly technical tools for highly technical people. But there’s a huge opportunity out there for people to deliver those insights to non-technical folks. And that’s really what we’re focused on.

    Latest Articles