By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Sep 26, 2022

Meero - Customer Story

Post by
Benedetta Cittadin
&

How the Meero Data Team cut troubleshooting time by half and evangelized business users about Data Observability with Sifflet

When Meero came to Sifflet, they were looking for a tool that could help them sustainably improve the quality of their data assets. The data team was submerged with questions from the business; why has this number changed from last Monday? Why do I see different ARR numbers? Are you sure that all the KPIs represented here are accurate? Etc.

In addition to monitoring the quality and accuracy of the data assets, Meero was also looking to ensure the reliability of their data pipelines and observe schema changes and other metadata-related metrics. More than anything else, Meero wanted to know the root cause of each anomaly detected and conduct proper incident management reporting. Sifflet ticked all of their boxes.

 

Meero is at the forefront of innovation and creating tools to deliver businesses customized photos and video through end-to-end shoot management at top quality and unbeatable delivery times. The Data team reflects the company's level of sophistication, using a data stack composed of: BigQuery, Looker, Airflow, and in-house Python scripts.

Cloud warehouse, BI tool, transformation layer and alert channel
Meero's Data Stack

The Challenges

Data Reliability 

As the company was scaling fast, it relied heavily on data to drive business decisions. For the data team, that meant that ensuring the reliability of the data assets was critical. 

Common data quality issues like freshness (is the data up-to-date?), volume (are there missing values? Missing rows? Missing data records? Incomplete pipelines?), schema (did the structure of data change?), etc. were simply not tolerated. 

Root Cause Analysis & Incident Management

According to Laurent, Head of Data, getting an alert when something breaks was only part of the solution. He wanted a tool for his team to make them proactive when dealing with data quality. Oftentimes, the team was able to do manual testing and know when a pipeline was breaking. Still, the biggest challenge was to assess the downstream impact and act accordingly before it became a business problem. Furthermore, as their platform continues to evolve, keeping up with the lineage of the data was simply impossible.

Sifflet in Action

Sifflet was able to connect to Meero’s data platform in less than 15 minutes, covering even their less standard and custom-built assets. The Auto-coverage feature allowed them to expand ML-based anomaly detection to thousands of data assets with the click of a button. After setting up the alerts, the preferred channel of communication, and the frequency, the team quickly started receiving alerts about a data set feeding a critical Looker dashboard. This gave the data team the visibility they needed to take appropriate action and be proactive before the business realized there was an issue. 

In addition to Incident Management, Sifflet allowed Meero Data Engineers to troubleshoot problems quickly by pointing them directly to the root cause. Thanks to Sifflet’s data monitoring and lineage, the data team at Meero saves between half a day to a full day of research every time there is an error in the data.

Key Results

Time

Sifflet significantly reduced the time the data team spent troubleshooting and getting to the root cause of a data problem, saving them about 50% of the time.

Laurent recalls a specific incident:

“On occasion, we manually import quite a few data sources at once. These data sources are handled by business teams and can often induce errors that impact some key tables. One time, Looker was showing the wrong amount for some critical bills, everything was double or triple what it was supposed to be. There wasn't a specific logic to it, making it quite difficult to debug. Sifflet caught the issue by sending us an alert on Slack, pointing us directly to the root cause, and it took us 30 minutes to understand and solve the problem. Without Sifflet, we would have spent countless hours going through the 30 external sources one by one and trying to figure out where the anomaly came from.”

Increased Trust & a Data-Driven culture

Thanks to Sifflet real-time dashboards, the data team can monitor the health status of the data assets, allowing them to focus on value creation. In addition, Sifflet enables data consumers to leverage data to make informed decisions without worrying about the data assets' reliability. Thanks to the friendliness of its UI, Sifflet allowed even the least technical personas to be informed about the health status of the data, bolstering a Data-Driven culture within the organization. 

Increased NPS of the Data Services

As the Head of Data, Laurent is aware of the importance of data quality. However, this is not as obvious for data consumers in other departments or teams. For him, giving access to Sifflet can help raise awareness among less technical data consumers about the importance of data accuracy and quality - consequently exporting these best practices to other departments.

The Modern Data Stack is not complete with an overseeing observability layer to monitor data quality throughout the entire data journey. Do you want to learn more about Sifflet’s Full Data Stack Observability approach? Would you like to see it applied to your specific use case? Book a demo or get in touch for a two-week free trial!

Related content

Stay tuned !