LinkedIn just like any other social media platform deals with massive amount of data. But, this huge database exactly caused it a problem, when the career website and professional network decided to attempt a redesigning the LinkedIn app for its social media platform. The main problem was to evolve a suitable solution for handling the huge data from the users of the social media platform.
The way the mobile app performs is not only to display the user data, but also getting ingested with LinkedIn dataset, which is based on the important aspect of the user behaviour. It is this ingested data that gets served to the users as well as becomes available to the internal analysts.
The changes that are made to the LinkedIn app crops up challenges before the internal team and these are:
It is not only the user experience that is impacted by the change. The reports and the analysis of data are also affected. A small change in the mobile app will have a rippling effect on the upstream and downstream applications and data.
The changes will affect the whole of the data stream.
Extensive use of data is made by the larger organizations. The same data is addressed by different names, while applications are built by the various application developers. The siloed process for different apps increases complexities.
The need of the hour is structure data that is easy to maintain. The maintainable data is made by good schemas only.
LinkedIn was about to make a choice between the two options - the old data model and the standardized data.
There were obvious benefits of the older model, as it requires no movement of the data to a newer approach. But, cost was a factor as the developers would require some replication efforts for the old code. LinkedIn estimated this as 5000 worker days, for the completion of the project.
The best option for LinkedIn was to standardize the data. Even though this required a greater amount of work, this will help to reduce the data modelling efforts, with the updated technologies. There were some constraints too like the migration to the new model by the consumers and higher up-front investments on development. An evaluation stated 3000 worker days for the project completion. Obviously, the number of worker days favoured the standardization of the data approach.
While planning for the LinkedIn app, a data ecosystem was recommended by LinkedIn that could effectively manage change. All these were in place with the help of standardized data core entities, a maintainable contract between the consumers and the data producers and ensuring a proper dialog between the producers and the consumers of the LinkedIn app. The company also created a set of tools for monitoring the contracts and also maintenance of the code base.
LinkedIn also came up with the idea of an internal visual tool, for compliance purpose that will let the employees know the facts, needed for tracking the application. A monitoring app sits on top of this tool, solving the problem when there is a mismatch between the guidance for tracking and the data emission. Another effort was a library for framework-level tracking, used by the product teams while developing the LinkedIn app.
A standardized process was also created by LinkedIn, so that the team becomes aware about the consumers of data.All these work on the backend data foundation and retooling only led to making the work on the mobile app appear simpler.
James is a writer and editor at MobileAppDaily and he is famous as a tech journalist at MobileAppDaily. He focuses on the mobile app startups & ventures and brings them to the light. He has started his career as a tech writer 6 years ago just after completing his degree in Broadcast and Digital Journalism.