In today’s rapidly evolving business environment, small and midsized businesses (SMBs) have redefined the competitive landscape and have become high-growth organizations by virtue of their acquisition strategies, geographic expansions and increased service offerings. Over the last few years, we’ve seen myriad business deals in various forms, a large number of them coming from SMBs in emerging nations like China, India and Russia. In this changing scenario, it is more important than ever that such organizations embrace appropriate technologies that will help them make strategic and tactical decisions that are aligned with their long term organizational goals.


Adoption of business intelligence (BI) technologies occupies an important place in the strategic direction of an organization. BI addresses most needs related to accurate information and data presented at the right place and at the right time. Traditionally, BI adoption had been considered a privilege of large organizations that have the time and money to invest in it. This is changing, and many SMBs have successfully implemented BI programs. However, many SMBs still wait to attain critical mass before they are willing to invest in any meaningful BI initiative. Organizations with an early start and a focused approach are more likely to succeed in the long run by laying the core foundation for a strong and scalable BI initiative. This article discusses some key strategies that SMBs could possibly take advantage of to institute a simple BI/data warehousing (DW) program for their organization.


DW/BI Products, Platforms and Architecture


What is the best set of technologies and product platforms that offers SMBs the quickest ROI at the lowest cost? This is perhaps the most pertinent question that such organizations will be faced with, especially after the wave of BI and performance management product consolidations seen last year. Organizations planning to start BI initiatives will be dazed and puzzled about how to ride this wave and align their overall BI strategy with one or more vendor roadmaps. A wide range of products, platforms and architectures are available today, each addressing specific slices in the BI space.


Packaged analytics versus best-of-breed tools: Large organizations with enterprise architecture implementations may find it easier to address this by recognizing inherent value in preserving the flavor of the overall ecosystem and thereby implementing a packaged BI solution. Additional customization to the stack is done on an as-needed basis, for example to plug in a high-end analytics application, extract data from non-enterprise resource planning (ERP) sources or address other functionality that does not score well in the existing BI stack. SMBs with ERP operational systems already in place will likewise tend to adopt a packaged analytics approach with some amount of customization. Packaged analytics add immense value to the BI program by virtue of tight integration with the underlying ERP and are best suited for operational BI, and providing near real-time information.


For organizations that have implemented custom applications to automate their business processes, a mix and match of BI tools can be considered that will serve various functions of the BI/DW solution. A product evaluation process can be used to identify products that best address these functions, but proper care must be taken to ensure that the number of products shortlisted do not clutter the ecosystem and increase maintenance overheads. In other words, invest selectively in a very small set of products that will cover maximum functionality while providing the greatest flexibility and scalability.


SMBs that are trying to establish a first-time BI initiative on a shoestring budget will not have an abundance of options for product platforms, but they will still have enough options to set up a basic architecture.Carefully follow the basic tenet of a centralized data warehouse while building a custom data warehouse, and take utmost care to avoid development of stovepipe subject areas that do not integrate well with one another. The information discovery and delivery medium can be improvised with a clever combination of existing applications in the office suite and the Web medium. Again, carefully segregate canned reports from ad hoc queries after thorough analysis of the requirements to minimize duplication of effort and redundancy. At the minimum, however, invest in a good, robust ETL tool. The alternative of hand-coding scripts for data extraction and integration makes for bad design and is generally not recommended. Once a basic framework with good design principles is set in place, it will be much easier to plug in a standard product at a future time.


Bus architecture versus hub-and-spoke model: From an architecture point of view, various options are available today - ranging from the Ralph Kimball-prescribed bus architecture (bottom-up) to Bill Inmon’s hub-and-spoke model (top-down) and various hybrid models thrown in between. There is value in each of these approaches, and various factors need to be carefully evaluated while selecting the architecture for new BI initiatives. Some organizations may adopt the Kimball model to realize quick ROI by way of departmental data marts rolled out in a phased manner. On the other hand, shops with ERP applications and packaged analytics will traditionally tend to adopt the enterprise data warehouse-style approach by virtue of the built-in architecture. For limited-budget BI initiatives where all requirements are not known upfront and organizational dynamics call for quick wins, it will be prudent to build a data warehouse that integrates the best of both worlds - the agility of the bus architecture combined with the robustness of a virtual EDW-design. It is very important to note that whatever the approach, the architecture must score high on scalability and flexibility parameters without compromising on performance and speed.


Operational BI and Strategic BI


Operational BI is widely accepted as one of the most efficient means to realize the demonstrated value of pervasive BI by getting actionable information down to every last employee in the organization. The reach of operational BI is not limited to employees in the organization but can be extended to a much larger audience to include vendors, business partners and customers. For SMBs that are currently in various stages of automating their core processes, now is the big chance to embed intelligent analytics into the business transaction processes. There is tremendous value in the proposition that front-line workers will be able to access decision-making information from their transactional workspaces without having to launch additional interfaces and tools. Combined with other upcoming concepts like geographic and demographic BI, these will go a long way to augment business process effectiveness and, more importantly, enhance the customer experience.


The benefits of operational BI are far reaching, and it is known to have provided huge gains in the form of quicker decision-making via tactical querying, increased business via cross-selling and higher productivity via reduced data latency. Product vendors are moving up the value chain too, and a few offer technology that enables operational BI, albeit with different flavors. Best-of-breed products are offering technologies that will help bring near real-time data into the warehouse; on the other hand, a few packaged BI product vendors have demonstrated via roadmaps the intent to address embedded BI with tight integration to the core business process.


