“It’s just the sheer performance of Coalesce. Our nightly batch processes are 75% faster. We also now have everything in Git, giving us version control and branching. And our deployments are 12x faster.”
A big basket of challenges
Challenges
Data siloed across 18 different brands with multiple transactional systems
Legacy ETL processes struggled to handle 10x peak holiday demands
Building or updating data sets in legacy system was cumbersome and resource-intensive
Behind the scenes at 1-800-Flowers.com, John Paradiso manages the BI and data management team that serves all 18 sub-brands as well as the corporate umbrella brand. His team of data engineers and analysts supports everything from Harry & David to Things Remembered to, of course, 1-800-Flowers.com.
The company’s growth-by-acquisition strategy had created a patchwork of disparate, siloed data systems. At one point, they juggled six or seven transactional systems. Each acquired brand brought varying levels of data maturity. Harry & David, for example, arrived with a data warehouse and CRM, while others had minimal data infrastructure.
As its business grew and evolved, 1-800-Flowers.com found its existing architecture increasingly problematic. “We get a lot of requests to add new data sets and new capabilities. In our previous system, it was quite extensive for us to be able to build this out,” says Paradiso. With increased focus on efficiency and optimization of their internal resources, the legacy system threatened to bottleneck the company’s data capabilities.
That legacy system also struggled to handle the company’s dramatic seasonal swings, when transaction volumes spike tenfold during Christmas, Valentine’s Day, and Mother’s Day. It would take four hours to rebuild the data warehouse, limiting updates to just a few times daily.
Creating a better arrangement
Solution
Evaluated CI/CD-specific tools but quickly discovered Coalesce could improve entire ELT process, not just CI/CD
Conducted focused one-month proof of concept with internal team and Coalesce experts
Successfully executed aggressive 4-month migration of 88 pipelines with help from consulting partners
When Paradiso and Ed Looft, Data Management & CRM at 1-800-Flowers.com, attended a 2024 Snowflake World Tour event in New York, they were simply looking for a CI/CD pipeline tool. But when Lee Dirks from DigBI Consulting declared during his onstage presentation, “I love Coalesce more than coffee, and I really love coffee,” they decided they had to dig deeper.
“We saw how much faster we could develop with Coalesce while still getting all the CI/CD capabilities, and that was what really sold us,” Paradiso says. The key differentiator was Coalesce’s seamless integration with Snowflake. “All the compute is happening on Snowflake, so your ELT is leveraging and gaining Snowflake’s elasticity,” explains Looft. “In the past, it was software running on a VM that had to pull data out, transform it, and then push it back—so processing was always limited by the capacity of that VM.”
With existing contract deadlines looming, the 1-800-Flowers.com teams embarked on an ambitious four-month migration of 88 pipelines containing 1,332 objects and 37,787 attributes, with support from both Coalesce and its partner Kipi.ai. “It was a much easier migration than I thought it would be. The learning curve for Coalesce was minimal; it’s very intuitive and easy to use,” says Paradiso. “The ability to create a custom node (object) to fit Coalesce within our existing patterns was important because a lack of flexibility would have been a deal-breaker—we weren’t looking to start from scratch.”
Delivering smiles all around
Results
75% faster nightly batch processing
Warehouse rebuilds cut from 4 hours to 1, enabling hourly refreshes and dramatically improving data freshness
12x faster deployments, enabling accelerated releases
A more productive team, operating with fewer resources than before
The impact of the Coalesce migration was immediate and dramatic. “Our nightly batch processes are 75% faster,” Looft reports. “Rebuilding the warehouse used to take four hours, meaning we could only refresh data a few times a day.” With Coalesce, that same process now runs hourly, fundamentally changing how fresh data flows throughout the organization. “Our deployments are 12x faster, so our ability to release changes is significantly improved.”
Beyond raw performance gains, the team has transformed how it manages and deploys code. With everything now in Git, the team has version control and branching, enabling continuous integration and deployment capabilities that were impossible with their previous system.
The productivity improvements the team sees are equally significant. “It’s early on now to quantify, but obviously, productivity gains in our ability to implement changes are much faster,” Looft says, adding, “I don’t have a number for that one, but it’s significant.” For a team that has shrunk over time, these efficiency gains are critical. “We need to do more with less. This actually helps us be a lot more productive with the time and resources we have,” he says.
Looking ahead, the team sees Coalesce creating new opportunities for data governance and testing. “We can now have built-in data quality tests,” says Looft. “The ability to add regression tests, to fully test changes against controlled data sets is exciting.”
The team also sees the potential to expand Coalesce access beyond technical users. “I think there is an opportunity to build out documentation so that business users can better understand what data is available to them,” says Paradiso.
The real payoff will likely be seen as the company starts leveraging its newfound agility. “We get a lot of requests to add new data sets and new capabilities. Coalesce makes it much easier to not only change what we’ve already built, but also to add additional capabilities,” Paradiso says. “Our leadership team will notice the difference in two ways: first, by getting access to data more frequently, and second, by how much faster we can develop and react. In short, our time to market for development efforts will shrink.”
With Coalesce in the basket, 1-800-Flowers.com is set to deliver not only smiles, but also insights, faster than ever before.
“It was a much easier migration than I thought it would be. The learning curve for Coalesce was minimal; it’s very intuitive and easy to use. The ability to create a custom node (object) to fit Coalesce within our existing patterns was important because a lack of flexibility would have been a deal-breaker—we weren’t looking to start from scratch.”
1-800-Flowers.com