It's More Than Code: The Hidden Complexities of Building Integrations
July 15, 2024

Michael Zuercher
Prismatic

On average, companies use 130 SaaS applications (BetterCloud 2023 State of SaaSOps Report) to manage operations, leading to fractured workflows. Eight in 10 (SalesForce 2023 Connectivity Benchmark Report ) say integration challenges are slowing their digital transformation. B2B SaaS companies add enormous customer value to their products with built-in integrations to their customers' other tools. These connections reduce friction and streamline tasks.

Could your SaaS company use your existing developers to build integrations entirely in-house? Absolutely. But should you? Possibly not.

On the surface, connecting systems might appear to revolve around writing code. But so much more goes into delivering integrations, and these additional facets are what truly drain developers' time. Let's explore these "other" details required to build a B2B SaaS integration environment.

Infrastructure

In addition to building the actual integrations, your team must figure out infrastructure and deployment strategies. First, you'll need to provision compute resources that can run your integration code. This approach differs based on the system you're using.

Once you outline infrastructure, you must figure out how to deploy and update code, which also requires considering numerous options and scenarios. Swapping out containers or Lambda definitions works for container or serverless deployments. Servers, VMs and EC2s will require more advanced blue/green deployments. Teams must also answer questions like:

■ What happens if your integration is executing when a container is replaced?

■ Can you easily roll back to a previous version when issues happen?

■ Can you quickly deploy different versions to different customers?

Scaling infrastructure is yet another challenge to consider. Data moves between applications in bursts, and your system must be able to handle the ebb and flow of these transfers. At the same time, misconfigured third-party apps can overload your servers, which requires creating additional configurations to protect you from scaling resources unnecessarily.

Integration UI for customers

For many B2B SaaS products, the ultimate goal is for customers to activate integrations themselves. Sure, building a basic listview screen is straightforward for a front-end developer. But what happens when your team adds more integrations? This brings about questions like:

■ Will the front-end developers need to be involved with every change?

■ Who maintains the list of integrations?

■ Can you display different lists of integrations based on pricing tiers?

Next comes the integration configuration experience, which is more complex than a customer providing you with their third-party app credentials. You must:

■ Decide between creating custom code for each configuration or building a generalized solution.

■ Track all these configurations for each customer, likely requiring you to build new APIs to ensure the integration code has access to the proper configuration.

Depending on your process, your backend team may need to be involved with each integration.

OAuth 2.0

Many applications use the OAuth 2.0 authorization mechanism. It's simple and easy for users but requires significant backend work, including:

■ Setting up a predetermined callback URL to handle the return from the third-party app's consent screen.

■ Implementing a mechanism to track user and integration information using a state property within the callback URL.

■ Establishing infrastructure for generating OAuth authorize URLs and receiving authorization codes from users.

■ Ensuring that the OAuth callback URL can exchange authorization codes for access tokens with the third-party apps.

■ Developing an OAuth service that periodically refreshes tokens for all integrations and alerts you if a token refresh fails.

Webhook infrastructure and event scheduler

Event-driven integrations are ideal for moving data between apps. You can use a webhook payload to notify your app that a third-party app event has happened. To support this function, you need to build a service that:

■ Determines which customer and integration a webhook payload is for.

■ Dispatches the request to the appropriate compute resource.

■ Queues up requests if resources are not appropriately scaled.

You also need to define retry logic for failed requests. If an application does not support webhooks, you must build a service that invokes integrations on a schedule.

Data logging

All of the previously-mentioned services will create logs, which need to live somewhere. The infrastructure you build must capture the logs from your runners and supporting services, which usually involves installing agents from your log provider on your infrastructure and keeping those agents up to date.

Monitoring

It's an unfortunate reality that all software will fail at some point. When something goes wrong with an integration, you want to detect the problem before your customers do. Your log provider might include monitoring features, but they'll require fine-tuning to detect failure accurately. Depending on your needs, you may require an additional monitoring service to send alerts to your team when an integration isn't behaving as expected.

Testing

Your integrations will require a way to quickly deploy new versions to a production-like testing environment. Integrations are iterative, so you need effective strategies to deploy and test new code efficiently without tying changes to core product updates. Adding to the challenge, you can't always test against a vendor's test endpoint, requiring you to create a mock endpoint.

Security

Managing security between applications is difficult. Properly protecting data requires building infrastructure to encrypt third-party credentials in transit and at rest — and designing data protocols so integrations can only access relevant credentials and configurations. You'll need to consider:

■ How you monitor the infrastructure.

■ What normal vs. abnormal function looks like.

■ How to detect and respond to incidents.

Certain customers might require your integrations to comply with strict industry-specific regulations, adding additional complexity.

The cost of in-house integration building

B2B SaaS development teams are entirely capable of accomplishing all of these integration tasks, but at what cost? Skilled engineers spend months working on these behind-the-scenes integration needs rather than focusing their energy on enhancing the core product.

There's no doubt about it — integrations are business-critical and harder than they look. Rather than draining engineers' time and creativity, consider how your company can reduce integration burdens with technology — like embedded integration platforms — that are capable of doing some of the heavy lifting. Simplifying integrations benefits your engineers, your product and your customers.

Michael Zuercher is CEO and Co-Founder of Prismatic
Share this

Industry News

April 17, 2025

GitLab announced the general availability of GitLab Duo with Amazon Q.

April 17, 2025

Perforce Software and Liquibase announced a strategic partnership to enhance secure and compliant database change management for DevOps teams.

April 17, 2025

Spacelift announced the launch of Saturnhead AI — an enterprise-grade AI assistant that slashes DevOps troubleshooting time by transforming complex infrastructure logs into clear, actionable explanations.

April 16, 2025

CodeSecure and FOSSA announced a strategic partnership and native product integration that enables organizations to eliminate security blindspots associated with both third party and open source code.

April 16, 2025

Bauplan, a Python-first serverless data platform that transforms complex infrastructure processes into a few lines of code over data lakes, announced its launch with $7.5 million in seed funding.

April 15, 2025

Perforce Software announced the launch of the Kafka Service Bundle, a new offering that provides enterprises with managed open source Apache Kafka at a fraction of the cost of traditional managed providers.

April 14, 2025

LambdaTest announced the launch of the HyperExecute MCP Server, an enhancement to its AI-native test orchestration platform, HyperExecute.

April 14, 2025

Cloudflare announced Workers VPC and Workers VPC Private Link, new solutions that enable developers to build secure, global cross-cloud applications on Cloudflare Workers.

April 14, 2025

Nutrient announced a significant expansion of its cloud-based services, as well as a series of updates to its SDK products, aimed at enhancing the developer experience by allowing developers to build, scale, and innovate with less friction.

April 10, 2025

Check Point® Software Technologies Ltd.(link is external) announced that its Infinity Platform has been named the top-ranked AI-powered cyber security platform in the 2025 Miercom Assessment.

April 10, 2025

Orca Security announced the Orca Bitbucket App, a cloud-native seamless integration for scanning Bitbucket Repositories.

April 10, 2025

The Live API for Gemini models is now in Preview, enabling developers to start building and testing more robust, scalable applications with significantly higher rate limits.

April 09, 2025

Backslash Security(link is external) announced significant adoption of the Backslash App Graph, the industry’s first dynamic digital twin for application code.

April 09, 2025

SmartBear launched API Hub for Test, a new capability within the company’s API Hub, powered by Swagger.

April 09, 2025

Akamai Technologies introduced App & API Protector Hybrid.