Building a scalable data stack: Three lessons from Backcountry

The outdoor gear retailer uses a modern data stack to overcome data silos, modernize legacy systems and improve customer experience.
January 12, 2023

Retail is a cutthroat business — and keeping a competitive edge requires the ability to use data and analytics to react in real time to emerging trends, customer preferences and backend operations. To maintain this edge and answer pressing business questions, Backcountry.com, a premium online retailer of outdoor gear and apparel, looked to create a scalable data infrastructure that would play a critical role in their digital transformation.

Using data and analytics as the thread that carries their digital transformation efforts, Backcountry wanted to overcome key business problems that were roadblocks to successful data initiatives. These challenges included data silos, complexities with legacy systems and managing customer expectations. 

As a team of six, Backcountry’s data engineering team handles everything from data engineering, reporting, visualization, web analytics and data science. They needed to better serve the needs of internal and external stakeholders by ensuring they have access to a centralized source of truth for fast, reliable insights and reporting.

Here are three ways Backcountry has built and leveraged a modern data stack to drive meaningful business outcomes.

[CTA_MODULE]

Lesson #1: Automate data ingestion and pipeline maintenance

Many organizations struggle with being inundated with a large and varied volume of data sources. This makes it difficult to analyze the data for insights, so realizing value from the data isn’t as fast or easy as hoped. Additionally, because of these challenges, moving data from one place to another has historically been constrained within organizations to the IT department.

At Backcountry, automating data movement with Fivetran has reduced pipeline issues by 30 percent, which has decreased the time spent on pipeline maintenance by engineers from 40-50 percent to 15 percent. 

The time savings are so significant that Backcountry says they save the equivalent of one data engineer’s annual working time. The accuracy of the data being provisioned through Fivetran is also better, which has allowed Backcountry to gain a more accurate and unified picture of the business.

The combined time-savings and improved accuracy have led to some significant improvements in how their data team operates. They can now: 

  • Shift focus from maintaining data pipelines to spending more time on activities that move the needle on analytics outcomes
  • Ensure data accessibility across more teams
  • Hold more productive meetings with business partners and drive data-driven decision-making across the organization

“In effect, we are offloading all of the heavy lifting back to Fivetran, so my engineers are focused more on delivering business solutions rather than just worrying about the ingestion aspect,” says Prasad Govekar, Director of Engineering at Backcountry.

Lesson #2: Build a modern data stack that’s simple and integrated

One of the key objectives for Backcountry was to ensure that the switch to a modern data platform didn't interrupt business capabilities, which meant that reliability and accuracy of the data were critical along with the uptime.

With these requirements in hand, Backcountry chose Google BigQuery as its cloud data warehouse and Fivetran for ELT as the foundational pieces of its modern data stack.

"What makes BigQuery unique from my perspective is that it enables real-time analytics," says Govekar. "There are a lot of streaming capabilities available within BigQuery, many of which Fivetran leverages to ingest the data as quickly as possible in real-time into BigQuery."

Fivetran’s deep integration with Google BigQuery optimizes data movement across its systems and makes Fivetran data pipelines available within the BigQuery user experience — all of which limits integration challenges and increases ease of use. 

Prior to Fivetran, Backcountry used Talend to load data to its Oracle data warehouse and Oracle Business Intelligence for reporting. This setup made it difficult to meet the demand of the business for real-time analytics. It also required a lot of manual work and maintenance. 

“We were constrained by physical infrastructure limitations, siloed operations, and sometimes…the inability to evolve,” recalls Govekar. “The data engineering teams were brute forcing their way to keep things operational, the focus was being diluted to keep the lights on. We were spending way too much time to keep up with the daily enhancements.”

[CTA_MODULE]

Lesson #3: Develop a data strategy that supports AI and ML

Many companies struggle with data initiatives like AI and ML due to data pipeline issues. In our recent global survey of senior IT and data science professionals, data scientists spend an average of 70 percent of their time working with and preparing data versus building AI models. For Backcountry, automating data ingestion has alleviated many of these data preparation challenges. Govekar notes that having the right data in the right format has been critical to enhancing their AI and ML capabilities —- and that Fivetran has played a key role. 

"You have to move data from where it is today in a very easy way, and Fivetran is the best in terms of how it does it effectively,” says Govekar. “I was really surprised with the ease of experience in terms of being able to pick the sources and set up the connectors."

Govekar adds, "It's compelling as an organization to get started and start experimenting because a lot of the innovation comes from trying out different scenarios, building out new prediction models and seeing what works.” 

Backcountry has been experimenting with its AI and ML efforts in pricing, promotion and markdown planning as well as helping with improving workforce planning, including creating a staffing plan for physical stores and their distribution centers.

“We can really leverage everything that is there within the Google Cloud BigQuery stack and ask our data scientist to come up with things that are even better and enable our business partners to do well,” says Govekar.

By leaning on the power of the modern data stack, Backcountry freed up its data team’s time to work on more strategic projects, created processes that empower other business users to self-service their data insights and ensured the retailer has the capabilities for continued innovation.

[CTA_MODULE]

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Blog
Blog

Building a scalable data stack: Three lessons from Backcountry

Building a scalable data stack: Three lessons from Backcountry

January 12, 2023
January 12, 2023
Building a scalable data stack: Three lessons from Backcountry
The outdoor gear retailer uses a modern data stack to overcome data silos, modernize legacy systems and improve customer experience.

Retail is a cutthroat business — and keeping a competitive edge requires the ability to use data and analytics to react in real time to emerging trends, customer preferences and backend operations. To maintain this edge and answer pressing business questions, Backcountry.com, a premium online retailer of outdoor gear and apparel, looked to create a scalable data infrastructure that would play a critical role in their digital transformation.

