Build v. Buy: Five Pitfalls of Coding Integration “In-House”

Build v. Buy: Five Pitfalls of Coding Integration “In-House”

It seems easy enough; You purchase a cloud/SaaS, application because you believe it will save you money, be easy to use and reduce dependence and strain on your organization’s IT department. You sign up and you’re off to the races…. Right? That’s when the elephant in the room rears its head, threatening to turn your sound investment into and endless time and cost nightmare. That 11-letter word: Integration. You thought it would be easy. Just let the IT folks write some code and “Presto!”, you’re on your way. But there are some serious pitfalls to coding integration and they can be a serious threat to your bottom line. With that in mind here are five of the major pitfalls to hand-coding integration:

  1. Application tunnel vision. It’s not enough to integrate your SaaS app to one system. An effective IT strategy requires that all of your applications be linked seamlessly, both cloud and on-premise. Can you say, “Scope creep?”
  2. Poorly-defined integration processes. Too often, IT builds integration on a “by the seat of their pants” basis. This does not take into account the organization’s vital – and often changing — business processes, nor the big picture of the company’s overall strategy.
  3. A multiplicity of rapidly-changing API’s. Not just the API’s of the cloud application (which often change four or more times per year) but those of every application that the SaaS solution is linked to. You also may have added custom fields or objects to your SaaS application, which complicates the re-coding task. Any break in the code is a break in the integration and a potential break in critical business processes.
  4. Monitoring and redundancy tools. Redundancy must be assured in the case of system breakdown, and performance monitoring is essential. Logs and alerts must be built and maintained. This is another layer added to the in-house integration task.
  5. Performance and process optimization challenges. Assuming that you’ve built the reports needed to provide excellent monitoring of system performance, what do you do when performance lags? SaaS applications provide unique integration challenges, like security considerations, limitations to the number of calls that can be made and limitations on how much data each call can contain. Factor in the changing nature of the API’s and the performance implications therein, and you’ve added yet another layer to the already complex process of trying to code integration in-house.

With the availability of easy-to-use tools (e.g. Jitterbit) that provide complete cloud integration solutions, attempting to code integration routines in-house – especially with SaaS applications involved – can be a costly proposition. Choose wisely to optimize your IT budget and resources, and above all, maximize the value of your cloud and infrastructure investments.