What Is Single Customer View?
Companies across industries aspire to understand their customers in data-driven world. Single Customer View (SCV) or 360-degree customer view is a powerful data engineering concept that lets companies combine consumer data from several sources. Integrating data points into a unified picture lets organizations gain insights, personalize experiences, and make data-driven decisions.
Telegraph Media Group
TMG’s free-to-air, advertising-based income model grew increasingly difficult. Like many news media publishers, Google cloud noticed long-term factors like declining print readership, content publisher ad yields, and ad revenue volatility, which made revenue estimates and growth unpredictable.
Google cloud was launched an ambitious plan in 2018 to become a subscriber-first business with great journalism to increase our subscriber relationships. TMG adopted a subscription model to increase income predictability and advertising returns. In August 2023, Google Cloud reached one million subscriptions, meeting their five-year objective.
They use TMG’s subscription data and a customer’s digital behavior across all domains in Google’s Single Customer View platform. It also uses data from partner retail websites and engagement goods like fantasy football and puzzles. These varied data sources help the company understand its consumers and provide a complete news experience.
The SCV building process can be envisioned in several stages:
Data collection: Data is collected from multiple sources and loaded into Big Query, a data lake. For centralized storage and processing, BigQuery imports data from data bases, APIs, and files.
Data transformation: Business rules are applied to BigQuery data in this stage. Data is cleansed, standardized, and supplemented for quality and consistency. In a new BigQuery dataset, the Single Customer View data model organizes the data for easy access and analysis.
Data presentation: The BigQuery data lake may organize data into data marts after transformation. Specific user groups or departments receive specialized and pre-aggregated data from these data marts for usage by third-party activation tools like email marketing systems and reporting and visualization tools for internal decision-making.
Initial data collecting
Sales force stores all TMG subscription data. They optimized data collection and BigQuery storage.
- First, They execute a pipeline of containerized Python apps on Apache Airflow (Cloud Composer) every minute. Pub/Sub, a Google Cloud messaging service, receives updated Sales force API data from this pipeline.
- Second, They used Data Flow to develop a real-time pipeline that reads Pub/Sub data and updates Big Query tables for real-time data insights.
- Data integrity is maintained by daily Sales force batch ingestion. This method gives us a complete picture of the data, correcting for real-time ingestion data loss.
Ingestion of Adobe Analytics, which records user behavior on TMG websites and applications, and Adobe Campaign, which tracks user behavior on communication channels, is identical. Since these data sets do not require real-time availability, batch processing is suitable for import and processing. They also use similar ingesting methods for additional data sources to create a cohesive pipeline.
Stage 2: Data transformation
They use BigQuery and SQL to transform Google Cloud data via the open-source Data Build Tool (DBT). DBT converts all business rules into SQL for effective data translation. Cloud Composer, based on Apache Airflow, orchestrates hourly DBT workflows. These data pipelines produce a BigQuery relational model with streamlined and organized data for analysis and processing.
They use many key pathways for data transformation:
Sales force Snapshot: This pipeline snapshots real-time and batch sales force data. The snapshot provides the newest Sales force data to other pipelines in the transformation process.
Source: This pipeline provides a table for source data, including original and new customer IDs. This data is vital for identifying clients in the original data source.
Customer: This pipeline builds a table using precise information to show their qualities and characteristics.
Contact: This pipeline builds several customer contact tables.
Content Interaction: This pipeline creates a table of customer digital behaviour, including content interactions, for further customer engagement and preference analysis.
Subscription Interaction: This pipeline provides a table to collect and store subscription-related events and their information, revealing consumer subscription habits.
Campaign Interaction: This pipeline provides a table with extensive communication behavior data from different channels to analyze customer interaction with campaigns and marketing efforts.
Stage 3: Data display
To translate Single Customer View data model data into data marts, DBT pipelines are essential, like the transformation layer. Data marts are useful for analysis and third-party applications. Current SCV data consumers include Adobe Campaign, Adobe Experience Platform, and Primitive.
Single Customer View data helps Adobe Campaign send relevant campaigns and personalized offers to customers. The rich customer insights from this data help Adobe Campaign optimize customer interaction and target marketing.
Adobe Experience Platform uses Single Customer View data to personalize website experiences. The Adobe experience platform uses extensive consumer data to customize the website experience to specific user preferences, improving customer satisfaction and engagement.
Primitive targets Telegraph website and app users with targeted ads using Single Customer View demographic data. Primitive segments customers and uses Google Ad Manager to personalize advertisements.
Before the Single Customer View, these consumers used data from multiple sources, which was often days old. Their ability to target clients numerous times a day was limited by this delay. The SCV gives them immediate access to near-real-time data, allowing them to use it every 30 minutes. This huge data freshness enhancement allows TMG to deliver more timely and relevant experiences to our target audiences.
Implementing a Single Customer View (SCV) presents hurdles, including building a data model that fits current and future demands. Google cloud prioritize cautious data model modification to accommodate new requirements within the framework whenever possible. Choosing the right Single Customer View data is also crucial. Businesses may want to incorporate all customer data, but they eliminate noise and include just important and valuable data to keep SCVs clean.
Managing consumer communication choices is another major concern. TMG is authorized to interact with SCV customers based on their preferences. These preferences can vary between channels, including third-party platforms, and clash with our first-party data. To ensure compliance and minimize legal risks, use hierarchical rules to set permissions for each communication channel.
Unifying clients across numerous sources requires efficiently matching third-party data to TMG’s first-party data. They use precise and fuzzy matching to solve this problem. BigQuery User-Defined Functions (UDFs) provide fuzzy matching with various techniques. Processing fuzzy matching on huge data sets takes time. They are investigating ways to balance accuracy and processing time to improve matching and customer data integration.
Conclusion
In conclusion, adopting a SCV on Google Cloud helps TMG exploit consumer data to expand, improve customer satisfaction, and stay competitive. Companies may make data-driven decisions and provide personalized customer experiences by using SCV insights. Overcoming SCV building obstacles lets firms maximize data potential and generate significant results.
