Once an organization decides to purchase a data warehouse appliance and identifies the specific product it plans to implement, the project management phase
Remembering that IT projects in general have high levels of failure, companies deploying data warehouse appliances should be sure that they develop a strong project plan. Another priority is becoming aware of the common pitfalls and problems that other companies have encountered on appliance deployments, so you can learn from their experiences. By planning smartly and identifying potential stumbling blocks before they occur, companies will be less likely to run into difficulties.
When looking at the hurdles that organizations can face on implementations of appliance-based data warehouse systems, both business and technical challenges need to be taken into account and addressed beforehand to help ensure a successful project:
Business challenges. Sometimes organizations overlook the business factors when deploying appliances due to the detailed technical demands of a data warehouse project. Aside from identifying the business problems or requirements that precipitate the need for a data warehouse appliance, and looking at the business rules associated with the data that will be warehoused, proponents should attempt to calculate the projected return on investment and total cost of ownership associated with the new or expanded infrastructure. The ROI and TCO figures can help justify the project and prove its overall value.
Unfortunately, some organizations are unable to demonstrate the potential value of data warehouse systems beyond the data cleansing and consolidation benefits they provide. The problem is that even if the initial appliance deployment is approved, future expansions likely won't be a priority without a direct attachment to tangible business value, such as cost savings or increased revenue and profits.
Technical challenges. At a high level, some of the common technical challenges that can be encountered during a data warehouse appliance implementation are based on project dependencies. Data warehouse teams should be aware of other IT projects that are taking place and how they affect the data in source systems – for example, whether any changes will result within the source data or to any related business rules. Those factors can affect project timelines and the ability to deploy an appliance with the proper data quality and integrity assurances.
Integration considerations also need to be identified – both from a data perspective and from a systems view in relation to the existing IT infrastructure within an organization. Setting internal hardware and software standards can lessen the integration challenges because processes will already be in place to guide the required work. Of course, databases by their very nature present basic technical challenges for project teams – and the ones built into data warehouse appliances are no exception.
Data preparation, initial data loads and table joins are all areas that are ripe for failure when implementing appliances, as is the process of developing data quality rules. Without adequate controls that help preserve the validity of data over time, even reiterative data-cleansing efforts won't resolve quality issues. And the adage “garbage in, garbage out” applies very strongly to data warehouse appliance projects: Business intelligence and advanced analytics require high levels of data quality to provide valid findings. Companies that focus on ensuring data integrity, and that develop business processes to support it, are more likely to succeed in their appliance deployments.
Whether potential pitfalls are business- or technical-oriented, it's important to assess how each could affect the implementation process and the resulting data warehouse systems environment. Developing a detailed deployment plan and checklist, and then following it to make sure that all of the required tasks are completed, can help an organization avoid problems and mistakes on a data warehouse appliance project. In addition, there should be adequate contingency planning to enable the data warehousing team to deal with any discrepancies between the project plan and the actual deployment.
Appliance vendors tout the bundles of data warehouse hardware and software as easier to deploy than traditional data warehouses are. As with any IT project, there are still roadblocks to watch out for – but planning and attention to detail can enable companies to get around them without major difficulties.
Lyndsay Wise is president and founder of WiseAnalytics, an independent analyst firm and consultancy based in Toronto that focuses on business intelligence and dashboards for small and midsized organizations.