By using tdwi.org website you agree to our use of cookies as described in our cookie policy. Learn More

RESEARCH & RESOURCES

Question and Answer: Business Intelligence Technology Holds Special Promise for Small and Medium Size Businesses

Up and coming technologies including in-memory processing and column–based databases, combined with advances such as inexpensive memory and 64-bit operating systems, are finally promising affordable BI to small and medium-sized businesses.

Column-based and in-memory technologies are finally making it possible to offer good, accurate, and reliable BI to small and medium-sized businesses (SMBs) with compromising on quality, says Elad Israeli, CEO and co-founder of business intelligence company SiSense, which focuses especially on SMBs. New technologies mean that it's no longer necessary to undergo lengthy data warehouse and OLAP projects simply to generate a few reports, Israeli says, greatly shortening projects times.

In fact, as he explains in this interview, it's only a matter of time until column-based and in-memory technologies completely replace OLAP, "simply because OLAP has no significant advantage where BI is concerned."

BI This Week: Your company largely targets the SMB market, and you've said that SMBs have different needs and a different psychology than enterprises in terms of business intelligence tools. What do you see as those differences?

Elad Israeli: There appears to be a BI myth that SMBs need less functionality because they are smaller. However, in many ways, smaller and medium-sized businesses have the same requirements and needs that larger enterprises do. Like enterprises, their diverse data is typically scattered across multiple locations and needs to be consolidated, cleansed, and standardized before it is accessed. Then, once the data is prepared, they, too, require the tools to enable them to extract intuitive, relevant information and share it across the business.

However, the reality for an SMB's operations versus an enterprise's is that SMB resources are dramatically limited in comparison to larger enterprises both in terms of technical staff and budgets, so a proposed BI solution must take these fundamental restrictions into account. If a solution is too complex, too expensive and requires a battery of external experts to set up and maintain, it simply is not a viable option for an SMB. Entry costs and the total cost of owning the system (TCO) become too high and the business becomes more reliant on external resources than it should be.

There's also a key difference in how business intelligence is perceived by SMBs. Many large enterprises view business intelligence as an infrastructural strategic system, such as enterprise resource planning (ERP), that is primarily focused on making business processes more informed and effective for the long term. In contrast, when a typical SMB turns to explore BI options, it is usually trying to solve specific business problems that require an immediate or short-term solution, and is not necessarily driven by the need to create a long-term framework for future business operations.

This is not to say the solution shouldn't be able to grow within the organization to cover more aspects of the business and service more users, but this needs to be done gradually, at a pace the business can handle. By their nature, SMBs cannot always plan years ahead so they must be able to implement a solution that is flexible and dynamic enough to expand as new requirements come along.

What role does new data technology play in business intelligence for SMBs?

New technology is very important, as it allows software developers to overcome challenges they could not with older technology. It also opens up the opportunity to develop products with simpler architectures and broader functionality. For instance, column-based databases and in-memory query processing allows handling of large amounts of data on less powerful hardware, by relying more on availability of computer memory than hard-drive access.

Technology is just one piece of the puzzle. Many vendors, for example, are using column-based or in-memory technology, or both, to enable analysis of large amounts of data on the desktop, targeting increased power user productivity. This wasn't possible before these technologies were introduced. However, many years of business intelligence experience has proven that having isolated data sets scattered around the organization is counterproductive.

Business intelligence has its greatest contribution by providing a single version of the truth. By encouraging a business to work with isolated "Excel files on steroids," are we making progress or going backwards? This may be a question for BI philosophers, but the fact of the matter is that an excessive number of Excel files in an organization was a major motivator in the introduction of business intelligence solutions in the first place.

Column-based and in-memory technologies have the potential to completely revolutionize the way business intelligence is implemented and used in organizations. It can significantly simplify architectures, thus shortening implementation times, lowering costs, and decreasing reliance on IT, all of which are the main entry barriers for SMBs. This must be done without forsaking well-established and proven BI concepts.

What effect will column-based databases and in-memory technology have on making BI available to SMBs?

In all honesty, most SMBs don't care whether the underlying technology is a complex in-memory database or hamsters, as long as it reliably responds to their business needs and makes a positive impact on their business operations. Technology is a means to an end. For SMBs, as with any other user, the technological focus should be squarely on matching their business needs and resources.

That said, I strongly believe that column-based and in-memory technologies finally make it possible to offer good, accurate, and reliable business intelligence to SMBs without compromising on crucial business intelligence functionality and premises. With these technologies, it's no longer necessary to undergo lengthy data warehouse and OLAP projects simply to generate a few reports. This is very important because it can mean the difference between a three-month project and a three-day project.

In addition, now that 64-bit operating systems are widely available and physical memory is relatively cheap, it's possible for companies to implement even more complex BI solutions without monstrous hardware to support them.

These are just two out of numerous examples of how these technologies help remove barriers of entry to the SMB space.

What does the future hold for BI technology?

It's only a matter of time, I believe, until these technologies completely replace OLAP technology. There is no doubt in my mind that a business investing in OLAP-based business intelligence today is actually investing in a dying technology. OLAP will eventually be completely replaced by column-based and in-memory technologies, simply because OLAP has no significant advantage where BI is concerned. Instead, there are only disadvantages, including complexity, rigidness, and high IT maintenance.

Unlike the initial BI boom that affected primarily large enterprises, I believe the next boom will take place in the mid-market. There is a steadily growing demand for business intelligence in the SMB sector, and the new technologies we have discussed make it possible to address this demand.

It's my guess that during the next decade, BI will become highly affected by standards and requirements posed by smaller businesses, not just Fortune 500 companies, and that surrounding technology will continue to evolve to service these demands. More emphasis will be placed on how to use these cutting-edge technologies to achieve speed of deployment, cost efficiency, and flexibility -- with perhaps a little less emphasis on enforcing rigid BI processes. That gives more power to users, while keeping IT in control.

How does what SiSense offers tie into what we've discussed here?

Our flagship product, PrismCubed, enables non-professionals to merge, cleanse, and centralize data without a dedicated data warehouse project or OLAP cubes, then create reports, dashboards, or even pre-defined guided analytics applications without a single line of code or SQL.

The underlying technology, called ElastiCubes, is based both on column-based storage and in-memory query processing. Working in conjunction with a visual development environment, it provides an end-to-end platform for implementing all stages of a typical BI project quickly and easily.

PrismCubed is our proof that even complex, enterprise-class BI applications can be implemented for SMBs in days, not months, without a battery of consultants and at a cost that fits their business model.

TDWI Membership

Get immediate access to training discounts, video library, research, and more.

Find the right level of Membership for you.