Data Management for SMBs

Posted on
Share on Google+Share on LinkedInShare on FacebookShare on RedditTweet about this on TwitterEmail this to someone

In today’s fast-moving business world, organizations have evolved to rely so heavily on the IT infrastructure that without it, they often cannot function. One of the most prevalent challenges small and medium businesses face is centralized data management.

Data management is defined as the ability to control and maintain data in an ever-changing environment. Many businesses start out small, with a data storage and management structure to fit the needs of current budgets and staffing abilities. Often, data management problems begin with the organization purchasing and designing its own structure to meet its needs, without regard to possible growth. This lack of scalability can mean deterioration and reduced efficiency. Even if your data structure doesn’t degrade, sooner or later you will find that it has rapidly become out of date. Although “outdated” doesn’t mean “non-functioning,” incompatibility can be just as damaging. IT organizations tend to treat the problem rather than the source.

For example, consider a medium-sized wealth management company. This organization’s data structure was very sound for its needs. AS400 and data terminals were the standard for centralized data 15 years ago. As the company’s needs grew and the usage of its infrastructure increased, it added PCs with terminal emulators, which should have allowed for growth and flexibility much better than terminals. The company continued to phase out dumb terminals and replaced them with more independent workstations.

Now the company is faced with two IT departments: PC support and mainframe administration. Its programs will eventually need additional functionality, and it will have to hire outside help for development. This company now faces a crossroads of whether it should continue to throw money and resources into an aging system or migrate to a more flexible, scalable solution. Often, companies make the mistake of keeping the aging system because of the daunting cost and effort involved in a major migration. So the wealth management organization decides to develop applications that run on desktops and on true servers. This choice is common because desktop applications can come pre-canned or are easily developed in-house.

This is where the problem starts. What might have been obvious while you were reading the story is that some of these choices aren’t the best for long-term data management. Unfortunately, companies find it very difficult to see many years into the future. Now, our wealth management company is supporting an aging mainframe, the software that runs on it and an entire client/server architecture with custom-built desktop applications.

This company is also faced with unforeseen complications, such as authentication, data collaboration, compatibility, etc. To further complicate the problem, problems are often “duct taped” through purchases of more applications. The real source of the problem—a decentralized and disconnected data management infrastructure—is never treated.

What Causes Poor Data Management?

There are three primary causes of poor, degrading data management:



  • Lack of Planning: This should not be a difficult task, but as companies grow, their demand for data can overwhelm the IT structure and force IT staff to make rapid, on-the-fly implementations. This can easily be solved by appointing a responsible party for approving and overseeing the implementation of any infrastructure change. This party can be a single executive in the IT hierarchy, but always remember, you want someone who is close to the floor. Otherwise, he or she won’t be able to accurately determine realistic project timelines. If you choose to make a group of IT professionals responsible for implementation, then be sure they have the appropriate scope of understanding and can accurately determine the needs of the organization. The planning aspect of data management should be handled like a standard IT project. It should involve a timeline determined by existing research or prior experience, along with cost analysis and risk assessments. When building your critical path, you should always make room for additional slack depending on the difficulty and scope of your implementation.
  • Control Over Changes on the Network: As your IT infrastructure grows and custom needs are identified, it is very easy to have little rogue projects pop up around the network. A file server here, a database server there—this seems to be the trend in medium growing businesses. These small items may not seem like much until the application that uses them becomes identified as critical. Now, there’s a system that might not meet the standards of the organization and might even be using non-standard hardware. I’ve even seen some situations where a desktop sitting in someone’s cubicle had been converted to a server that is running a critical database. Most of the time, you find out about these rogue projects after they have failed and there is no backup in place for them. This can be avoided by implementing a change control solution. This will ensure that any needs, no matter now small, require thorough research and correct, planned implementation.
  • Proper Implementation: Implementation can sometimes be the most detrimental cause of poor data management. Although planning and change control are important, ultimately, correct implementation will determine the level of functionality. This may seem obvious, but companies often do not realize the need for skills, hardware and software. Identifying the wrong program and using poor talent most likely will spell doom for your project.


Getting Data Management on the Right Track

The three areas of proper data management are people, products and maintenance.

You must have the right people or skills to correctly complete any project—and data management is no different. Your current staff may not be able to handle a large migration or implementation. You might be running a lean staff, or the hardware and software may not be familiar to your employees. There are few options available in these situations. First, you can consider training your existing team to handle the new project. This step must be built into the critical path of the project plan. Most companies that supply data management solutions will offer some sort of training. Make your decision early in the pre-planning stages of your project so you can build the training provisions into the statement of work.

There are other ways to acquire proper skills. You can always hire contractors to perform the implementation—but this approach comes with many pitfalls to be avoided. Be sure your contractors actually know what they are doing, and choose a well-known company to help with the deployment. Usually, the vendor that sells the hardware or software has a team of engineers who can assist or complete the project with you. Larger companies might find it useful to establish a paid partnership with the vendors of mission-critical applications and systems. This will allow for cost-effective rapid response when issues arise. You also should spend the extra money and time to have the hired guns train your staff on regular maintenance and operation. The last thing you want is to get stuck with a system that you don’t know or understand.

There are several disadvantages to hiring outside help for a data project. First, your IT department isn’t actually doing the work, and as you know, building a system is one of the best ways to learn to understand it. You also will be at the mercy of the vendor when problems arise, which can be expensive and cause unnecessary downtime.

You also should carefully choose your data management solution products. Spend plenty of time researching and contacting existing customers of various products. There are many solutions to choose from. The prevalent companies that specialize in data managemen

Share on Google+Share on LinkedInShare on FacebookShare on RedditTweet about this on TwitterEmail this to someone


Posted in Archive|