Considering an Embedded Integration Platform
If you're looking to buy an embedded integration platform (also known as embedded iPaaS, a term we're not a fan of), there are several things to consider. Buying an embedded integration platform is a High-Consideration decision: not only will the platform be customer-facing, but it will also deal with your customer data. Once you go with a platform, you will be with them for a long time. So, it makes sense to take a moment to think through what a promising integration platform is and how you decide to choose one.
While numerous evaluation guides are available, it's important to note that many of them tend to blur the lines between B2C and embedded integration platforms. These two are distinct entities, and most software review platforms lack the depth or expertise to provide an accurate assessment, making the evaluation process challenging.
With our extensive experience in this field and insights gathered from numerous companies, we can give you a unique perspective on what customers consider important versus what becomes crucial in the long run. We'll also equip you with an evaluation form to guide you in making an informed decision.
Does the platform have the app I'm looking for? What are my options if it doesn't?
If you're looking for a platform, you should have a customer or many customers looking to integrate with a specific app. These are usually the most common ones within a vertical (like Salesforce or HubSpot in CRM). If the platform doesn't have the app, it can be a non-starter. If you're talking with a unified API provider and they don't have the app, this should be an immediate disqualifier since there's nothing you can do other than wait and hope for the unified API to add the app. For other embedded workflow-based platforms like Integry, you can add your own app.
However, even if the platform "has the app," does the app do what you want? What if your use case is uncommon? What then? Does the platform allow for extensibility? How long does it take to add a new endpoint? What kind of expertise do I need? For a Unified API, you'll have to bypass their data model and directly hit the API (aka API Passthrough), which defeats the purpose of using a unified API. At that point, you're better off coding directly with the API in question.
Finally, some platforms give you their own connectors, which saves you a lot of time. On the other hand, other platforms require that you get your own credentials, which can be a very time-consuming and resource-intensive effort. Whitelabeling a connector is a great option, but having default connectors is a big time saver.
How do I build my integration?
An integration platform might do very little work for you. For example, it might only handle authentication and expect you to do everything else, like making API calls, writing the integration, and keeping up with changes.
A Unified API platform will give you access to common APIs, but you'll have to write the code for every integration. You will be responsible for maintaining that code, running those integrations, scaling them, and building monitoring.
An embedded workflow platform allows you to build the entire integration on the platform. This means managing changes, running the integration, scaling, and monitoring comes out of the box. This significantly reduces management and engineering overhead.
How will I monitor integrations?
As your integrations go live, customers will report issues. Many of those issues will be related to data. For example, you're pushing contacts to Salesforce, but the user only sees some of them at Salesforce. What's going on?
To debug this, you need to look at an integration's entire life cycle. When was it set up? What permissions were used? How many times did it run? Did it send all the data? Was there any data rejection or normalization?
If you're using a Unified API, you can't get this complete picture without writing significant logging in your code since you are responsible for running the entire integration. You must piece together API calls made to the app and the code executed on your side and see if there's a logical coding error, an API error, or both. Debugging with Unified APIs thus becomes really complex, and it requires the high-level expertise of a product engineer.
Contrast this with an embedded workflow like Integry. Because the entire integration lives inside Integry, you can see exactly when it was created, how many times it ran, what steps took place each time, which API call failed, or where the integration didn't behave as expected. Because this is built into the platform, you don't need to pull in a product engineer, and anyone from technical support can trace the flow and let the customer know what possibly went wrong or escalate to engineering, thus keeping a fast turn round time for support issues as well as reducing product engineering time consumed.
Embedding Options
Sometimes, you already have your own integrations and want to show a common list of integrations you built along with the platform's integrations. Sometimes, you want to show integrations in a certain context where the integration already has information (like which list you're setting up the integration on). This requires that the embedded UI be flexible and powerful.
Most Unified API platforms have fixed functionality where they authenticate users and ask for limited field information. You will have to build a UI for everything else. For example, what if you need the user to select a list on a CRM to sync to? If you're using unified API, tough luck; you'll have to code this yourself. The more platforms you integrate with, the more code you need to write and maintain.
With a platform like Integry, you get a UI editor built right in. This allows you to ask questions, surface data, and perform mapping. This eliminates the need to write any front-end or UI code and enables you to write.
How long does it take to go live?
Building an integration requires many steps, especially your first integration. There's front-end work, the integration logic, getting API credentials for your developer app, building receiving/webhooks, and testing.
The average time to take the first integration live is 2.2 months (G2 Report). On the other hand, on Integry, it takes a week for your first integration. Depending upon your use cases, the time to go live can be longer, but if the platform reduces your work and helps you be productive, it's a win.
Security and Data Protection
An integration platform must deal with your customer's data, making security and data handling important. Does the platform store your customer's data? Does it give you control over what is stored? Does it give you a clear retention policy?
Most Unified API platforms replicate the entire customer data on their own servers. This allows them to give you an API to query the data quickly and repeatedly. However, this is very heavy-handed. Why should your customer's entire data be stored on another platform? This is an architectural decision by the platform, and we believe this is a significant liability and, in many cases, a non-starter.
Platforms like Integry do not store any customer data. In fact, we've adopted a strict data-is-toxic policy. In essence, we believe data is a toxic asset, which means we want to keep it for the shortest time, reduce touches, and reduce any exposure. In the world of AI, where every company wants to use data for training, this is critical clarity to have. Integry does retain logging data for 30 days for your debugging benefit before permanently deleting it.
Lastly, you will want single tenancy or on-premise for those in regulated spaces or with even higher data protection requirements. These options typically incur a higher price or higher maintenance cost.
How is the platform priced?
How a platform is priced is an important consideration. You want pricing to be aligned with the value you're getting.
There are different pricing models out there, from charging by API Call, by users, by integration runs, by tasks, by several integrations, by data moved, by a number of machine licenses, etc.
Ultimately, a good metric is one that you can accurately predict, does not penalize you for growth, and is aligned with the value your customers are getting.
We price Integry by either Run, which is an invocation of an integration regardless of the amount of data it processes, starting at $0.009/run and going lower with volume. We also offer per-user pricing starting at $10/month, where the user can set up any number of integrations to any app and make any number of calls. Our price is lower because our architecture costs us less to run. For example, we do not replicate your customer's entire data whether you use it or not. This allows us to be dramatically lower in price and significantly more secure.
Summary Criteria
Here's a table for the final evaluation criteria that readers can use to evaluate each embedded integration platform.
You can make a copy of the Google Workbook by clicking here
