Analysts say over a third of projects fail
A recent study by Bloor Research put the failure rate for data migration projects at 38% - a staggering statistic. Especially when one considers that a failed data migration project can temporarily hold up vital business processes - which of course often impact customer-facing operations. So why do so many integration projects end in failure?
That question, and a look at how companies can avoid the most common mistakes, is something I focused on for an article in our recent CBR Data Quality Management Special Report (you'll find that on this microsite, along with a number of white papers and a video podcast featuring yours truly talking to Experian's Colin Rickard.)
Rickard, Experian's data management director, started by explaining why data migrations are undertaken in the first place. "There are really two main drivers for these sorts of migrations," he said. "The first is mergers and acquisitions, in which companies often find themselves with a plethora of systems doing the same or similar things. Mergers and acquisitions is a classic one really, because you often end up with two or three separate systems, and someone in the call centre is saying, 'were you originally with company 'A' or company 'B'? The operatives have to get to the right system of course.
"And the other is what you might call new IT innovation, for example legacy migration. Companies are asking how they can reduce cost across the board, but you only really get that cost reduction if you are able to transfer entirely off the old system."
So why does Rickard believe that the data migration failure rate is so high? "Well it partly depends on what you mean by fail," he said. "Do you mean it has failed technically, or failed to deliver the benefits? The most benefit clearly comes if you can switch off the underlying system you are migrating off, so you can get better response times from the new system. But if you can't migrate 100 percent of your data, you can't switch off the old system. The danger of course is that you end up with the best of no worlds, because the old stuff is still hanging about."
So why do too many companies end up down this cul-de-sac? "Often there has just not been enough analysis done at the start, so you end up with a lot of data problems at the end," Rickard said.
So, what factors should an IT department be taking into consideration when they are about to undertake a data migration project? Visit the microsite to read the full report, and also access a comprehensive list of the reasons for data migration failures.