The Dangers of Delaying Cloud Integration

I often hear from people implementing cloud-based applications and platforms that they’re going to put off integration until phase two. Where phase one is driven by the line of business (LOB), it often goes something like this:

“We don’t have technical resources, so we’ll figure out integration with on-premise systems in phase two of the implementation. Initially we’re going to focus on getting started with Accounts, Contacts, and Opportunities as well as the challenges of adopting a new sales process.”

Of course, if you’ve been involved in a situation like this you quickly discover how important data integration actually is to cloud application success. As one sales operations manager I spoke with put it:

“We learned early on that instead of building SaaS marts, we had to be SaaS smart!”

Okay, he also had a business intelligence background and understood some of the “one version of the truth” issues of data mart proliferation and data warehousing. At Business Objects years ago, we used to refer to the Gremlin analogy when it came to unsupported, departmental BI tools. The analogy seems to apply here as well – the SaaS app has the appeal of a cute, furry Gremlin initially. But without proper care, feeding, and planning the nightmare begins….

Inevitably in phase one of a cloud application implementation (especially in the case of CRM), you’re going to get questions about integrating with a quoting engine, price books, orders, and all kinds of on-premise databases and systems.  And if you’ve somehow managed to succeed in putting of integration until phase two,  adoption and productivity issues lie ahead as your users are faced with the daunting task of inputting to and accessing data from multiple systems. And I haven’t even mentioned data quality challenges…more Gremlins!

I’m curious about this “phase two cloud integration” phenomenon. What are the drivers for putting of cloud integration?  Here are a few possible reasons for this mindset:

1. SaaS application sales reps want to check the integration box and avoid slowing the deal down with potentially complex integration discussions (see my Wilfred Brimley post)
2. Prospects are told that data migration, synchronization and replication are easy thanks to a few free tools, Excel spreadsheets, and a Web Services API
3. LOB either doesn’t have technical resources or simply don’t want to have to get IT involved as they’ll slow things down (see my post on Cloud Politics)

The good news is that data integration delivered as an online service is helping to demystify and simplify the topic of application and data integration. LOB managers are discovering that there are self-service solutions that are available that incorporate all of the principles of software as a service (ease of use, rapid deployment, multitenancy, subscription pricing, etc.). At the same time, IT organizations are realizing that they can retain governance and control of LOB self-service (often a scary concept) and extend the use of the powerful data integration tools and technologies they’re already using.