Survey

Coordination, Cleansing Most Demanding Data Warehouse Tasks

Dave Andrews

Data warehouse projects can be a bit like cleaning a messy room. It's easy to justify the project, but doing the actual cleaning is the hard work. So say respondents to BYTE's survey on data warehouse projects, which is designed to gauge the level of difficulty for a variety of tasks. Response to the survey wasn't as high as previous ones. But 15 respondents who've completed at least one data warehouse project and five others currently in the pilot phase rated the tasks in terms of difficulty and took the time to write in comments about their experiences.

Their comments suggest that while defining goals and justifying a project are relatively easy ( see the charts ), data cleansing and coordination amongst various departments are the two most difficult tasks in a project. Respondents wrote that this collision of expectations (goals) and reality (availability of data) can come back to hurt you if you don't plan properly and determine what users need and what you can deliver to them.

"Data [that users] think is available frequently isn't or is in such poor condition it can't be easily used," says one respondent. Several said that warehouse planners should get a strong definition upfront of what data should be in the warehouse and what's available to ensure usable results. Knowing in advance what data is accessible is quite useful, as 60 percent of respondents rated integration with legacy data either "hard" or "very hard."

Along with technical expertise, knowing how to play the political game can help, too, says one respondent. Since coordination amongst various departments is hard, you'll want to ensure that the project sponsor wields a big enough stick to knock down barriers to success.

---

H ow Difficult Is ...

illustration_link (28 Kbytes)

Data cleansing and departmental coordination are the toughest data warehouse tasks, according to a BYTE survey.