Optimizing Looker with Kargo

Our impact

• Revamped Kargo’s Looker instance to enable BI at scale and upskilled Kargo to future-proof its use of Looker
• Achieved a 22% reduction in computationally heavy merged queries, reducing the load on Kargo’s database, and increasing efficiency
• Reduced query runtimes with 71% of all queries returning in under 10s (up from 58%), reducing time spent waiting and improving the productivity of its users

About Kargo

Kargo collaborates with advertisers to develop custom ad campaigns that align with their brand objectives and leverage the unique capabilities of mobile devices to engage audiences. Its platform combines data-driven targeting, creative expertise, and advanced technology to deliver impactful advertising experiences.

Kargo also offers services such as creative production, ad verification, and data analytics. These aim to enhance the effectiveness and measurement of mobile advertising campaigns. 

The challenge

Kargo has grown its BI capabilities rapidly and like many rapidly scaling functions, it wanted to assess its Looker instance with best practices in mind. It did not have a unified client-facing data platform across its various suites and agencies. This meant it needed to rely upon its internal users to provide clients with ad hoc reporting on a by-request basis.

Furthermore, its internal developers wanted to increase their experience and skillsets to work efficiently with embeds and APIs. A priority feature that Kargo wanted to provide to its users was an external-facing data platform that empowers clients to visualize their campaign performance across multiple Kargo business lines in one place and create custom data reports.

Finally, Kargo wanted to boost its efficiency by ensuring its Looker instance is always performant and reducing LookML validation times.

With these challenges in mind, Kargo wanted to develop a self-serve UI, leveraging Looker’s embed functionality, while ensuring strict user roles and permissions are implemented and enforced. This included enabling its developers as quickly as possible to increase capabilities and efficiency, and future-proof its Looker instance.

Our solution

Initially, Datatonic conducted a Looker Health Check to assess Kargo’s Looker instance. After identifying the areas of improvement, Datatonic worked with Kargo to:

• Implement recommendations from the Health Check, leading to a more performant and easy-to-navigate Looker instance
• Reduce the runtime of the top 5 explores with the longest-running queries by 30%
• Implement data groups, leading to increased results returning from the cache by 5% reducing the load on the database
• Achieve a 22% reduction in computationally heavy merged queries from 57% to 35%
• Reduce average query runtimes with 71% of all queries returning in under 10s (up from 58%)
• Develop an Embedded Analytics MVP (Kargo Application)
• Create sample workflows to provide Kargo developers with confidence on all the considerations to be taken into account with the embed implementation
• Prioritize Security & Governance (i.e., roles & permissions guidance in Looker) 
• Provide guidance and best practices on how to implement comprehensive roles and permissions

This work has enabled Kargo to implement BI best practices in its Looker instance and increase the speed at which its teams can use Looker to unlock key business insights, improving decision-making across the business.