Savvy SMBs plan for data management

Last month, in the first of a series of articles exploring the choices that smaller and mid-tier companies face in the burgeoning market for data management products and services, we scoped out the range of technologies available to small and mid-sized businesses (SMBs).

“Looking to gain from the experience of larger enterprises, many SMBs are considering the application of sophisticated data management strategies to improve their own systems. Prominent among these is SOA, which offers an approach to data management rather than a specific technology.”

In this article, we will consider data management strategies – service oriented architecture (SOA) in particular – then explore how and where such deep technological approaches can be applied to the SMB market.

Starting with the end in mind

How should smaller companies approach data management?

Before selecting data management technologies and strategies, SMBs need to begin with an assessment of their current and future business requirements, according to several industry experts.

They need to clearly articulate the reasons for building or updating their data management system. What business problems will and will not be addressed by the system? What are the information needs of the enterprise? These questions must be asked of any data management project in any size company, not just a large enterprise.

Smaller and mid-sized companies tend to favour proven solutions to immediate business problems, but are just as interested as large companies in adopting any technology that will advance their business goals or provide an advantage in the marketplace. It is even much better if they can harness their existing infrastructure to achieve this. This is where SOA comes in.

Not SOA simple

Looking to gain from the experience of larger enterprises, many SMBs are considering the application of sophisticated data management strategies to improve their own systems.

Prominent among these is SOA, which offers an approach to data management rather than a specific technology.

Wikipedia, the online encyclopedia, refers to SOA as “a software architectural concept that defines the use of services to support the requirements of software users.”

Stressing the virtues of interoperability and reusability, SOA promises greater flexibility and responsiveness in data systems and the business processes they support.

For these reasons SOA might appear to be a perfect fit for smaller businesses, for which nimbleness and the ability to change direction on a dime are often key to success.

But, as Joanne Friedman, principal at Toronto-based consulting firm ConneKted Minds Inc., points out, the risks are correspondingly much higher for SMBs than they are for large enterprises. SMBs may not be able to absorb the kind of process change required as easily as larger organizations.

“For the small- and medium-sized enterprise, a multi-purpose technology that solves an immediate business problem while creating a future economic benefit is really what the technology was intended to do.” Joanne Friedman Principal, ConneKted Minds Inc.

Ironically, said Friedman, if an SMB decides to re-architect its entire infrastructure along SOA lines, it might actually lose flexibility – at least in the short term. In her experience, most SMBs realize this and resist the type of overhaul that SOA implies. “They don’t see the safety net.”

She believes that the long-term benefits are attractive, however, particularly where companies are able to use their existing infrastructure to create an advantage.

“For the small and medium sized enterprise in particular,” she emphasized, “a multi-purpose technology that solves an immediate business problem, while creating a future economic benefit, is really what the technology was intended to do.”

Friedman believes the real value of SOA will come in the 2007-09 timeframe, when SOA becomes the enterprise architecture on both the business and technology fronts.

It’s not too early to begin creating a blueprint for the future, she said, but companies need to carefully examine the implications for processes in every area of the business before they act.

“You have to define a strategy before you go down the road,” Friedman counsels. “If you don’t, it’s going to end up not giving you the value and the outcomes you sought to achieve.” There are easier ways for SMBs to get into SOA. One of these is to look for technologies built on an SOA foundation when upgrading or adopting new systems.

Many vendors have embraced SOA as an architectural paradigm and now offer SOA-friendly features in their hardware and software products. IBM, for example, recently announced the purchase of Data Power, a builder of XML-service appliances with software and hardware designed on SOA principles for an SOA environment.

This move supports the company’s expressed interest in taking a holistic approach to SOA, one that enhances business process integration, whether applied to staff, systems, customers or business partners. The idea is to reduce cost and time by easing the movement of information between applications, based on business rules.

Joanne Friedman clearly thinks this is a good place for SMBs to jump into the SOA world. “We view those types of products, and things that can be embedded in hardware, as being very significant in driving the SMB and mid-tier forward.”

Best of breed vs. integrated systems

Like larger enterprises, SMBs face choices when selecting a DM strategy or product suite. They can put together a “best-of-breed” solution — based on proven products from various vendors — or they may opt for a suite of products offered by a single vendor. The former approach may yield superior performance, but is usually more difficult to integrate and manage, whereas the integrated suite promises smoother interoperability and simplified management.

The best of breed approach also appeals to IT managers in smaller enterprises because it can be implemented in stages, with the cost spread out over a longer period. As a result, rapidly growing companies often find themselves with a hodge-podge of systems and standards. Such diversity may be acceptable at an early stage of a company’s development, but integrating these disparate threads can become increasingly onerous for IT staff.

One solution is to acquire robust middleware, such as IBM’s MQ Series, to bridge data sources and servers, but the cost can be prohibitive for smaller companies.

The same is true of other best-of-breed systems developed to handle specific aspects of data management. George Goodall, a research analyst at London, Ont. based Info-Tech Research Group, points out that their appeal may be tempered by the high acquisition and operating costs.

For these reasons, a standards-based architecture, such as SOA, appeals to many CIOs of mid-tier companies, because they’re not in a position to be standards-agnostic, according to Goodall. They usually don’t have the depth in their IT staff to support multiple standards.

Without the bench strength needed to integrate multiple best-of-breed products, SMBs tend to go for integrated offerings. These may sacrifice performance in certain tasks, but they offer a consistent interface and a single point of contact when support is needed.

Major vendors, such as Oracle and Microsoft, are now actively courting the SMB market with packaged products, supported by channel partners who provide the fine-tuning.

