Traditional BI Systems A re N ot Des igned For A gility

The traditional approach to business intelligence (BI) has reached its limits. Over the past 20 years we have developed a set of procedures and technologies that allow us to aggregate,
cleanse, prepare, and report on enterprise data. While there have been incremental improvements in efficiency and speed over that time, the fundamental approach to BI has not changed much. During that same period, however, businesses and their information needs have changed dramatically. Enterprises have transformed themselves from rather isolated entities, often with an inward focus on operational efficiency, to players in geographically dispersed, multi-partied ecosystems that need to have as much understanding of their customer’s interests and partners’ operations as they do their own operational efficiency. The information they collect, transport, and analyze has changed accordingly and, more importantly, so has the speed with which they need to act.

Today agility is the goal of most organizations regardless of their industry and it is a top goal of business leaders whether they are in sales, marketing, manufacturing, engineering, customer care, or service delivery. Businesses are competing not just on efficiency, but on their ability to sense market conditions and quickly respond. That is exactly where BI and the business needs have diverged.

In many organizations, BI systems are used almost exclusively to generate standard monthly or quarterly reports. These reports deliver great value – certainly most organizations could not survive without them. But traditional BI systems require an expensive and time-consuming process to identify all the answers the business users ultimately want, build a data model that captures that information and unify data from disparate sources into that data model. Often, multiple cycles of this process are required before the needs of the business users are fully addressed. More than anything, a traditional BI approach is designed to structure data, ensure consistency across systems, and efficiently handle large amounts of structured data – goals that seemingly run counter to the need for agility.

Q&A on SaaS BI Technology

SaaS is a service model that sits on top of cloud computing environments and is designed to leverage the power of the cloud infrastructure. The software applications are accessed via a client such as a Web browser. They are managed and maintained by the SaaS vendor company removing much of the administration costs generally associated with on-premises software solutions.

SaaS and cloud computing are dependent on one another to bring their unique value to the market.

What types of projects are most often addressed with SaaS BI? What BI projects aren't good candidates for this technology?

The types of projects best suited for SaaS business intelligence are evolving quickly. Several years ago, the most basic functions of BI were the only ones best suited for the SaaS environment. Recently, SaaS vendors have taken great strides to deliver sophisticated feature sets and are now branching off into corporate performance management suites and even on-demand predictive analytics. Reporting and analysis still lead the way as the most heavily adopted feature sets within SaaS solutions, but as the technology continues to evolve, so will the demands of the SaaS customers.

Because of the obvious data integration challenges presented by SaaS Bi applications, real-time data is still difficult to leverage in a SaaS model. Leading integration vendors are starting to deliver new solutions that have greatly reduce data access time. As these innovations continue, real-time business intelligence will become easier.

One of the benefits SaaS promises is that it reduces an enterprise's IT costs. Is SaaS really less costly than traditional solutions?

In many cases, SaaS BI is less expensive than traditional on-premises software. Recent research by Enterprise Management Associates (EMA) shows that 76 percent of the organizations implementing SaaS have realized significant ROI. Upfront capital costs of hardware, along with the less expensive operating costs of SaaS, have contributed to these savings. There are exceptions, especially when the number of user licenses is high. Many companies have moved to a total cost of ownership (TCO) model when analyzing the impact of SaaS versus on-premises solutions. Many variables are involved in these scenarios and determine which direction is financially best for a company.

What other benefits can BI professionals expect?

There are many benefits to SaaS BI from a monetary standpoint. Both the capital and operating expenses of IT are reduced for most SaaS implementations. This upfront cost savings can translate into reduced risk for companies that are trying to experiment or deliver proof-of-concept (POC) projects. The elastic qualities of SaaS are a perfect fit for companies that experience seasonal highs and lows, as they don’t need to purchase additional hardware to handle the changes. Some studies have shown that SaaS BI solutions have fostered greater adoption among business users, adding to a more diverse business intelligence community.


You've discussed some compelling benefits. Are there any drawbacks to SaaS? Given that data is now "off site," isn't security a problem?

Security has always been a hurdle for SaaS. Customers are vigilant about securing their data as well as governance, regulatory, and compliance issues that surround it. The vendor community recognized this challenge early on and has addressed it on two fronts. The first is third-party auditing and certification. Leading SaaS BI vendors are SAS-70 certified and some have also attained Systrust certification. Both certifications relate to how the company controls client data and the systems and processes surrounding their data infrastructure.

The second front is innovation around data integration. Many of the vendors are finding ways to keep the data where it is while still leveraging the power of SaaS and cloud computing platforms. Data virtualization firms have also entered the market, providing trusted data federated environments that reduce the security risks of off-premise systems. In the end, the most reliable security feature for SaaS vendors is their own desire to prosper. EMA research has shown that 83 percent of the respondents would be unlikely to work with a firm that has had a security event. SaaS vendors know this and take every precaution to secure customer data and the applications they run on.


Is SaaS a real alternative for enterprise-sized companies? If so, what are the challenges to success?

Small to mid-size companies make up the largest portion of SaaS BI customers. Enterprise-sized companies have been slower to adopt SaaS BI. Large companies are often not early adopters and take a wait-and-see perspective. This, coupled with the need for a greater level of sophistication around feature sets, made it difficult for these firms to see SaaS as a viable alternative to on premise solutions early on. Enterprise-level successes such as SalesForce.com have shown large firms that SaaS can fit their needs and expectations, causing them to look hard at what the SaaS BI market has to offer. There is now a thriving community of firms big and small delivering SaaS BI to enterprise clients such as AON Insurance, ACNeilson, DHL, and Citrix. Both pure-play upstarts and the biggest BI vendors in the world are now offering on-demand SaaS BI products and applications.

What best practices can you offer to overcome these challenges and for successfully managing a SaaS BI project?

Companies looking into SaaS BI solutions need to consider many things, including security, licensing, TCO, feature sets, training, and service management. I highly recommend that before jumping into the deep end of the pool, a customer should scope out a proof-of-concept project that closely aligns with their overall business intelligence needs.

Like any BI project stakeholders, IT staff and the line-of-business users must take part in the process. It’s important that the customer understands both the flexibility and the restrictions of a SaaS application. Customization in a SaaS environment can be costly or impractical. Support and service are also important in a SaaS relationship. Service events should be built into the POC to assure that the needs of the clients are met.