Meet Soda, a data monitoring platform that is supposed to help you discover problems with your data processing setup. In this way, you can react as soon as possible and ensure that you keep the complete data picture.
If you are building a digital-first company, you and your customers can generate a ton of data. And you can also leverage that data to adjust your data yourself – hotel pricing, finding the right restaurant on the food delivery website, applying for a loan with a fintech company, and more.
“Companies build a data platform – as they call it – in one of the big three clouds [Amazon Web Services, Google Cloud, Microsoft Azure]. They made their data available there and they make it available for analytics and others.
You can then tap into those data lakes or data warehouses to display analytics, visualize your data, monitor your services, but what if there is a problem with your data workflow?
It may take you a while to realize that there is some missing data, or that you are missing some stuff. For instance, Facebook missed the average video view for many years. When you look at that issue, a significant part of your business may be affected.
Soda wants to capture data issues as soon as possible by monitoring its data automatically and at scale. “We sit at the front, closer to the source of the data” said Maschelen.
When you set up a soda with your data platform, you immediately get some alerts. Soda tells you if something is off. For example, if your application today generates only 6,000 records, while you usually create 24,000 records in 24 hours, then there is a possibility that something is wrong. Or if you usually get a new entry every minute and 15 minutes have not entered, your data may not be refreshed.
“But it only covers a small part of the data issues being considered. There is more logic that you want to test and verify, ”Maschelin said.
Soda lets you create rules to test and validate your data. Basically, think about the test suite in software development. When you create a new version of your app, your code needs to pass several tests to ensure that nothing significant is going to break with the new version.
With soda, you can immediately check the data and get results. If the test does not pass, you can react programmatically – for example, you can stop a process and quarantine data.
Today, the startup is also launching Soda Cloud. It is a collaborative web application that gives you visibility into your data flow across the organization. In this way, non-technical people can easily browse the metadata to see if everything is flowing correctly.
Originally, soda customers use Soda SQL, a command-line tool that helps someone scan data, along with Soda Cloud, a web application for viewing Soda SQL results.
Beyond those products, Soda’s approach is that software products are becoming a whole category. Development teams now have a ton of dev tools available to automate testing, integration, deployment, versioning, etc., but there is much potential for tools specifically designed for data teams.
Soda recently created a $ 13.5 million Series A goal ($ 11.5 million) led by Cingular, a new Paris-based VC fund that I covered earlier this week. Soda seed investors Point Nine Capital, Hummingbird Ventures, DCF and various business angels also participated.