Integration in the Real World

5 minute read

Integration is the elephant in the room these days. As Gartner predicted way back in 2012, it’s consuming more resources than ever before because it’s more necessary than ever before. The explosion of cloud-enabled SaaS products means most companies are spending more on integration than development.

Just how many SaaS products are we talking about? CIO Magazine recently reported, “The average Global 2000 employee uses 35 different tools a day, switching back and forth more than 1,100 times.” This same article on ‘app glut’ features an astute observation from our own Tim Kulp: “Too many apps is usually the sign of a missing strategic roadmap.”

Successful Integration: It’s All About the Data

You might think of integration as connecting technology tools to streamline business processes. But a big part of integration is making sure data is accessible. It needs to flow where users need it, when they need it. Data doesn’t become information until you put it in the hands of your people. Only human experience can turn data into intelligence.

Before you even attempt integration, you have to do the hard work of understanding your business processes, inventorying your tech tools, and mapping those tools to your business functions. Developing that ‘missing strategic roadmap’ is a four-phase process you can learn all about in our white paper, Tech Platform Architecture: A Human-Centric Approach.

Once you’ve done that initial discovery work, integration allows you to begin reaping the rewards of actionable insight. Successful integration drives the leveraging of data visualization and business intelligence (BI).

Iterative Integration

A successful integration doesn’t have to be one miraculous big bang of magical alignment. More power to you if you can make all the pieces fit together seamlessly in one go-round. But it’s okay if your roadmap requires pit stops for reflection and fine-tuning. Just make sure to connect the systems most critical to your users’ vital processes first. From there, internal communication is key. Users will cut you some slack if they understand both the value of where you’re headed and how you are planning to get there.

Breaking integration up into bite-sized pieces can make it more palatable and a lot less intimidating. It’s really just a series of decisions. At base, you have to choose between two models for integrating each system:

  1. Direct Connect: A workflow automation tool like Microsoft Flow or its big brother Azure Logic Apps – or, if you’re not a Microsoft shop, Zapier or AWS Lambda – connects two systems directly. Specific actions trigger data movement. For example, a user completes a task in a SharePoint list. A manager approves. That approval triggers a data flow to Dynamics to kick off the next action.
  2. 3rd Party Data Move: When two systems cannot be directly connected, you can schedule data to move between the two as frequently as your business process requires.

Obviously, a direct connect is far preferable to a data move because there is no data disconnect; you never have to wait for a data run. But some systems just don’t play nicely together. Luckily, we can now depend on BI to bring data together across systems and after the fact, which can mitigate the effects of a suboptimal integration.

Integration Considerations

So you’re directly connecting everything that can be and relying on scheduled data moves for the outliers. How do you decide which tool you should use for each of those direct connects? There are three main considerations to help you determine whether you need a small team or enterprise-level solution:

  1. How big is the workflow? If we’re talking about Bob from Accounting’s to-do list or the purchasing department’s standard protocol, anyone who’s comfortable with Office 365 can build a Flow to connect the systems involved. But if you have a behemoth 300-step workflow, it’s time to call in the developer to build it in Logic Apps.
  2. How many units are involved? Anything that is interdepartmental is probably headed toward Logic Apps, Zapier or Lambda. An insurance company’s claims management system is a textbook example where enterprise-grade integration tools are necessary.
  3. How catastrophic are errors? We know it makes you queasy just thinking about it, but the first thing to plan for is failure: What happens when something goes wrong? How do you build for resilience? If Accounting Bob gets an email that says, “Your flow didn’t run,” he can tackle it in the morning; no sweat. But if a system failure warrants alarm bells and weekend phone calls to the entire IT department, again you have to go with one of the more robust tools.

Ready to Plot Your Course?

It’s hard to talk about application integration in general terms. Every integration project is unique because companies have unique needs and platform configurations. But taking a decision tree approach can make the process more manageable. In the real world, strategic integration happens one decision at a time.

If you need help developing your roadmap, determining your speed of business, or finding the integration that will support it, check out Tech Platform Architecture: A Human-Centric Approach. We walk you through the whole process, step by step.