GLS Gains Faster Insights for Improved Delivery Forecasting with Snowflake’s AI Data Cloud

Delivery company GLS understands the importance of parcels. It knows that they connect people to each other. Which is why the company doesn’t just see itself as delivering packages — through its own infrastructure and strong partner networks it’s delivering the hopes, dreams and ambitions of businesses and consumers across 220 countries and territories worldwide. 

One of the most important aspects of any delivery service is making sure things arrive on time. To this end, GLS uses analytics to gain insight into delivery status and share that insight with its customers.

“It’s very important that people can rely on the delivery forecasts we provide,” says Dirk Herzog, Senior Manager of IT Product Management Business Analytics at GLS. “Behind every parcel there’s a story, and a human being who needs what’s in the parcel.” 

GLS has a long tradition of being data driven, but Herzog saw an opportunity to improve how the company was working with data and to be more future facing with its insights.

“We have a huge amount of data,” he says. “We were always able to look back and see where delays had taken place. But we were not able to look into the future and predict what would happen. That was our next step.”

To remedy this, GLS replaced its existing database management system with Snowflake’s AI Data Cloud on Amazon Web Services (AWS).

GLS’s existing, on-premises system had its drawbacks. Primarily, high maintenance requirements meant time that could be spent working on new features was instead dedicated to keeping systems up and running.

“We also observed extended periods of reduced processing,” says Herzog. “And that could be caused by something as simple as a big query.”

As GLS’s data platform reached the end of its service life, it presented the perfect opportunity for the company to do something new — which meant some big decisions had to be made about the next steps on the company’s data journey.

“We didn’t just want to take our old setup to the cloud,” Herzog says. “We wanted to take a completely different approach to analytics.” 

GLS was attracted to Snowflake for several reasons. As a cloud-only database, Snowflake requires near zero maintenance, it offers scalable performance so big queries would no longer disrupt operations, and it would provide a cheaper alternative to GLS’s previous solution.

GLS was able to migrate and shut down its existing data warehouse over to Snowflake on AWS in just 21 months — a time period which included six months spent updating the data structure of its core parcel system. 

As well as acting as a central store for the company’s data, this new data platform also provided an opportunity for cultural change. Using a data mesh architecture, GLS can now provide satellite teams with access to data, rather than having one team serving each individual reporting request. 

The company uses AI to enable natural language interactions with Snowflake, translating written requests into SQL queries and presenting the results in visualizations. It’s a process Herzog relates to one of his favorite shows, “Star Trek.” 

“It removes the problem of learning an interface,” he says. “Just like Jean-Luc Picard, you can say ‘Computer, can you tell me what the starship is doing?’”

This is the first step on the way to data democratization, and a much faster way of providing insight to business users. 

“Before, my team would be asked to produce a report on a certain topic and it would just go in a queue,” says Herzog.  “We realized that if we had this great technology in Snowflake we had to change the way we were working. That led to a data mesh approach, and Snowflake provides everything people need to do it the right way — it’s a very easy-to-use technology.”

One of the top priorities for GLS was ensuring data protection. To meet this need, Snowflake has implemented and continuously enhanced robust security measures to ensure that every aspect of the company’s data flow is handled safely, reliable and transparently.

With Snowflake, GLS has been able to bring new efficiencies to its analytics work, including a reduction in the time required to create last-mile reports from 12 hours to just one. Similarly, monthly closing reports that once took an entire day are now completed in just two hours. 

“Even though I had studied how Snowflake operates, seeing it in action during our POC was impressive — some queries ran up to ten times faster than our previous ones, without any optimization and on the smallest warehouse sizes available.” says Herzog. “Overall we’re twice as fast as we were, and that speed comes at a lower cost.”

The benefits of GLS’s new data platform extend beyond its data teams to its end customers, who now experience an even more reliable service and more accurate delivery forecasts.

“We have more transparency in our process now,” says Herzog. “We are faster in seeing where things went wrong and we can look into the future. That means we can better plan our delivery routes, understand what will arrive in our depots when, and ensure the forecasts we give to our customers are reliable.”