According to Gartner Inc. analyst John Hagerty, there are two “stories” that people tell about midmarket business...
intelligence (BI). In one, small and medium-sized businesses (SMBs) are too “simple” to need enterprise-class BI capabilities and too resource-constrained to support them. In the other, midmarket companies have the same business needs as larger organizations and thus should follow the same standards in deploying and managing BI systems.
The truth is, said fellow Gartner analyst Kurt Schlegel, that business intelligence best practices in the midmarket aren’t that different from those in large enterprises. For example, there are, or should be, cross-functional BI teams that include IT and business representatives as well as workers with analytical skills, plus lots of end-user training and a focus on change management. “There are some differences, but I’d emphasize the similarities,” Schlegel said.
However, simply getting started is the key step for SMBs. Schlegel said there is one notable difference between them and larger organizations: The latter typically have been investing in BI processes and tools for a long time, but in many cases that isn’t true of the former. “As a result,” he added, “getting the business justification for an initial investment is a current challenge for midmarket organizations, where often there is no budget and no staff for a BI program.”
According to Schlegel, the best bet for demonstrating the benefits of business intelligence as you start to move forward on a midmarket BI project is to follow the aphorism “Start small, think big.” In other words, set substantial goals for the BI program, but be sure you can deliver features and functionality to business users in useful increments.
And once a project is off the ground, there are other important decision points and BI project management approaches that can help SMBs avoid BI problems. Here is a checklist of midmarket BI best practices recommended by Gartner and other consulting firms:
Don’t forget to have a plan. Jeanne Johnson, global head of the BI consulting group at KPMG LLP, said that BI projects fare best when the implementation process is well planned and moves ahead in an orchestrated fashion. “Be very staged,” Johnson said -- echoing Schlegel’s advice, she recommended that BI teams “do something quickly to get people excited” and then move ahead in a series of achievable steps. For example, you might mock up multiple versions of BI dashboards for prototyping purposes while prioritizing one of them so it can be put into use ahead of the others. “If people learn to trust BI, they’ll put more resources into it,” Johnson said.
Figure out what you need to buy -- and what you don’t. According to James Kobielus, a senior analyst at Forrester Research Inc., one of the first things that midmarket organizations should do is determine whether they need a separate BI platform or whether the BI functionality built into their existing business applications are sufficient. For example, most ERP and CRM systems include reporting interfaces, Kobielus said. If an organization requires only basic reporting and trending capabilities for business intelligence purposes, it may be able to rely on the built-in tools and avoid the need to license, install and manage standalone BI software, he added.
Keep it simple. As a general rule, Kobielus said, SMBs should try to start simple and stay simple on their BI strategies and deployments unless they have a compelling reason to build a more complicated BI system. “The vast majority of BI is just focused on delivering basic reports,” he said, noting that midmarket organizations in particular might not need “fancy dashboards, predictive models or continuous [data] updates, because you may have just one or two data sources and only a few users.” Many BI vendors are now offering “stripped-down” versions of their product suites geared to SMBs, Kobielus said, recommending that midmarket BI teams carefully evaluate those packages if they do plan to purchase BI tools.
Empower end users, and save on IT resources, through self service. Both Gartner and Forrester recommend a self-service BI approach that enables business users to build their own views of standard reports instead of having to rely on IT or a BI team to do it for them. “We see that as a front-and-center issue, especially in midmarket companies where people need to be more enabled and self-service-oriented,” Hagerty said. With the self-service approach, users should be able to create new reports faster than if they had to wait for IT’s help, and they can personalize reports based on their individual needs. But, Hagerty cautioned, “when you look at a BI package, make sure it has features that can actually be used by the users.”
Assess your data needs and determine whether BI can “do it alone.” Kobielus said that if a midmarket organization has multiple applications handling different sets of data, a BI initiative might be a good reason to try to unify reporting and better correlate the data, conceivably in a data warehouse or integrated data marts. However, he warned that building and managing data marts or a data warehouse can be a big undertaking for SMBs, typically requiring a dedicated group within IT. Consolidating, cleansing and integrating data and developing standardized reports also require resources. If all that seems too daunting to take on internally, “consider outsourcing,” Kobielus said. The available options include hiring consultants to develop and perhaps manage a data warehousing and BI system for you; deploying your servers at a hosting facility while continuing to manage the BI environment internally; and using Software as a Service BI applications that run in the cloud and are priced on a subscription basis.
Don’t let costs get out of hand. It’s no surprise, of course, that avoiding unnecessary spending would find its way onto a list of midmarket business intelligence best practices. Keeping BI costs under control boils down to doing a good job of identifying the things that an organization really needs as part of the process of gathering requirements and building a BI business case, Kobielus said. Not overbuying on hardware is one good way to achieve the cost-containment goal, he added; adopting an outsourcing approach for at least part of a project might be another.
Alan R. Earls is a Boston-area freelance writer focused on business and technology.