I have heard that data integration can be a big problem in SaaS cloud. What type of problems can happen? What data integration features, services and assurances should we expect a SaaS provider to offer?
A big problem compared to what? Doing it in-house? Hmmm. Data integration is a tough issue, regardless of who does it or where the data lives. The only true thing I can say is that in-house and SaaS are different. Here’s what you can do.
Find partners in independent third parties that have done similar jobs before. We’ve seen the same approach since the mainframe days and it works, for a price. But also consider the kind of integration you might want or need. Is real-time a necessity or can some kind of batch processing work for you? Depending on the application a good batch interface might be all you need. For instance, data vendors supply company profile data in mostly batch mode simply because while the data changes it doesn’t change that fast. Also lead data coming from a digital marketing application to a sales application can do so in a few hours and everyone will be fine. That wouldn’t work in a hospital admissions system, but that’s not CRM either.
You should also consider appliances or interfaces in a box. Companies like Cast Iron Systems (now a part of IBM) produce blade servers that have integrations for popular SaaS applications like Salesforce already built in and they take a great deal of effort out of the process.
Regardless of the way to marry up two systems, the assurances that a vendor provides should be the ones most important to you and to the merged systems’ stated mission. Whatever it is, get it in writing. And don’t compromise at the outset. Compromise there insures you will not get what you want.