The physical world around us is rapidly evolving - be it a new housing complex around the block, a new grocery store around the corner, a kiosk, or ATM. As a leader in places data, SafeGraph knows exactly where points-of-interest (POI) are located on a map, how the places are connected to the surroundings, and how and when consumers interact with these places. Through its open-platform design, SafeGraph offers users access to over 9 million POIs and counting, enabling enterprises to make more informed and actionable data-driven decisions.
SafeGraph, founded in 2016, provides clean spatial POIs across the US, Canada, and the UK, along with essential metadata to its customers. SafeGraph uses Airbyte for its business analytics - allowing stakeholders to get just-in-time analytics on who, when, and how SafeGraph is used. As their platform grew, several key challenges needed to be solved.
Deeper understanding of the customer journey
For SafeGraph's growth teams, it was essential to track customer journey touchpoints across their platform accurately. This would help them answer questions like "What attracts customers to SafeGraph, what drives customers to sign up, and what customers do after they sign up?". A significant challenge for SafeGraph's team was that valuable customer journey data was strewn across several SaaS applications, from Salesforce and Marketo to Unbounce and Webflow. It was important to rethink their data architecture to solve this.
We want to see how customers arrive on our website, which tabs they utilize, and at what point do they sign up. This information helps us better understand the user journey. We are a data company, and so we have a lot of data sitting in a lot of different places.
With over 5000 clients worldwide and growing, SafeGraph had pockets of data scattered across multiple database locations. SafeGraph had to consolidate and analyze large amounts of dispersed client data to determine different points of interest for customers. Their old data architecture failed to meet these requirements and had several problems.
Unable to query across multiple databases
SafeGraph utilized an open source BI tool called Metabase for business analytics. Unfortunately, Metabase was not designed to query across multiple databases, and it did not meet our changing data needs.
We were using an open source BI tool called Metabase. Metabase does not have compute hosted on it, and it is not built to query across multiple databases. With Metabase, there was no query optimizer and intermediate storage of queries, which made it challenging to meet the requirements of our use case.
The inefficiency of cross-data syncing
Before redesigning their data platform, SafeGraph managed data synchronization operations using multiple tools. Segment was used to extract data from Salesforce. However, querying the Salesforce API in this way was incredibly time-consuming. A large part of the data was kept in a single RDS instance. Therefore, several scripts were written to synchronize data from different databases like Amazon Aurora and other RDS instances. Due to inefficient data syncing practices, SafeGraph could not quickly analyze data on customers to better understand how their business operates.
We had to manually pull data out of Salesforce, which was a nightmare. We even tried writing a script using SOQL to extract data, but that didn't work out too well.
SafeGraph had several issues with its old architecture. The company used Airbyte to resolve these issues and consolidate data, ensuring that everything ran smoothly.
Consolidated data provides a holistic view for easy analysis
SafeGraph handles a lot of data daily. This information is stored in multiple databases and is scattered across various networks. With Airbyte, SafeGraph can bring data from different sources into a single platform, providing a holistic view of the data. In addition, Airbyte's data transformation capabilities powered by dbt enables SafeGraph to transform data as it is moved and create data pipelines tailored to their needs.
We needed a tool where we could put all of our data into one place. Where Airbyte shined for us was the case of a multitude of small data sets that you have in different places and the ease of putting all that together to understand our business better.
Increased integration flexibility
SafeGraph has various data use cases, including streaming data, appending data, iterating projects, replacing tables, and more. Using Airbyte, SafeGraph's team was able to get more flexibility and choice. In many cases, with Airbyte's rich connector ecosystem, integrating it with existing systems was a simple plug-and-play process.
In other cases, SafeGraph utilized a 2-step process, connecting Airbyte with a third-party tool for indirect integration. For example, Wasabi and Airbyte were used to sync important customer information to an RDS database in the same data pipeline. Together, the direct and 2-step integration processes allowed SafeGraph to identify and prioritize their use case to migrate to Airbyte.
For us, native integrations are something we appreciate. Airbyte is a really fast solution to 90% of your data engineering problems. Whenever we need to extract data, we have to check that there's an integration with Airbyte, and since we already have it set up, we can add it to our data pipes easily.
Looking into the future
Today, Airbyte is an essential component of SafeGraph's platform. In the future, as SafeGraph scales its use cases, it is confident in Airbyte's capabilities and the support that it can get from experts in the community.
Being able to ask something in the slack community, get individual attention, and have someone answer super fast was good and helpful. One of our goals is to model our community after some aspects of the Airbyte community.
Start breaking your data siloes with Airbyte.
Agridigital, a digital supply chain platform, massively sped up their release cycle with Airbyte.