Bloor researcher Philip Howard points out in a recent column that data-migration projects are notorious for running over time or budget - in fact, 84 percent of them do.
Most place the blame on inadequate scoping, which, according to Howard, means they didn't properly or fully profile their data.
This is a problem that could be avoided, Howard says, with a data-quality platform.
Data quality isn't something you do once - it's an ongoing analysis repeated throughout data's life cycle. Hence, Howard spends the rest of his post writing about integrated data-quality solutions.
Normally, analysts talk in large generalities or about every specific vendor. They seldom actually recommend a solution. So, I was surprised when Howard pretty much advocates outright for Trillium Software Systems' integrated data-quality platform.
Howard's rationale is that only Trillium offers a truly enterprise-wide solution. The other providers are smaller, specialized companies. He argues that while it might be tempting to go with "best of breed," in the case of data quality, there are a number of compelling reasons to opt for a less-effective, but enterprise-wide solution.
Really, this column could have been two pieces: The first a review of Trillium Software System latest release, version 11; the second a comparison between the pure-play data-quality vendors and the vendors who integrate with ETL tools. I found the second part most interesting, though, after reading it, Trillium's new release seemed more significant than it had moments earlier.
His main argument against ETL vendors is there are plenty of incidents where data-quality requirements aren't at all related to data integration or ETL, such as when you want to embed data matching in a call-center application. Why, he asks, bother with the baggage of an ETL vendor when a pure data-quality product can offer the functionality in any situation?
It's an interesting assessment of a technology that's seldom written about in the main tech press - and a technology that's apparently often neglected by organizations as well. For more background, read the related article on data migration.