Operational BI only complements strategic BI; it does not replace it. Whereas the former enables front-line workers to make timely event-driven decisions by way of right-time information the latter ensures that historical information provided to power users and analysts can help develop strategies to reduce costs and increase revenues and margins. Finally, SMBs may consider adopting a 50/50 approach toward the two forms of BI. Both must come under the purview of the BI competency center executive sponsorship, which will be discussed later in this article. The single most important factor for SMBs is to realize that the right time to implement such a technology is now - before operational systems grow to attain scale and magnitude, rendering them too complex to embed tactical BI.


BICC, Data Quality and the MDM Movement


A BICC provides for a structured and centralized governance model and is characterized by high activism from both the business community as well as IT teams. Lately, many organizations have begun to explore opportunities of instituting a BICC to increase the effectiveness of their BI initiatives and to streamline the overall delivery process. Unfortunately, many do not do so early enough in the DW/BI program lifecycle to be able to realize the full potential of it. This is the single most important factor that has resulted in the creation of federated environments and disparate BI systems in multiple organizations.


SMBs must initiate BICC setups with whatever resources are available and establish the standards and processes for a successful foundation. If there is scarcity of funds to set up enterprise platforms, R&D labs and other infrastructure, the focus must be on the other dimensions of the BICC - core processes, delivery excellence, architecture and, most importantly, governance. Carefully planned execution will ensure that the BI initiatives do not assume an IT-centric form but have active participation and ownership from the business users. Setting up a well-represented governance structure with the right executive sponsorship will ensure higher BI maturity level, even for SMBs that are just getting started. It will also eliminate the need for radical transformations at later stages, when the BI systems grow rapidly to become commensurate with the overall business.


SMBs, like any other organization, must approach the data quality and master data management (MDM) conundrums as philosophies and continual processes, rather than as slick IT solutions that can work in autopilot mode. Data quality issues trickle down to all areas of the application landscape, but unfortunately, the pain is most noticed and felt in the BI domain. Active participation of all departments and business functions is essential to establish strong data stewardship and governance processes. Starting early when the systems are still small will ensure that fractured data is rooted out of the organization and centralized repositories are built, providing one version of the truth. Again, SMBs that have not gone too far ahead setting up operational systems must embrace the data quality and MDM movement upfront to realize major benefits - the message should be loud and clear about the absolute necessity to do it once, and to do it right.


Emerging Trends: Database Management System Alternatives, End-User Experience and On-Demand BI


New trends that enhance the end-user experience, like specialty data storage devices, interactive visualization, and on-demand BI, are pushing the BI paradigm to entirely new dimensions. Several organizations have selectively embraced one or more technologies that drive these trends. Unfortunately, few of these are premium products, and they have lower penetration among SMBs enterprises. However, SMBs can consider investing in such products after having formulated a priority ranking of their BI needs. Any such evaluation should be carried out keeping the long-term organizational direction and goals in mind under the overall leadership of the BICC governance.


Fast-growing organizations clocking several million transactional records on a daily basis will quickly grow to reach multiterabyte warehouses bursting at the seams. They will outgrow almost any conceivable traditional database design, dragging down performance and speed, much to the chagrin of the end users. If such a growth trajectory is indeed foreseen, SMBs might strongly consider alternatives such as the worthy data warehouse appliance. In addition to providing huge throughput and unlimited size, such alternatives are cost-effective in the long run. There are many flavors of DBMS alternatives available today that bring together one or more niche technologies like in-memory analytics, column-store databases, dictionary-based compression and several others. It is worthwhile to make such investments in specialty devices, because substantial savings can be realized in the medium and long term.


Interactive visualization is another factor that influences BI adoption in a major way, and there are plug-and-play products available in the marketplace. The theme of these products is “No IT involvement,” but the downside is the cost involved. However, many small organizations that do not wish to invest in such premium products have successfully demonstrated highly interactive custom front-ends and rich Internet applications (RIAs), leveraging AJAX-based technologies. It must be remembered that no matter how sophisticated the BI system is at the integration layer, end users will attach a lot of importance to the information delivery and presentation medium, and rightly so.


BI as a service is another popular trend that is catching the attention of many organizations, especially the smaller ones. As cloud computing matures and more vendors start offering software as a service-based products, on-demand BI systems will evolve further, providing organizations with a spectacular array of services ranging from simple data hosting to more complex end-to-end DW development/support. On-demand BI will enable SMBs with limited budgets but high analytics requirements to take advantage of state-of-the-art BI systems without having to invest in enterprise platforms. The concept of on-demand BI can also be extended to small organizations with no in-house BI infrastructure. Organizations can sign up for on-demand BI to supplement their in-house BI investments to be able to conduct such activities as one-time data analyses, prototyping, marketing campaigns and so on.


It is important to note that there is no rule of thumb here, and nothing can be gospel in the BI world. However, there are guidelines that will help organizations devise their overall BI approach, keeping in mind the organizational direction and goals. More importantly, this article addresses strategies that take into account the gradual transformation to next-generation BI. It is essential that SMBs develop detailed medium- and long-term roadmaps and a properly devised plan around which their BI initiatives will grow. While the BI paradigm is evolving and there is a great degree of convergence bringing together operational systems, analytics and business process management, the most important factor to run a BI initiative must remain undisputed - to maintain a business value-driven focus.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access