How Many Integrations Do You Need?

When choosing a new CRM or other software for your company, you naturally want to know which of your systems it will integrate with. But sometimes choosing a platform based on available technology integrations can lead you astray.

What Is An Integration?

A software integration, also known as a data integration, systems integration, or IT integration, is the process of connecting two or more pieces of technology so that they share data at set intervals, and create a more coherent whole, (hopefully) to the benefit of the users.

Why Are Integrations Important?

Almost every company uses more than one software system. Often, the core elements of a tech stack include marketing automation, CRM, financial management & inventory control using Enterprise Resource Planning software (ERP), business intelligence tools, communication and productivity management. Any and all of these systems contain data, and some of that data is the same across multiple systems, such as some customer information.

It can be helpful for this data to be shared among some systems, so that everyone is operating from the same information and so that data doesn’t have to be entered or maintained in an excessive number of locations.

Just because “an integration” exists doesn’t mean that it will work, or do what you imagine...

In an ideal world, all of this data would be coordinated across all the platforms, and when it’s updated in one place, it would be updated in all the others. This “single source of truth” is the gold standard in the eyes of many tech professionals who see that as the only “right” way to do it.

Unfortunately, in the real world, with the number of tools being used across an organization, integrations are much more complex than it looks on the surface, and sometimes integrating software packages is not your best or most productive choice. In other cases, integrating them can actually cause more problems than it solves.

5 Reasons You Shouldn’t Integrate Everything

The allure of a fully integrated tech environment is strong. It’s easy to imagine that having every technology smoothly communicating with all of the other technologies would make your life and that of your employees easier.

The major technology companies have capitalized on this dream by promising integrations for everything you can imagine, each of them vying to become your single source of truth.

It’s to their advantage to push this agenda, because each integration you add to their platform increases your “stickiness” for them. If all of your data is retained in their platform, and you’ve invested thousands, hundreds of thousands, or millions of dollars in those integrations, you will be very reluctant to go elsewhere.

And when your user base is very sticky, you can do things like raise the prices without justifying it - as Salesforce is notorious for doing.

But the reality is that in most cases, you neither need to nor really want to integrate everything.

• Integrations are costly
  Salespeople at major software vendors love to tout the large number of integrations they offer. What most buyers don’t realize is that the integrations don’t necessarily work right out of the box and will require major, costly technical development resources to implement correctly.
• Integrations are complex
  Your different software platforms may share the same data, but they almost certainly don’t share the same data structure. What is listed as a “Company_Name” in one software may be listed as “company” in another, to state a simple example. And the more complex data structure, the more complex your integration is going to be.
• Integrations are time-consuming
  Due to the complexity of getting your data integration and migration right and avoiding problems, integrations are never simply plug-and-play. You will have to navigate the structure of your data and it will take time - sometimes months or even years. And it’s a moving target as vendors add new features and expand their data structures, or change their APIs (Application Programming Intefaces) - the way systems interchange information with each other.
• Integrations can “break” your data
  If you get the integration process wrong, your data can get messed up fast. Once your systems start “talking” to each other, they have the power to overwrite each other’s data. If you’ve messed up the connections between fields or the architecture of the integration, one system could be writing wrong information over the information in another system. It’s possible to not realize this is happening until the damage is catastrophic. This can even happen after the integration is complete, when someone changes something in the architecture of one system without updating the integration.

How to Know Which Integrations to Invest In

When you buy a piece of software, naturally you want to know which of your other systems it will integrate with. But before you ask the vendor that question, you should get clear about which integrations you actually should invest in.

Start by asking:

• Who wants the integration?
• Is it the integrator or vendor pushing for it?
• Is it the actual business need that calls for it?
• What is the business benefit? Critical or a “nice to have”?

Integrators make money from every integration, and the ongoing maintenance of it, so naturally they are incentivized to encourage you to do as many as possible. The vendors want more integrations because it makes their product more sticky.

If it’s an actual business need, then naturally you may want to enable an integration. But even then, it’s worthwhile to ask:

How critical is the integration?

Integrations can be expensive, time-consuming, and potentially risky, and sometimes there are better choices to accomplish the business need with less trouble.

For instance, in Membrain, we often help customers get access to ERP data using a simplified version of integration, where the ERP shares data in a view-only format within the Membrain workflow. This is a fast and cost effective way to give salespeople exactly the info they need, in an easy-to-digest format, without risking data contamination in either direction.

Take a look at this side-by-side comparison of Membrain vs Salesforce here  »

Which Data Needs to Integrate?

If you decide to move forward with an integration, your next step is to define the data you want to move, down to a very granular level. Which data do you need each system to have, why do you need it, who needs to see it, and how much detail do they need.

The less data your systems share, the less complex the integration will be. If you don’t understand the structure of the data and the integration needs, you can quickly get bogged down in months- or years-long projects that never end and that cause more problems than they solve.

Save Yourself Some Trouble

I can’t tell you how many times a customer has come to us saying that they absolutely have to have a certain integration. But once they’ve implemented our system and onboarded, they discover they really don’t need it and they never ask for it again.

When a salesperson tells you that their software has an integration, it doesn’t mean what most non-technical people think it means. It just means that it’s possible to make the systems talk to each other - it doesn’t mean it will be EASY to make them talk to each other, or that it’s important, or that it won’t create unexpected nightmares.

Asking yourself questions about your data and your integration needs up-front will prevent you making bad decisions based on which integrations are available.

Do you have data integration horror stories to tell? Or more tips for getting it right? I’d love to hear about them.