Joybird reduces engineering time spent on customer data integrations by 93% with RudderStack

Highlights

• Joybird retooled their customer data stack using RudderStack for real-time data collection, Snowflake for their cloud warehouse, and Iterable for email automation
• RudderStack’s Event Stream and Reverse-ETL pipelines enabled Joybird’s marketing team to iterate faster on personalized email campaigns
• With RudderStack, Joybird’s data engineering team is able to focus on value-added activities instead of spending time managing brittle pipelines

Key Stats

• Data Engineering reduced the time spent on building new integrations and managing data pipelines by 93%
• Marketing team is now able to spin up new campaigns in an hour, down from 2 weeks
• Front-End team significantly reduced total time spent on implementing new event tracking dimensions on any given sprint from 15% to 1%

Overview

Founded in 2014, Joybird is an online-first direct-to-consumer manufacturer of high-quality, customizable furniture inspired by mid-century modern designs. Now a La-Z-Boy subsidiary, the company operates showrooms in Brooklyn, Washington, D.C., Chicago, and Los Angeles, but the bulk of its sales go through its website.

Each month, Joybird.com receives hundreds of thousands visitors. The company was keen to understand how visitors navigate the website, experience their customer journeys, and complete purchases. Joybird also wanted to connect users' actions on the website to downstream platforms, triggering automated emails and actions like adding people to their CRM.

The company was already tracking events on its website and wanted to build a unified event stream that fed downstream destinations, including Facebook Pixel and Kustomer. But it had trouble building the needed integrations with its previous customer data platform, Segment.

When Joybird moved its email marketing platform from Bronto to Iterable, it saw an opportunity to migrate its customer data platform from Segment to RudderStack at the same time.

Joybird's engineering, product, and marketing teams now enjoy seamless integration between RudderStack, its downstream destinations, and its Snowflake data warehouse. This new data stack has dramatically reduced the time to perform operations like turning destinations on and off, defining new dimensions, and adding new services. RudderStack has turned hours of data engineering time into minutes and weeks into hours.

Challenge: Moving Away from Brittle Integrations and Unreliable Data

Joybird's engineers were spending too much time on one-off activities. The Bronto integration stack was brittle, and they found themselves custom-coding integrations that broke whenever something on the front-end or back-end changed. It took weeks to respond to a simple ticket like adding a new event tracking dimension because they had to pass requirements back and forth and prioritize the many end-user requests.

Although there was a glut of customer data from different sources in Joybird's Snowflake warehouse, it wasn't reliable. Moving it from the front end to the back end and shuffling it between platforms and applications yielded inconsistent data definitions and schemas.

The data team didn't have confidence in the results: "We couldn't confirm that an order completed on Facebook was an order completed in Google Ads and on Pinterest," said Joybird Director of Analytics, Brett Trani. "We lacked the multi-touch attribution tools to gain insights into our customers' online behavior. "

To counter these inconsistencies and inefficiencies, Joybird jettisoned Segment for RudderStack. However, adopting a new platform to route customer data across its entire stack was only the beginning. RudderStack upended how Joybird manages customer data, leading the company to rethink how it used its stack.

Solution: Taking Control of Customer Data

Joybird used RudderStack Event Stream and Reverse ETL to take control of its customer data, leverage the full potential of its modern stack, and engineer a warehouse-first approach atop event-driven architecture. In doing so, the company transformed its Snowflake data warehouse. Instead of a silo that hampers the free flow of customer data, it is now a single source of truth that houses front- and back-end event data and routes it to downstream destinations.

“We have a very good modeling layer and consistent definitions across our entire ecosystem,” says Trani. “With RudderStack Reverse ETL actions, we can push event data downstream to Iterable and our CRM, Kustomer. With Segment, we were feeding customer and event data into Snowflake, but we couldn’t turn it into actionable information that could drive our marketing efforts.”

Joybird’s marketing team is now using RudderStack to test new activation platforms. Instead of spending valuable engineering time to build new instrumentations to validate every hunch, the company’s marketers can toggle RudderStack destinations on and off. They can experiment and iterate faster, test integrations, and launch new campaigns in real-time. In addition, to ensure data consistency regardless of destination, the team cleaned up the company’s website code, replacing multiple SDKs and Pixels with RudderStack’s SDK including GTM snippets and calls to random Facebook Pixels.

In addition, the marketing team is using RudderStack Reverse ETL to push enriched customer data from Snowflake to downstream destinations like Iterable. Instead of waiting for the engineering team to develop a custom integration, Joybird marketers can submit a destination to the data engineering team, which will spin up a campaign in an hour or two.

“We didn’t realize the power of RudderStack Event Stream and Reverse ETL at first,” adds Trani. “Either in terms of saving engineering time and allowing our downstream teams to work and iterate faster. Our CTO is very excited about using RudderStack to realize our event-driven architecture vision.”

Results: Saving Time, Money, and Effort

Joybird’s data warehouse is no longer a liability but a powerful reservoir of actionable information. “We ingest a lot of data, put it into Snowflake, and use RudderStack to get it downstream,” explains Trani. “More importantly, we can guarantee that the data is clean, consistently defined across our entire stack, and compatible with all existing and future platforms.”

RudderStack’s easy integrations also simplified Joybird’s move to a new email platform: “It took less than a month to migrate from Bronto to Iterable,” adds Trani. “We instrumented all the events, architected the workflows, and switched everything on in that time.”

“We were up and running with hundreds of automated email processes, and we used RudderStack to test and validate the reliability and consistency of our Iterable workflows, events, and destinations.”

Speed and efficiency aren’t the only benefits of RudderStack. The final consideration is price. RudderStack’s event-based pricing is trans- parent and provides full access to the platform’s Event Stream and Reverse-ETL tools, making it more attractive than platforms that offer pricing based on monthly tracked/active users.

“We have hundreds of thousands of people using our site every month,” concludes Trani. “We wanted a platform that would allow us to capitalize on that volume by capturing and analyzing their customer journeys across all our marketing channels. RudderStack allows us to do that without breaking the bank.”

Joybird Data Stack

Destinations: Snowflake, Iterable, Indicative, Google Analytics, Google Tag Manager, Bing Ads, Kustomer (webhook), Facebook, Criteo, Qualtrics, TV Squared, Posthog, Google Optimize

Sources: JS SDK, PHP Server SDK, Python SDK

Warehouses: Snowflake

Learn more about how Joybird is using RudderStack

Watch our on-demand webinar with Joybird’s Director of Analytics, Brett Trani, to learn about the Joybird team’s experience implementing and using RudderStack. You’ll also get to hear insights from experts at Snowflake and Iterable.