Want to know more?
Are you interested in this project? Or do you have one just like it? Get in touch. We’d love to tell you more about it.
Saving $3m with better data health
Media and entertainment
How a global media business made Video on Demand services more effective using our Data Health Check
Our customer is a multinational entertainment and broadcasting company based in the US.
The organisation wanted to improve its service for Video on Demand customers and turned to Equal Experts to improve its underlying data processes and insights.
Our Data Health Check helped the customer to speed up data ingestion times from 24 hours to under 2 hours, while consolidating platforms and adding automation to analytics and reporting has driven $3 million in savings and improved data quality by 20%.
Outcomes
$3 million
savings achieved
20% increase
in data quality
About the client
Our client is a global entertainment and broadcasting company, with headquarters in the US. It operates broadcast television channels, a radio network and a video-on-demand streaming service.
Industry
Media and Entertainment
Location
USA
Challenge
Understand and improving customer experience
The entertainment industry is changing rapidly, with consumer adoption of streaming services and video on demand.
Our client, a global media and entertainment brand, needed to understand and improve consumer experiences on its video-on-demand platform. To do this, they needed to make content more discoverable, by collecting metadata from thousands of individual video assets and then mapping each video to a unique, industry-standard identifier.
This code would help consumers discover and search for content on demand, but that was only the start of the journey. By tracking and analysing this data, the broadcaster would be able to gain valuable insight into customer journeys. With better insight into what customers searched for or watched, and rates of customer loyalty and churn, our client would be able to improve the consumer experience and deliver better services in future.
Getting a handle on large data volumes
The first challenge our customer faced was the sheer volume of data. Capturing metadata for just two years of video meant analysing more than 121 TB of data. It was important that the process of collecting metadata, mapping to new codes, and integrating this information with the customers’ analytics and reporting tools was as seamless and fast as possible.
Without data expertise in-house, our client was struggling to optimise these processes. Equal Experts developed a new data platform for the client in 12 weeks, that was capable of processing data queries in under 5 minutes, costing less than $5,000 per month.
Solution
Delivering the data health check
As part of our work with this customer, Equal Experts Data Studio carried out a Data Health Check, which identified a number of challenges that were slowing down the upfront collection, mapping and analysis of video data.
A data health check is an independent assessment of a client’s data processes, that is usually carried out over a period of two weeks. Each check is unique and tailored to the customer’s particular challenges and goals. An expert team examines all aspects of the data journey to identify issues around data quality, performance and efficiency. At the end of the health check, clients will receive a checklist that includes recommended improvements and advice to help the client improve their data processes.
Our data health check identified a number of challenges with the media company’s data processes relating to this project. In particular, there were inconsistencies around how data was imported into the system, leading to less reliable analytics data. Some data was imported to the system manually in CSV files while other data was pulled from a database in a materialised view.
The issue was compounded because data pulled from the database lagged, leading to a 24-hour delay in some cases and creating inconsistencies between different data sets. In addition, our customer used three different tools for data analytics and reporting, which meant that some reports needed to be manually created by pulling data from multiple platforms, as well as increasing the risk of data inconsistencies and duplication.
Results
Recommendations
As part of the data health check, we provided recommendations to the customer on how to fix or mitigate these issues. First, we advised the organisation to switch to a fully automated data import process. This immediately improved consistency and reduced the data lag from 24 hours to just under 2 hours.
We also advised the client to standardise on the Metabase reporting tool, which combined:
- Native SQL report creation
- Easy to use dashboards with the ability to embed dashboards on external sites
- Slack and email notifications
- Good user management
- Anomaly detection
Making this change further improved the accuracy and reliability of analytics data.
Our client recognised the data health check would be most effective as more than a one-off check. One year after our first report, the EE team conducted a follow-up Data Health Check to help the customer understand the potential for further improvement in its data processes.
Once again, our team performed a comprehensive check of the customer’s data systems, processes and reporting. We identified two key areas where additional improvements could be made.
Data Pipeline
First, we advised the customer that thinking about data as pipelines would help to deliver faster, more accurate data analytics. Previously, data imports were being managed within a single Clojure project. This offered limited flexibility to change schedules or create new data ingestions.
Adopting a data pipeline approach allows for data from multiple sources to be automatically moved to a single destination as required. This would allow the organisation to schedule, re-schedule and manually trigger data ingestion, making the process more efficient. In addition, a pipeline would give the customer an overview of the status of current and completed data ingestions.
Business logic
Second, our data health check revealed that the application of business logic in data analytics was slow and complex to change. To improve this process, we recommended using a tool called DBT, which allows organisations to rewrite business logic and then create or view associated documentation in a user-friendly interface.
This new approach allowed our customer to:
- Centralise the analytics business logic
- Enrich data transformation with testing
- Create and view documentation for data
- Improve data processing times from 2 hours to 10 minutes
As a result of the changed made through these Data Health Checks, our customer has achieved savings of $3 million, with a 20% increase in data quality, using optimised insights.
The result is that the company is able to be more competitive, offering a positive experience to consumers.