Testing the BI Waters with Phased Deployment
Taking an incremental approach to BI deployment can help those hesitant enterprises jump into BI head first.
By Michael Brockway, Consulting Director, Dimensional Insight
What's keeping so many companies from deploying business intelligence? Budgetary reasons are, of course, a main concern, but many companies also fear a long, drawn-out implementation that consumes time and resources and leads to too many headaches and infighting. Unfortunately, this fear often prevents the companies who need BI the most from adopting solutions that would add immediate value and insight to their businesses.
How can companies conquer this fear and realize the benefits of BI? In this article, I'll take a look at phased deployment, a strategy that favors an incremental approach to rolling out BI solutions, typically starting with a single department or application within a company. Although phased deployment is certainly not the answer for every company, it can help those hesitant to jump into BI head first.
The Pros and Cons of Phased Deployment
The phased deployment approach makes sense on several levels. First, it allows companies to dip their toes in to a BI vendor and its platform on a small scale before committing to a more involved relationship. For example, customers can figure out first hand the quality of the consulting engagement, how a vendor handles support requests, how suitable the platform is for solving the key strategic and operational problems confronting the business, and how easy the application is to scale when adding additional users, departments, and business units.
However, this strategy is not for everyone. For companies that typically consider only a few capital-intensive IT budget requests annually, the smaller incremental requests to fund phased deployments may disrupt the established budgetary process.
Second, departmental infighting may act as a roadblock to a wider rollout regardless of demonstrated initial success. Finally, for customers with an urgent need for an enterprise-wide solution, an incremental approach will not prove valuable. They are acutely aware of their needs and they needed a BI solution "yesterday."
Despite these potential drawbacks, phased deployment can be an attractive solution to CFOs who balk at committing to expensive enterprisewide BI rollouts that often span months or years before demonstrating value or delivering significant ROI.
Keys to a Positive Phased Deployment
Not all vendors will support phased deployment for customers. For it to be viewed as a realistic implementation strategy, rapid BI application development, validation, training, and roll-out must be integral to the overall solution. Simply having feature-rich software is not enough, especially if the vendor's development environment proves cumbersome to work with or forces users to experience a steep learning curve before they can productively utilize the toolset. What should you look for?
A fully integrated, end-to-end solution. Many BI providers' application portfolios consist of acquired products that are rarely unified by a consistent end-user interface. This can slow down deployment because it's harder for customers to learn the technology. Solutions with a consistent and familiar look and feel across the system will help accelerate deployment.
Ease of use. Who wants to spend a long time learning a new software application? Any takers? I didn't think so. If you're rolling out software in a phased approach, it must be easy to use. For the best results, seek out approaches to data query and analysis in which no SQL code or other cumbersome scripting language is required to build sophisticated reports and graphical data displays. Simple point-and-click functionality ensures that even novice users can perform analysis and generate reports. Also, finding solutions with customization options allows users to develop visually impactful reports, all without coding.
Free yourself from the database. Databases can make life much more complicated -- and expensive -- which detracts from the benefits of phased deployment. Solutions are available today that can store data models on disks and do not require an underlying database to store or manipulate data. This frees customers from having to purchase additional database licenses or deal with database scaling issues each time new departments are added.
Focus on self-sufficiency. Solutions that enable you to be self-sufficient will ultimately work out much better in a phased deployment than will solutions that force you to depend on others. For example, self-service reporting ensures that IT help desk requests for report creation will decrease. Database independence frees IT personnel from maintenance overhead. Also consider your vendor's approach to training. Does it have a "train the trainer" approach so your training resources are located in-house, or must you heavily depend on the vendor's consulting and technical support teams? The more you can do on your own, the easier rollout will be.
Even if a phased deployment strategy delivers initial success, budgetary and capacity constraints often limit further initiatives. You can mitigate these constraints in several ways, including:
- Look for software lease options with a flexible timeframe to cover additional user licenses until the next budget cycle (or longer)
- Determine if there are options that provide functionality at a low monthly subscription rate
- For on-premise installations, ask your vendor if it has a scalability road map that enables you to seamlessly expand existing server capacity to accommodate anywhere from dozens to hundreds or even thousands of additional users
Michael Brockway is a consulting director at Dimensional Insight's Florida office, where he is responsible for all consulting activities in Florida and other locations. Michael is a senior IT executive with over 34 years of experience in the areas of business intelligence, supply chain, healthcare, and government. He has direct leadership experience in all areas of IT, including strategy, business relationship management, large system implementation, educational services, electronic commerce, and IT governance. You can contact the author at firstname.lastname@example.org.