Data-Driven Approach to the International Supply Chain
Data-Driven Approach to the International Supply Chain
Flexport is a licensed freight forwarder for modern logistics teams that move freight around the globe by air, ocean, rail, and trucks for the world’s leading brands. Their software is a driving force behind managing the complexities of international trade.
When unforeseen factors arise, Flexport needs answers on how to alter shipping routes in a manner of minutes, not days. To ensure they see quick turnarounds on data analysis, Flexport relies on Periscope Data to provide its real-time data analytics platform, with a flexible workflow that enables its analysts to instantly create new analysis without proprietary data modeling.
Flexibility Supports Data Team Growth
Now the Director of Data Science, Patrick Steigler initially joined Flexport in 2015 as their first data scientist, responsible for leading efforts to ramp up the company’s growth in data engineering, operational engineering, business and product insights, data science and advanced data modeling. He has since brought on a team of data analysts, data engineers, and data scientists to support a comprehensive data analysis effort as the company grew from 30 people to more than 1,000 in just three years.
In the early stages of building out the Data Science team, Flexport rolled out Periscope Data as its platform of choice for data aggregation and analysis. Steigler says Periscope Data’s Views and modeling layer enabled his team to accelerate quickly by building out its ETL in an iterative cycle and reacting quickly to engineering changes.
“The smaller you are, the faster your code base is changing – so we had a high-velocity code base,” said Steigler. “Having the strong modeling layer from Periscope Data was essential for propelling our framework in the initial growth of Flexport’s Data team.”
Additionally, just as the modeling layer enabled Flexport’s backend development, Periscope Data’s analyst-focused UI, with a focus on table-level contextual information and filters, fit perfectly into the workflow of Flexport’s growing data team.
“The nature of our work at Flexport is that we have a lot of tables, for all the supply chains and tariffs and lists of ports that we have to track,” said Steigler. “So Periscope Data’s flexibility for creating dashboards around those tables has been huge. It really supports our operational processes and ultimately accelerated Flexport’s growth overall.”
Today, more than 50 percent of the company logs in to Periscope Data daily to track trends and query data, while more than 90 percent of Flexport employees log in to Periscope Data monthly to view high-level reports.
“I’m a huge supporter of Periscope Data – they’ve really improved the state-of-the-art in the analytics space,” said Steigler. “I’ve been working on data science teams for almost a decade, and it’s just a whole different world in terms of the tooling that I have now with Periscope Data - it has drastically improved my workflow.”
Ad Hoc Analysis for Supply Chain Events
Steigler divides the company’s use of Periscope Data into four main use cases. The first includes standard company and group-level metrics or KPIs, which are accessed by hundreds of employees to understand how the business performs. The second is dataset management, and the third is operational reporting.
“We do a lot of financial processes and business analysis functions, and Periscope Data serves as a portal to manage all our datasets that are internal to Flexport,” said Steigler. “And our operational reporting is very table heavy, so we rely on Periscope Data to inform users of the tasks they have to complete that day, sorted by group or role and ranked with some contextual information directly in our software.”
The fourth and often most critical use case is for ad hoc analysis, which comes in handy for instant analysis when new questions arise. Every day brings new conflicts and situations around the globe that change shipping plans for Flexport and its customers. To prepare for those events that always seem to pop up, Steigler developed a series of calculators that exists as views within Periscope Data where he can quickly plug in new factors and determine the direct impact on Flexport’s clients.
“The global supply chain has to be incredibly resilient,” said Steigler. “There are political conflicts, there are crazy weather events, there are pirates – when you’re looking at the entire globe, these things actually happen very frequently, so we needed a good way to respond and help our customers make adjustments on the fly.”
One of those early data models was built to understand the impact of port strikes. Steigler once arrived at work to the news that a port strike in Asia had stalled all the ships carrying goods to and from that region. Flexport needed to quickly advise its customers on how to reroute goods to arrive at their final destinations on time without increasing costs.
Luckily, with Periscope Data’s model-as-you-go approach, Steigler and his team of data analysts were able to jump right in around 9 a.m. and deliver recommendations to the executive team by noon on how to reroute goods around the world.
“Periscope Data gave us the filters we needed to select a location, create a 50-mile radius around that location, and really capture everything that might be affected for a date range going through that area,” said Steigler. “With Periscope Data, we really have the flexibility to create instant models like that very quickly, and we can now use them to answer new questions about other strikes as they come up.”
Most recently, Flexport’s data team developed a “Tariff Calculator” to help clients understand how the federal government’s newly proposed trade tariffs would impact their bottom line. The calculator is built off a giant data set of tariff codes called “HTS codes” and available via an intuitive dashboard in Periscope Data that Flexport’s marketing and operations groups can filter to see the impact on their individual customers. Now, each time the government announces new tariffs or revisions, Flexport’s data science team simply uploads the new rates into Periscope Data and can immediately enable teams to see how the impact filters down to individual clients.
“With each supply chain event, there are hundreds and potentially thousands of people who need to take quick action based on these delays to reroute their goods and supplies,” said Steigler. “Analyzing and visualizing this data in Periscope Data helps our marketing team explain the magnitude of impact each tariff makes on the supply chain, and helps operational groups and client managers filter for just their clients and make recommendations directly to them on how to take action.”
“Periscope Data gave us the filters we needed to select a location, create a 50-mile radius around that location, and really capture everything that might be affected for a date range going through that area. With Periscope Data, we really have the flexibility to create instant models like that very quickly, and we can now use them to answer new questions about other strikes as they come up.”
Permissioning and Data Sharing with Git Integration
Heading into 2019, data democratization for Flexport’s 1,000+ employees is a major goal for Steigler and his team. Despite the growth of the data team, Flexport’s 14-person data team don’t have nearly enough time to satisfy the increasing hunger for data analysis on all the company’s products and services.
One way they are expanding access is by leveraging Periscope’s “Data Discovery” feature, an intuitive drag-and-drop interface that allows non-technical users to easily explore data, perform basic calculations and aggregations, create rich visualizations and share dashboards without any coding knowledge. Data discovery was recently rolled out to 50 new users at the company who are exploring and filtering data on their own, based on curated data sets provided by the data team.
“The biggest thing we’re trying to wrap our heads around is how to empower the business users who are closest to the problem, to investigate and create solutions for the challenges they encounter with data,” said Steigler. “So the Data Discovery feature really lowers the barrier to entry for those users.”
Another feature that Steigler and the Flexport team have been eagerly anticipating is Periscope Data’s new Git integration, which allows customers to connect any Git repository to Periscope Data. Once the connection is made, content created in Periscope Data is automatically synced to the related Git repository and saved as a series of files and folders that store reusable code and configurations for Dashboards, Charts, Views, Snippets, Filters, and Alerts.
Currently, Flexport uses multiple Periscope spaces as its data marts, giving each department in the company its own space. He says that while that approach helps with governance around data access, it has also created a roadblock for how departments share data with each other. However, with Periscope Data’s Git integration, he’s now able to build a much clearer workflow with proper permissioning.
“In my mind, the Git layer is really the unblocker for enabling easy collaboration,” said Steigler. “Our goal is to have the Periscope Data main space as a production environment, where the only way anything gets published to that is through Git and is confirmed by the data team. Everything else can stay in the department level spaces, where users own all the data in that space.”
The final step for Flexport’s data democratization efforts is expanding their use of the Periscope Data Cache, which will enable users to more easily interact with their data and speed up the time-to-value for Flexport.