Using data and analytics as the thread that carries their digital transformation efforts, Backcountry wanted to overcome key business problems that were roadblocks to successful data initiatives. These challenges included data silos, complexities with legacy systems and managing customer expectations. 

As a team of six, Backcountry’s data engineering team handles everything from data engineering, reporting, visualization, web analytics and data science. They needed to better serve the needs of internal and external stakeholders by ensuring they have access to a centralized source of truth for fast, reliable insights and reporting.

Here are three ways Backcountry has built and leveraged a modern data stack to drive meaningful business outcomes.

[CTA_MODULE]

Lesson #1: Automate data ingestion and pipeline maintenance

Many organizations struggle with being inundated with a large and varied volume of data sources. This makes it difficult to analyze the data for insights, so realizing value from the data isn’t as fast or easy as hoped. Additionally, because of these challenges, moving data from one place to another has historically been constrained within organizations to the IT department.

At Backcountry, automating data movement with Fivetran has reduced pipeline issues by 30 percent, which has decreased the time spent on pipeline maintenance by engineers from 40-50 percent to 15 percent. 

The time savings are so significant that Backcountry says they save the equivalent of one data engineer’s annual working time. The accuracy of the data being provisioned through Fivetran is also better, which has allowed Backcountry to gain a more accurate and unified picture of the business.

The combined time-savings and improved accuracy have led to some significant improvements in how their data team operates. They can now: 

  • Shift focus from maintaining data pipelines to spending more time on activities that move the needle on analytics outcomes
  • Ensure data accessibility across more teams
  • Hold more productive meetings with business partners and drive data-driven decision-making across the organization

“In effect, we are offloading all of the heavy lifting back to Fivetran, so my engineers are focused more on delivering business solutions rather than just worrying about the ingestion aspect,” says Prasad Govekar, Director of Engineering at Backcountry.

Lesson #2: Build a modern data stack that’s simple and integrated

One of the key objectives for Backcountry was to ensure that the switch to a modern data platform didn't interrupt business capabilities, which meant that reliability and accuracy of the data were critical along with the uptime.

With these requirements in hand, Backcountry chose Google BigQuery as its cloud data warehouse and Fivetran for ELT as the foundational pieces of its modern data stack.

"What makes BigQuery unique from my perspective is that it enables real-time analytics," says Govekar. "There are a lot of streaming capabilities available within BigQuery, many of which Fivetran leverages to ingest the data as quickly as possible in real-time into BigQuery."

Fivetran’s deep integration with Google BigQuery optimizes data movement across its systems and makes Fivetran data pipelines available within the BigQuery user experience — all of which limits integration challenges and increases ease of use. 

Prior to Fivetran, Backcountry used Talend to load data to its Oracle data warehouse and Oracle Business Intelligence for reporting. This setup made it difficult to meet the demand of the business for real-time analytics. It also required a lot of manual work and maintenance. 

“We were constrained by physical infrastructure limitations, siloed operations, and sometimes…the inability to evolve,” recalls Govekar. “The data engineering teams were brute forcing their way to keep things operational, the focus was being diluted to keep the lights on. We were spending way too much time to keep up with the daily enhancements.”

[CTA_MODULE]

Lesson #3: Develop a data strategy that supports AI and ML

Many companies struggle with data initiatives like AI and ML due to data pipeline issues. In our recent global survey of senior IT and data science professionals, data scientists spend an average of 70 percent of their time working with and preparing data versus building AI models. For Backcountry, automating data ingestion has alleviated many of these data preparation challenges. Govekar notes that having the right data in the right format has been critical to enhancing their AI and ML capabilities —- and that Fivetran has played a key role. 

"You have to move data from where it is today in a very easy way, and Fivetran is the best in terms of how it does it effectively,” says Govekar. “I was really surprised with the ease of experience in terms of being able to pick the sources and set up the connectors."

Govekar adds, "It's compelling as an organization to get started and start experimenting because a lot of the innovation comes from trying out different scenarios, building out new prediction models and seeing what works.” 

Backcountry has been experimenting with its AI and ML efforts in pricing, promotion and markdown planning as well as helping with improving workforce planning, including creating a staffing plan for physical stores and their distribution centers.

“We can really leverage everything that is there within the Google Cloud BigQuery stack and ask our data scientist to come up with things that are even better and enable our business partners to do well,” says Govekar.

By leaning on the power of the modern data stack, Backcountry freed up its data team’s time to work on more strategic projects, created processes that empower other business users to self-service their data insights and ensured the retailer has the capabilities for continued innovation.

[CTA_MODULE]

Google Cloud and Fivetran: Open cloud meets automated data movement
Download the ebook
Backcountry leverages Fivetran and Google Cloud to fuel their data-driven transformations
Watch the webinar
Power your data movement for Google BigQuery with Fivetran
Sign up for a free 14-day trial now

Related blog posts

Build vs. buy data pipeline: Four factors Autodesk considers
Blog

Build vs. buy data pipeline: Four factors Autodesk considers

Read post
Why the modern retailer needs a modern data stack
Blog

Why the modern retailer needs a modern data stack

Read post
Five reasons to attend Modern Data Stack Conference 2023
Blog

Five reasons to attend Modern Data Stack Conference 2023

Read post
No items found.
Driving self-service analytics on the modern data stack
Blog

Driving self-service analytics on the modern data stack

Read post
The modern data warehouse: Benefits and migration strategies
Blog

The modern data warehouse: Benefits and migration strategies

Read post
Build vs. buy data pipeline: Four factors Autodesk considers
Blog

Build vs. buy data pipeline: Four factors Autodesk considers

Read post

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.