It makes sense for small to mid-size organizations to go with such integrated tools where possible, and to apply best of breed products sparingly when necessary.

BI Driving Change

Regardless of size, companies today know the importance of their data as an information resource. Adoption of business intelligence (BI) and data analysis tools — used to reach into their data stores to extract actionable insights — is growing rapidly.

“BI in a Web services environment and in an SOA would actually start to create a more focused operational business context around the data.”

This is partly driven by the reporting requirements of new legislation, such as Sarbanes-Oxley, a U.S. law that is affecting some Canadian companies as much as their American neighbours. But as the volume of data companies must manage continues to grow exponentially, BI is also providing the means, and the impetus, to improve data quality and derive meaningful information from the masses of data.

For example, Kitchener, Ont.-based Crawford Adjusters Canada, an independent provider of claims management tools to insurance companies, recently upgraded much of their data infrastructure with products from EMC, IBM and Microsoft.

Scott Sutherland, Crawford’s vice-president of information technology explained that the primary goals for this upgrade were to ensure high availability for their customers and adjusters, and to comply with regulations, such as Sarbanes-Oxley. But now that their infrastructure and databases are in good shape, the company’s interest has turned to the analytical tools to provide BI to their management team and field staff.

Although doing BI with Microsoft Excel reports still works for them, it is a largely manual process, and one that Sutherland concedes is archaic. Ideally they would like a BI application with a Web interface for users to select fields and generate the report themselves.

The impact of SOA is being felt here as well. Info-Tech’s Goodall points out that companies now looking to implement a top-tier BI application will probably be adopting an SOA approach, whether or not they are aware of it. The have to be tightly linked, he contends, because it is very difficult to implement a service-oriented architecture in the absence of an actual application initiative. Without it, the business case is just not as strong.

The value of this linkage is clear. As Joanne Friedman of ConneKted Minds puts it, “BI in a Web services environment and in an SOA would actually start to create a more focused operational business context around the data.”

In fact, for many companies, the BI application may become the core of a data management strategy that drives consolidation and process change throughout the data stream.

Getting the job done

As SMBs start to consider a comprehensive, standards-based data management strategy, one of the first hurdles they encounter is a skills gap. Very few mid-sized or smaller companies have a staff that is familiar with the standards, can design the data architecture and take on development of the required systems.

Outside consultants with the right kind of experience and skills may be needed, but they can be very expensive. However, many of those same consultants say there are steps that companies of all sizes can take to improve their data management systems before getting outsiders involved.

Learning as much as possible about the subject is a good starting point for companies determined to tackle their own data-management Everest.

Goodall related the experience of a resource-stretched company that wanted to improve its data systems, but knew that it was beyond its internal capabilities. The company had attempted to build a data warehouse using open source tools, then used the knowledge gained through this experience to evaluate vendors and consultants.

Putting your house in order is also an essential step. Like Crawford Adjusters Canada, companies can start by ensuring that their data is clean, and that their hardware and software are able to handle the current load, plus any anticipated increase.

Working through the implications of proposed changes is essential. As Friedman explains, “take a business problem and extrapolate across all of the functional areas of your business — how it’s going to [affect] people, process, technology and services. Out of that you will show the value of doing this architectural type.”

Throughout the process, a clear focus on the business needs and goals must be maintained.

Business-driven deployment

Such a focus was what Toronto-based Guardian Group of Funds (GGOF) sought to maintain all through its recent “data management” initiative.

GGOF is a major Canadian mutual funds company offering a wide variety of funds to investors. Their eleven investment management teams collectively manage over $750 billion.

What began as a project to improve the reporting capabilities of Guardian’s CRM tool, soon became a full-blown review of the company’s data management system, followed by a major BI implementation.

Julie Gill, Guardian’s assistant vice-president of information services recalls that the company put in a lot of thought and planning up front, but the project was not driven by a design philosophy such as SOA. They were just trying to addresses business issues, such as reducing the cost of data management. The business issues were always paramount.

In the end, Guardian chose Information Builders as consultants and integrators, and selected its Web Focus software as the BI application of choice.

We have seen that data management paradigms, such as SOA, may be adopted in their entirety, or they may infiltrate the organization in a piecemeal fashion, as with SOA-based appliances. In either case, companies that start with clear business outcomes in mind already have the firm underpinnings for a winning data management strategy, whatever its flavour.

Conclusion: Best practices

As the Guardian Funds deployment demonstrates, a clear articulation of business objectives is an excellent starting point for an effective data-management strategy. Knowing their goals, SMBs will be in a far better position to assess the challenges and opportunities ahead and the resources they will need when planning a data management initiative.

Outsourcing one or more phases of the project may be an option SMBs would want to consider very seriously – especially as they are unlikely to have in-house staff with the skills to implement a comprehensive data management system. Guardian, for example, found that their chosen vendor also offered consulting services that met their needs and fit their culture.

“As a general principle, SMBs should go with integrated tools wherever possible, and apply best-of-breed products sparingly when necessary.”

The decision of whether to opt for “integrated” or “best of breed” systems for data management is not an easy one. As a general principle, SMBs should go with integrated tools wherever possible, and apply best-of-breed products sparingly when necessary.

Like many smaller organizations, Guardian adopted this strategy. It selected an integrated product suite, mostly from a single vendor. And, reflecting a growing trend, that suite centred on the need to provide relevant business intelligence to a widening pool of users at all levels of the organization.

Next month, we will be taking a closer look at the elements of a good data management solution, including technologies, people and planning. Are there trade-offs to be made? What provisions for the future — such as scalability — need to be included?

Don’t miss related articles, white papers, product reviews, useful links and more! Visit our Spotlight on Data Management