Why deep linking isn’t enough for mobile app integration

This article initially appeared on VentureBeat on February 1, 2015.  Please click here for the full article as initially published.

Image Credit: Stock-Asso / Shutterstock

Imagine your big night out on the town being completely planned and paid for with no more than five taps on your mobile phone.

The seconds and minutes saved may seem trivial, but aggregated across thousands of users and instances, “deep linking” has the potential to save an immense amount of time. The investment in deep-linking player Button last week shows the economic value of more intuitively streamlining how consumer mobile apps work together.  Suggesting and even executing actions on behalf of users saves time, increases app engagement, and is a step in the right direction.

This approach of pre-built app integration is not unlike the approach enterprises have taken to make their business apps work together seamlessly. Triggers in one app drive actions in another, and data from one is mapped and transferred to another in a predetermined manner.

However, these pre-built integrations have fallen short of solving the problem for business customers. For example, Intuit introduced an integration product to connect Salesforce and QuickBooks. Three years later, Intuit’s explanation for shutting it down crystallized the shortcoming of pre-built integrations:

“Our original intention for the Salesforce for QuickBooks and Salesforce for QuickBooks Integration business was to serve the needs of our Desktop (Pro/Premier) base with an out of the box solution…Instead, we have seen a high volume of use by customers who have needed a higher level of customization than our solution can provide – and our solution was unable to meet their needs.”

This had nothing to do with Intuit’s ability to program, code or collaborate with Salesforce. Instead, the company could not support the massive number of variations different customers had for how they needed Salesforce and QuickBooks to work with each other. The challenge in connecting consumer mobile apps is not that different. The model Button is using, while an exciting step in the right direction, doesn’t fully address the issue at-hand: People use combinations of apps in unique, infinite ways. It is impossible to predict their usage in terms of features, sequencing, and preferences.

Currently, deep linking requires serious engineering on a custom, one-off basis. App vendors must create mini-products for each app-to-app linkage. They must be prepared to handle the unique variations for how customers want to make their apps work with other apps. This engineering-intensive model inherently limits how much deep linking can occur.

Out of necessity, cloud business apps have evolved to make integration more adaptable, providing clues for how mobile app integrations might do the same. First, APIs have made it easier and faster to connect cloud-based apps, but APIs for mobile are not as prevalent or functional. Creating a more systematic, secure API approach would rapidly accelerate flexible deep linking. From here, app vendors must empower their users to link their apps as they want. Intuit showed us what can happen in the enterprise space when trying to meet the idiosyncratic needs of many with a single, standardized solution.

In the case of Button, smarter integration beyond simple automation is a step in the right direction. But such technologies must enable the myriad variations for how people want their apps to work together. Moreover, to make this work at scale, app users should be empowered to link apps themselves. Otherwise the promise of interoperable mobile apps will remain an elusive one for most users.

Leave a Reply