Business intelligence is an unsolved problem. At every company we’ve ever worked with, managers lack the information they need to make intelligent decisions. Why is this, after the rise (and eventual fall) of an entire industry around business intelligence technology? Will Cloud Computing change the sorry state of business intelligence for customers?
2008 saw a number of next-generation BI providers move their offerings from niche to mainstream. Companies like PivotLink, LucidEra, Good Data, and Panorama announced a maturation in their offerings and important customer wins. Even Business Objects and Cognos were able to advance their on-demand offerings from within their on-premise parents.
But most business still lack basic access to critical information. For most people in most companies, it takes a request to IT (and perhaps a call to your implementation parter) to do something as simple as add a field to a report. In our eyes, this is a market ripe for disruption from cloud computing in 2009.
Despite the fact that on-demand business intelligence has been slow to take off, there are fundamental problems in business intelligence that, in our view, can only be solved through cloud computing. We believe that the increased availability of cloud computing platforms in 2009 will make BI the next application category to reach the tipping point of on-demand adoption, fueled by data from SaaS applications.
BI is a classic “bursty” application. BI requires infrequent access to massive computing power-- a perfect application for cloud computing.
Thanks to Google, users have been trained to expect nearly instantaneous access to any piece of information. Instead of spending hours formulating a carefully crafted query to run once, users expect to be able to navigate iteratively through their business information. They expect to run an initial search, see what comes back, and explore from there.
This is simply not possible with any legacy BI system: vendors brag about achieving sub-minute response time… a far cry from the sub-second response time that an interactive application requires.
As a result, BI's impact in the enterprise has been limited... only a few power users are able to use it effectively. The impact and value of BI would be exponentially greater if business users could search and navigate their company's data in the same way they search and navigate the information on the internet.
Why can't in-house IT departments deliver this level of performance and usability using legacy BI systems? In part, the issue is infrastructure. BI infrastructure is overpriced and tremendously under-utilized. Many BI systems sit completely idle 90% of the time, only to be hammered on well beyond capacity in the days (or hours) leading up to a management meeting or end-of-quarter analysis.
These are perfect conditions for the shared infrastructure of cloud computing. Shared computing power is the only way that companies can deliver the responsiveness their users demand without a prohibitive investment in hardware. On-demand BI vendors, with their roots in modern web-based technologies and user interfaces are well positioned to create usable BI applications that will finally unlock the value from companies' transactional data.
Why BI from SaaS data?
We also predict that the winner in on-demand BI will emerge from an initial strength in analyzing data from SaaS business applications. Why? Analyzing data from SaaS applications overcomes 3 of the key barriers to adoption for on-demand BI: Security, transformation, and integration.
Security: Your business data are your crown jewels, and nobody wants to be the first to put that into a shared infrastructure. That’s why we think that data generated by on-demand business applications is the likely starting point for on-demand BI…. After all, this data already lives in shared environment.
Transformation: Query, reporting, and analysis is not the hardest part of BI—getting data into a format suitable for analysis is a far harder challenge, especially for an on-demand solution that’s trying to bring together information from behind the firewall. This is less of an issue for BI running off SaaS applications where the data is already in the cloud.
Integration: Most analytical applications require bringing together business data from multiple sources. Most modern SaaS applications offer more open APIs than on-premise software.... on-demand BI has the potential to be better integrated, and therefore more valuable, when paired with a SaaS application.
Implications for customers
Of course, it will be difficult for customers to justify investing in an entirely new BI infrastructure, especially since most are still trying to justify the investment they made in their last round of on-premise BI infrastructure and 2009 looks to be a year of frozen or slashed IT budgets.
That’s why we expect to see on-demand BI enter the enterprise from the line of business, not the IT department. Business users will be fed up by the inability of their IT department to support their basic demands for information. If they’re already using a SaaS solution, they’ll be tempted to try an on-demand BI solution that they can get running without IT support, out of their operating budget (exactly how on-demand CRM initially penetrated the market).
This is what we expect to see in 2009: Low ticket, low risk, on-demand BI solutions, built on cloud platforms, with adoption driven by business analysts hungry for information.
Our advice for enterprise IT? Let it happen. There isn’t yet a clear winner in on-demand BI, and you have much to gain and little to lose from these experiments. These are not requests you’ll be able to service in 2009 anyway, and early experimentation will leave you well positioned to jump on a winning solution.
What do you think?