AWS Partner Network (APN) Blog
Building a Complete Customer 360 in Amazon Redshift with RudderStack
By Brooks Patterson and Eric Dodds, Product Marketing – RudderStack
Kiran Singh, Sr. Partner Solutions Architect, Databases – AWS
Sai Prakhya Tata, Solutions Architect, Customer Experience – AWS
Harmeet Nandrey, Sr. Partner Solutions Architect, MarTech – AWS
 |
| RudderStack |
 |
In the age of the attention economy, brands that can’t master personalized customer experiences struggle to stand out. Brands that deliver personalized end-to-end experiences in real-time create a hard-to-copy competitive advantage.
However, mastering personalization today is no small task. It requires sophisticated applications of customer data across business units and systems. At its core, personalization is a data problem, and it’s one that siloed tools and fragmented data systems can’t solve because it requires a complete customer view.
A complete customer view—referred to as Customer 360—is a comprehensive, unified view of every data point from every customer. It enables teams and tools to work from a single source of truth with rich, up-to-date information about every customer. It also enables brands to build strong relationships with their customers at scale.
Building a Customer 360 view involves collecting data from every customer touchpoint, unifying the data, and modeling it into activation-ready profiles. Historically, this low-level, resource-intensive work has kept Customer 360 out of reach for all but the most advanced companies, but Amazon Web Services (AWS) and RudderStack are changing that.
With RudderStack Profiles and Amazon Redshift, every data team can now build a Customer 360 directly in their data warehouse without the complex query and modeling work.
RudderStack is an AWS Specialization Partner and AWS Marketplace Seller with the Amazon Redshift service ready designation. It was founded upon these convictions: that the data warehouse should be the foundation of the customer data platform (CDP), and that the CDP should be owned by the data team.
Power of Customer 360
A Customer 360 is a comprehensive, unified view of every data point from every customer that can be analyzed and distributed to every team in the business for activation. It typically takes the form of a table in the data warehouse with one row per customer and columns for every customer attribute.
An actionable Customer 360 helps data teams create a tighter feedback loop between data and decisions. It helps the data team get closer to the business and the business get closer to the data. Therefore, it can serve as a mechanism that shifts data teams from primarily reactive tasks—such as answering data requests and closing tickets—to more proactive work.
A Customer 360 enables data teams to leverage their creativity, technical expertise, and knowledge of the organization’s data to deliver unexpected value.
Figure 1 – Customer profile in Amazon Redshift.
The use cases for Customer 360 impact every part of the organization:
- Analytics: Robust customer profiles enable analysts to build detailed user journeys to answer how and why things happened, not just what happened.
- Advanced marketing personalization: With every available relevant first, second, and third-party data point unified and exposed through a Customer 360, marketing teams can go beyond basic templated personalization to automate delightful customer engagement experiences.
- Advanced segmentation and audience building: Customer 360 data gives marketers an edge when it comes to targeting. Rich customer profiles enable teams to create audiences based on sophisticated segmentation, enabling better targeting and more cost-efficient advertising.
- ML recommendation engines: With ready-to-use features from a single Customer 360 warehouse table, machine learning (ML) teams can pull data directly into their models for product and content recommender systems.
- Churn prediction: Customer 360 can unlock sophisticated ML-based churn prediction methods that help customer success and marketing teams proactively address red flags and retain customers.
- Account expansion: Customer signals that indicate expansion potential can go unnoticed without a Customer 360. It can deliver relevant product usage data, lifecycle marketing data, and website behavioral data directly to tools which sales and customer success teams use every day to provide timely alerts around expansion opportunities.
Technical Challenges of Customer 360
To understand why it’s hard to build a 360-degree view of the customer, let’s consider the whole data activation lifecycle. For a business to realize the value of a Customer 360, data must be collected, unified, and activated. Each of these stages involves its own unique challenges, but it’s the unification process that most often thwarts customer 360 projects.
Next, we’ll overview each step below with a special focus on unification.
Data Collection and Centralization
Data collection is the first hurdle to creating a successful Customer 360, which should include first-party user behavioral data and first-party data from your software-as-a-service (SaaS) tools and internal systems. This data will span various sites, apps, and systems that contain data from touchpoints throughout the entire customer lifecycle. Each of these data sources requires its own ingestion mechanism.
On top of that, a Customer 360 will, ideally, be enriched with relevant second and third-party data, such as data from ad platforms or purchased demographic and firmographic data. AWS Data Exchange makes it easy to find third-party sets to include in your Customer 360. Once collected, this data must be centralized in Amazon Redshift to prepare for the next stage.
Data Unification
Data unification is the most complex part of the data activation lifecycle. It’s also the most important because proper unification delivers the foundation of exceptional data activation—complete customer profiles.
This stage includes the identity resolution process of stitching every data point for each customer together under a single canonical identifier, building an identity graph, and building user features (attributes).
Data unification requires data modeling, managing user identities, computing semantic features, and keeping up with metadata for those features. This typically requires writing and maintaining excessive amounts of SQL and quickly gets untenable.
