Finding the key to efficient ETL
Working with data is at the heart of many companies' businesses, and that's true of Nokē ("no key"), but the company is strongly planted in the physical world as well. It manufactures smart padlocks, door controllers, and other physical security devices, and develops the software that powers them.
Hardware Product Manager Preston Wily is responsible for hardware development, supply chain, and QA at Nokē — and because the company doesn't have a dedicated data team, he also handles data analytics. Nokē replicates data from sources such as MySQL, Postgres, and Jira to a Google BigQuery data warehouse for business intelligence (BI) reporting.
"I was surprised that moving my data from Google Cloud SQL instances to BigQuery was so difficult," Wily says. "I started using Stitch to replicate data from a custom view I had set up so I could limit the number of rows and filter out unnecessary data. When that taxed the CPU I started using binary replication, and that's been awesome (at the expense of having to clean my data more in BigQuery)."
Wily says working with Stitch is easy. "After I found Stitch I had it up and running within a day. I was impressed with how easy setting up Stitch and creating the first replication was. I also really liked that data replication for the first 14 days is free — that makes it easier to justify bringing in a huge database of transactions.
"Now that we're using Stitch, our product dev and QA teams can analyze customer feedback and device telemetry data quickly to make sure we're solving problems before our customers notice them. This data-driven approach began in our hardware department, but the rest of our company is taking notice. I believe we'll have other departments investing more time and resources in leveraging data to make better decisions as they see our results."
Just how much of an impact has Nokē's data-driven initiative had? "It's hard to quantify how much Stitch has helped us in terms of efficiency or hours saved," Wily says. "Most of the data that our hardware, supply chain, and QA teams are using for BI just wouldn't have been analyzed before since querying in our native databases was so inefficient. Most of the improvements we've made to our hardware over the past six months have been data-driven; we've turned product weaknesses into strengths, which is really driving up the value of our brand."
I was impressed with how easy setting up Stitch and creating the first replication was.
Hardware Product Manager