Architecture & DesignThe Hard Things about Hard Integrations

The Hard Things about Hard Integrations content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

By Chase Doelling, Product Marketing Manager, Cloud Elements

The ecosystem of data within SaaS applications is compounding. More data is generated from each micro-transaction on customers and potential customers than ever before. This charge is led by data-hungry departments and curious leadership teams; each touchpoint is tagged, flagged, and bucketed to see what numbers are impacting the organization.

Data is also expanding outwards, back into the broader ecosystem, combining a vast array of different touchpoints to paint the picture of a user and customer. This in turn informs future product, architecture, and even executive decisions. The extension of data is led at the API level, the easiest access point both internally and externally to communities.

The sophistication of data is getting deeper. The ability to orchestrate connections among hundreds of systems is often coupled, cleansed, and executed by using a variety of specific and sometimes proprietary logic. It is this ability to orchestrate data on the fly, at the API level, that gives organizations the ability to go from stick figures to Monets in painting a picture of your users. It’s also driving major challenges for developers tasked with untangling the hairball.

Enterprises today use upwards of 1,000 Cloud and SaaS applications in operations. Not every application needs to connect to every other application; this eases the shock of big numbers. But, the connections that do need to be made, need to be meaningful. And this is when an application connection needs to become a hardened integration.

Let’s look at the typical life of a SaaS customer:

  • An Ad platform serves content created in a marketing automation platform.
  • Contact information from a download is sent to a CRM.
  • Automated e-mails get sent based on information collected.
  • Encouraged, a new trial sign-up updates the marketing platform and the CRM.
  • If the product meets the needs, contacts are updated with payments and accounting records.
  • Questions through chatbots are updated with support tickets.
  • Happiness is met with renewals, upgrades, and professional services.
  • Pain is met with downgrades, more tickets, and finally an updated churn metric.

That’s just the happy path, the “normal” path. Now, start extending this through more departments with different priorities and sources of truth. Add additional automations and niche use cases for customer ABC.XYZ to add to your snowflake-like value proposition. The complexity becomes compounded and the signal gets lost in the noise.

There is a volume issue and, adding to the fun, not all APIs are built the same. In most instances, they vary widely from subtle differences like: GET/contacts instead of GET/Contact; to large discrepancies like: SOAP vs REST APIs, Web hooks versus incessant polling or access to a set of APIs are only available through a SDK. Many of the largest SaaS players have added the excitement of certifications and enterprise-level costs for full access accounts: Oh you need that in production? Let’s talk to your sales rep. Each of these variations add to the headache and development scope creep. And that is just one enterprise-level integration. Let’s look at a common integration from SalesForce to NetSuite:

  • SalesForce has its own set of REST APIs, known as APEX.
  • NetSuite has its own set of SOAP APIs, known as SuiteScript.
  • SalesForce also doesn’t do Web hooks; it’s actually sophisticated polling.
  • Each has different call limits. Don’t worry; you’ll find out when you move from test to prod.

Now, we have a ton of applications, built differently, that need to talk to each other, intelligently. No problem, because we have a top notch squad that love challenging projects. Great!

Now, let’s talk security…

With the recent Equifax hack, the urge by consumers and organizations for digital security is paramount. Yahoo had to shave off a clean $1,000,000,000 of the purchase price due to its lack of security controls. It is the number one technology challenge that needs to be solved. This is why the second reason is easier integration between tools.

Chart of API technology
Figure 1: Chart of API technology

The good news is that, once you navigate through each of these obstacles and create an authenticated integration, you just have to maintain it. For product managers, it takes people off core feature work. For developers, let’s hope they are the same ones that built it in the first place! Otherwise, API maintenance asks them to deliver in unfamiliar technologies, and quickly, because the data doesn’t stop. Oh, there also is no applause when done right (because it worked fine before) and a firestorm of git blame calls when done wrong (because it worked fine before). That’s just for parity. What about enhancements? SalesForce only has 451 endpoints right now.

If you’ve made it this far, there is some actual good news: There are more integration platforms available in the market to help manage your spaghetti connectivity and turn it into a streamlined process (maybe even pair it with wine!).

About the Author

Chase Doelling is a product marketing manager at Cloud Elements, with a decade of experience marketing, developing, and marketing to developers (since Gmail was in Beta). Today, Chase speaks and writes about APIs and their role in integrating the data of the modern age. He lives in Denver, CO, and actively avoids mountain activities to make beer instead.

# # #

This article was contributed and is the expressed opinion of the writer. Data is based on The State of API Integration Report 2017 done by Cloud Elements.

Get the Free Newsletter!

Subscribe to Developer Insider for top news, trends & analysis

Latest Posts

Related Stories