While these are all solvable problems, they require laborious work that requires so much bandwidth from the data team that it can ruin their ability to proactively work with business teams on the data initiatives that move the needle.
Data Activation
Once you’ve managed to collect and unify your data into a Customer 360, it’s time to put it to work. Data activation uses your data to create value across the business, but the technical challenge of data activation lies in democratizing access.
The value of your customer data is essentially limitless once it’s in the right hands. Getting it into the right hands requires piping it out of the warehouse and into the systems and tools used by different teams.
RudderStack’s Reverse ETL pipeline enables you to easily serve your customer data to the whole company so every team can harness its power.
How RudderStack Profiles Solves Customer 360 Challenges
Combining Amazon Redshift with RudderStack enables you to take full advantage of Redshift’s highly scalable, highly performant data warehouse service and RudderStack’s end-to-end customer data platform to effectively eliminate the engineering challenges of building and maintaining a customer 360.
RudderStack Profiles is a data unification tool that effectively eliminates the low-level work required for data unification to produce a customer 360, freeing data teams up to partner with business teams on initiatives that deliver unexpected value.
Profiles makes it easy for you to build customer 360 projects natively in Amazon Redshift. Once configured, Profiles leverages the power of RudderStack’s end-to-end platform to produce a customer 360 table and activate the complete customer profiles in over 200 business tools.
Accelerate Time to Value with Out-of-the-Box Assets
When you ingest user data through RudderStack’s Event Stream pipeline into Amazon Redshift, the schemas, unique identifiers, and table relationships are already known. That allows RudderStack Profiles to automatically generate an identity graph and over 30 event-based user features out of the box.
You simply need to specify the data sources, run a Profiles job using our Customer 360 template, and traits like days_since_last_seen, first_campaign_name, and active_days_in_last_7_days will immediately be available in Amazon Redshift. These enriched data points can speed up analytics projects, and you can sync them to business tools through RudderStack’s reverse ETL pipeline.
Figure 2 – Customer Profile generated by RudderStack Profiles.
Configure Identity Graphs for Any Entity
If your business has multi-faceted identity resolution requirements, you don’t just need to reconcile disparate user identifiers across data sources. You also need to model relationships between various customer or object entities using more than just event data. Your individual users might belong to multiple business accounts, separate subscribers might belong to a single household, or multiple connected devices may belong to one customer.
Modeling these relationships with hundreds of unique identifiers at scale often outstrips the capability of no-code tools. Profiles has no such limitations. In addition to a robust interface, it empowers you to customize identity models through RudderStack’s code-based Profile Builder that fits within your existing version-controlled development workflows.
Figure 3 – RudderStack Profiles code-based Profile Builder.
Go Beyond Identity Resolution
Once you create your identity graph, you need to complete the picture to build a true 360-degree view of the customer. You can do this by combining known traits (like userId and email) with computed traits, like the number of days a user has been active and the average dollar amount of items in their most recent cart.
RudderStack Profiles eliminates the need to use complex SQL to generate these data points. Instead, you can write simple metrics definitions, and Profiles automatically computes the values natively in your warehouse. You can then add them to the Customer 360 table, where they’re ready for analytics and activation in business tools.
Figure 4 – Customer 360 table in Redshift with features generated by RudderStack Profiles.
See the Full Customer Journey with Funnels and Semantic Events
Customer journeys happen sequentially over time, and they’re defined both by what users do and what they don’t do. Building funnels in SQL is notoriously difficult, as is modeling semantic events like abandoned carts.
RudderStack’s templates automatically generate common semantic events and funnels from RudderStack event data, immediately delivering analytics-ready data in Amazon Redshift. For more complex use cases, you can model the full customer journey using simple definitions and human-readable operators like “before” and “after” in version-controlled config files.
Automate Historical Snapshots and Metadata Management
Maintaining a historical record of how data points change over time is one of the hardest parts of maintaining an identity graph and a Customer 360 table. This metadata is important for historical customer journey analysis, and it’s a crucial component of the machine learning models used for churn prediction and recommendations.
Profiles automates this metadata management entirely. It creates a snapshot table in Redshift every time a Profiles job is run, giving you full visibility into the time series changes of your customer profiles.
Conclusion
A Customer 360 approach enables brands to master personalization, leveraging their customer data to build strong relationships with customers at scale. It’s also a mechanism that helps data teams position themselves as strategic partners with every business unit. While historically out of reach for all but the most advanced companies, Amazon Redshift and RudderStack make it easy for every team to deliver an actionable Customer 360.
With RudderStack Profiles, data teams can take full advantage of RudderStack’s end-to-end platform alongside Amazon Redshift, enabling them to build a Customer 360 directly in their warehouse environment.
Request a demo to learn more about RudderStack Profiles and see it in action. You can also learn more about RudderStack in AWS Marketplace.
.
.
RudderStack – AWS Partner Spotlight
RudderStack is an AWS Specialization Partner and warehouse-native customer data platform that helps companies to get value from their warehouse or data lake.