All right

Masthead Data đť—‘ All Right

Masthead Data Helps All Right Improve Their Marketing Data Quality and Regain Trust in Data-Driven Decisions

All Right is a platform for English language learning, designed for children four years or older, which combines live lessons with teachers and homework with AI-powered “tutors”.

All Right’s mission is to help kids all around the world to learn and practice speaking English. All Right’s unique approach has been proved by exponential growth in several markets. From day one, the All Right team has relied on a data-driven culture to guide the platform’s growth and product development.

Oksana Sheidaiva, the Head of All Right’s Analytics, emphasizes that having accurate and reliable data is essential for growing the business and guiding the learning success of the All Right students.

Ensuring Data Reliability

Before Masthead Data, All Right did not have a system to identify data irregularities as they occurred until it caused bad consequences. All Right relies on demand generation platforms like Facebook Ads, Google Ads, Instagram, and so on. It also means that demand-side generation platforms (DSPs) are the core data for their decision-making in both customer acquisition and customer retention. The data from DSPs is pretty static regarding the schema, though the data delivery pattern can change. Having up-to-date and high-quality marketing data enables marketers to optimize spending and allocate marketing budgets toward best-performing campaigns. Any data issue regarding marketing performance is a high-priority data issue directly affecting the marketing bottom line.

The biggest challenges occurred when the marketing team pinged the analytics team repeatedly on inconsistencies on the Marketing Performance Dashboard. At the same time, Oksana’s team could not know immediately what exactly went wrong. All Right decided to use Masthead Data to monitor their data health at scale in real-time.

Masthead Data platform enables All Right data team to quickly identify and fix issues as they happen before they cause any downstream problems. Masthead Data also allows All Right to track the health of their data over time, so they can proactively prevent issues.

Data Integrity: Building a Scalable Data Infrastructure is a Challenge

The amount of data and the number of operations built upon it are growing exponentially. At All Right, the data team supports all aspects of the business — not only the marketing operations but also product development, customer success, sales, and all other processes aimed at improving every stage of the user journey. It’s hard to understand what errors, in which data tables affect the data quality of a particular dashboard. Oksana says,“ Before Masthead, I was the only one who knew the table dependencies in BigQuery by heart; that’s only because I architectured them. Not to say that it was not a scalable approach, but having column-level lineage helped us understand all the table dependencies. At the end of the day, it helped the data team ensure data integrity at All Right.”

Fixing a problem in a dashboard might not look like a problem at first glance, though it could be challenging to fix and check all data upstreams. A minor change in the upstream table schema could impact ten executives’ dashboards at once. Understanding schema changes in the upstream table or initial data source was an issue that caused fire drills for the data team. “Having Masthead Data in place reduces the root cause analysis of any schema changes and any upstream changes for our data analytics team. Masthead column-level lineage helped us better understand dependencies inside Google Big Query (GBQ) and enhance collaboration inside the data team and beyond when it comes to data activation.’’

Data Quality Solution: Masthead Data

In short, every team faces the buy or build dilemma. At first, it seemed like an obvious choice to implement dbt and SQLs tests. However, this would increase the workload on the team and take up precious time. Instead of delivering insights to the business, the team would have been building, implementing, and maintaining in-house solutions. Oksana understood that this was not a strategy to bet on. Before making the decision, Oksana, as Head of Analytics, decided to run a POC with Masthead. 

Thanks to Masthead’s no-code deployment, it took her only 30 minutes to integrate the platform with All Right’s data infrastructure. Masthead does not require downgrades in security standards, as it does not query data tables and monitors data health using only Google Cloud Product (GCP) native logs.  Masthead is HIPAA compliant so the decision was easy to make.

Masthead helped the All Right team answer questions such as:

Is the data in GBQ up-to-date?

Did we receive enough rows?

Is the data complete?

Did any schema changes happen?

What columns and tables are affected by anomalies?

The Outcome

To achieve data reliability, All Right used automated data anomaly monitoring with Masthead, which monitored every entity inside Google BigQuery. The finding was that one of their most critical marketing dashboards was affected by duplication in marketing expenses. Masthead Data spotted the abnormal growth in volume in a particular table and instantly notified the data team, which helped prevent inaccurate and costly decisions. Masthead Data’s ability to spot spikes in marketing spend and promptly prevent inaccurate decisions has won trust among both data and marketing teams right off the bat.

Courtesy of Masthead Data.
Courtesy of Masthead Data.

Another significant advantage delivered by Masthead Data to the All Right team was column-level lineage. Once an anomaly happened, the data team at All Right received an immediate alert and saw the impact of the data issue mapped on the lineage. 

That allowed Oksana’s team to promptly identify data consumers impacted by the anomalies and work proactively to fix the data issue.  Understanding which upstreams and downstreams were impacted by the data anomalies on the column level has enhanced team collaboration, reduced the time spent on identifying and fixing data errors, and minimized data outage time.

Data Quality Approach with Masthead Data: The Results

For All Right, finding out about data issues in real time was a big moment of epiphany. They could react fast once the issue happened, trace the implications downstream, and react proactively to any unexpected changes happening in Google BigQuery. Most importantly, it didn’t require the data team to spend time setting up alerts or tests to validate data because the Masthead’s ML algorithm automated that part. Masthead’s automated data monitoring and lineage became All Right’s go-to data governance and quality management solution. 

 

With Masthead’s reliability solution, All Right has achieved: 

  • Increased work time efficiency for the data team, having Masthead as the consolidated solution to observe data and issues at scale in every entity inside GBQ, which helped reduce the time spent on fire drills for the data team. 
  • Enhanced collaboration between the data team and data consumers. Masthead helps understand table dependencies inside GBQ and create an integrated picture of data usage patterns. 
 

With Masthead in All Right’s data stack, we look forward to our partner’s success by equipping them with the safest reliability solution. 

Special kudos to Oksana and the rest of the All Right team!