Top Ten Best Practices for Data Integration
Use these guidelines to help you achieve more modern, high-value and diverse uses of DI tools and techniques.
Edwin D'Cruz, a principal at Princeton Data Solutions, finished up his six-part series on managing information. Much of the series mentions integration, and I mentioned the third article, "How to Create Business Value Out of Master Data Management," in a previous blog post, as particularly pertinent.
In the last article, though, he focuses completely on integration, just as the title, "Enhance Business Capabilities with Information Integration," suggests. That title also should clue you in that he's not talking about data integration or application integration; he's talking about information integration, which can span both and so much more. Because of that, his advice may seem to be overly broad if you're used to reading pieces dealing with specific types of integration.
On the other hand, the benefit of shifting to an "information integration" mindset is that the focus becomes much more strategic, with an eye toward the whole business and its needs-not just single silos or business functions. And let's face it: Despite all the warnings about why this is a bad idea, that's where many organizations tend to get stuck, and that comes at a price, as D'Cruz points out:
The impact of a lack of this centrally available information is manifold-organizations suffer from increased costs, lack of cohesive information and poor information quality, leading to ineffective decision-making. In order to provide an appropriate integration solution, a good place to start is with the business needs. Identify the pain points and the questions being asked by the business in terms of cohesive, centrally available and reliable information. This information is at the core of implementing an enterprise solution.
That's pretty much his central theme-if you want an information system that helps the whole enterprise, you've got to think about the whole enterprise from start to finish-and that means "building solutions that answer questions across various functional areas," he writes.
I wish I could summarize this piece as a list of tips or issues, but it's more complex than that. It's more of a loose sketch for an action or strategic approach to tackling the integration problem. That said, he does at times become surprisingly tactical-for instance, he offers several pieces of advice about how to make that process more efficient.
He does offer a list of specific high-level business topics you should be able to support post-integration. It's pretty obvious-things like monthly sales, products sold, basic customer data and profit questions-but it will get you thinking through the big-picture issues.
There's also a five-step outline for what you'll need to do after you have identified your information requirements. Each of those items could be a separate series on its own, I'm sure-but again, the point is to give you an outline for attack, not a detailed plan.
One item he specifically mentions is a "determination of quality and integrity of source." Ike Ononogbu, a managing partner with InforData Consulting, recently wrote a short, but excellent, piece on data integrity and how data governance can help. Ononogbu confesses to viewing governance as "a 'police-like' action" until becoming involved in planning and scoping projects for clients. I thought that statement revealed a lot about why governance is a hard sell in many organizations.