Business Intelligence: Last to Leave the Nest

Computerworld - Companies are outsourcing IT initiatives overseas to save costs, but is it worth it to send business intelligence activities offshore? The costs and benefits of such programs are difficult to pin down even when they're kept in-house. Data management consultant William McKnight, president of McKnight Associates Inc., says the offshoring trend will likely hit BI in the next few years. In the meantime, he recommends that companies take it slow and stick to an onshore/offshore model, in which the project managers, integrators and architects are homebodies so they can maintain some control over the projects, while only routine tasks are done offshore. Jean Consilvio talked with McKnight about how companies can assess their readiness to outsource such a core business function.

Why not send BI overseas? It's probably not a good first target for offshoring. Data warehousing and BI is very iterative and business-focused - it's not a technical exercise. That's why we got into so many failures early on, because we treated it like a technical exercise.
It's important that most of the [staff] on the project have business knowledge. And it's also a best practice to have a small team of five to seven that acts as a SWAT team and continually executes on deliverables, as opposed to -- what I think is a poor practice -- having a 25- to 50-person data warehousing team. Until you have your processes efficient, having that many people is just adding to inefficiency.
The offshore model makes sense when you have larger projects. If you consider that maybe 2005 or 2006 will be about when offshoring best practices emerge for technology-related projects, you're probably looking at another year or two before offshore BI has best practices.

How are companies deciding whether it's financially worthwhile to go offshore? How you deliver is a TCO question, it's not an ROI question. It has everything to do with the investment, but the business targets don't change, regardless of the delivery mechanism. And so offshoring, if viewed as a lower-cost approach, will increase your ROI because you'll have lower costs. But it won't do much for your returns except put them at a slightly higher risk, because the delivery may not work as expected unless you employ best practices.

What are some of those best practices? Set up tight specifications, and don't expect too much first off. I would recommend starting with a staff-augmentation-type approach and growing from there. Set up service levels, and set your expectations appropriately and ease into it. Keep an onshore