Fitness specialist CrossFit needed to find a way to integrate disparate data sources in order to bring its corporate information together and find new ways to interact with its customers.
To that end, the firm uses CrossFit uses Snowflake’s Data Cloud to store all its information. The company was introduced to Snowflake through Twilio. It was using Twilio’s customer data platform (CDP) Segment to collect client profile information.
CrossFit needed a place to store all the data it was getting from the CDP, according to Jay Meyering, Senior Manager for Software Development at the company:
At the time, we were just using MySQL. We started looking at warehouses and we’re an AWS shop, so I was thinking Amazon Redshift at first. But as I started looking around, Snowflake came up everywhere.
The procurement process started about two years ago. After looking at Redshift and PostgreSQL, Meyering recognised he wanted to put a straightforward approach in place. And after using the trial version of Snowflake, his team ploughed into the full implementation:
We met with the Snowflake team and they just walked us through everything. They made it clear how we could use it. There were a lot of people who were proponents of the technology and it also seemed like the easy approach. I saw I could just plug it in. And today, I hardly do any administration of Snowflake.
Meyering says getting data from their existing tools into Snowflake was simple, including third-party tools they were using through Segment. The company has a broad range of web-based and mobile properties. He recalls that it was more challenging to move data from some of their internal systems, including older PHP systems.
Now, all the information is in one place and Meyering says his team can use other tools – including the collaborative platform dbt and Segment – to turn data into insight:
I now have all this source data coming in. To make sense of it, that’s where we’re leveraging DBT to take all this source data, transform it, clean it up, get a consistent view, and present it as analytics schemas. Segment has this new feature called Reverse ETL, so I’m able to leverage all the analytics data and pipe it back into the tools that we’re using, like Salesforce.
Rising to the challenge
Meyering says CrossFit’s data efforts are mainly focused on its customers. One of the key ways the data is being used is through the marketing communication platform Braze:
We wanted to enable our marketing teams to self-serve. Now, we’re feeding them data points about our users. We’ve consolidated those in the clean warehouse and then we push those into Braze. And then they can just activate information within Braze and say, ‘Well, I want to collect this particular group of people who participated in our CrossFit Open in 2022.
He gives an example of how the new approach has supported change. One of the firm’s analysts in the United States wanted to know how many CrossFit athletes who competed in the Open had also taken an education course. Running that analysis before Snowflake took days because the information was spread across many systems:
Our databases weren’t connected very well. They had incomplete information. We had duplicative sources. Bits and pieces of information were in different places and bringing it all together was really hard. We could get results, but it took time.
Today, it’s much easier to find answers to business questions. Segment brings data from the company’s web and mobile apps. The automated data platform Fivetran, meanwhile, is used to pipe data into Snowflake, where raw data is held in a consolidated location:
It’s just allowing us to get a picture of who our users are, what they’re doing and how we relate to them.
The biggest technical challenge was cleaning up the source data before the move to Snowflake. The team also took the opportunity to remove duplication, where two or three tools were doing similar things. Meyering says:
As I learned about all the mess we’d got with our source data, I was feeding that back to the development teams and saying, ‘Look, we can clean this up. There’s an easier way to do this. There’s these inconsistencies. Let’s try to fix our source systems and make it even cleaner.
The hardest part of the switch to Snowflake has been getting people across the business used to a new way of working. People were wedded to older tools, but top-level support from the senior team helped smooth the switchover, says Meyering:
Teams were holding on to their old way of doing things. But once they started seeing the benefits, everyone was psyched about the access to data. Having top-down support was huge and helped push our approach.
There’s a growing realisation across the business that people can do so much more with the right kinds of integrations, he adds:
Once we get the movement of data figured out, and the right tools in place, we can save ourselves so much complexity and effort. We won’t develop as much, maintain as much, and it actually put more capabilities in the hands of our business users.
The long-term aim over the next 12 to 24 months is to ensure customer personas and requirements are clear, whether that’s getting fit or coaching others. Meyering says it’s crucial CrossFit can talk to its customers about the things that are important to them:
Getting to know who our user base is the big value unlock that we’re trying to build as a business. We want to say, ‘Here are the things that really work and here’s how we spread the word and push more people to the gyms’.