Are you providing a generic customer experience? See how Snowflake can help.
Written by: Brendan Tisseur, Robert Guglietti
Problem: Organizations globally lack an understanding of their fan/ customer which makes it difficult for them to tailor customer experiences based on what customers want. Sports leagues are no different. They want to engage, acquire, retain, and grow their fan base which requires a robust customer 360.
Unfortunately, only 14% of organizations have achieved a 360-degree view of the customer, while 82% said they aspire to get there, according to Gartner, Inc.
The impact to companies is an inefficient use of sales and marketing budgets as well as a poor and inconsistent fan/ customer experience that results in higher fan/ customer acquisition costs, lower return on ad spend, and increased churn.
Solution: Snowflake recently worked with one of the largest sports leagues in the world to leverage data to build a fan/ customer 360 (see Figure 1).
This was leveraged to optimize their sales and marketing efforts by predicting the likelihood of a sports fan to attend an event (lead score — see Figure 2) through the appropriate communication medium (email, phone, no action required etc. — see Figure 3)
Workloads:
Such a complex use-case requires a platform that can help simplify the way you understand your fans/ customers. You need a platform that breaks down your data silos and offers near unlimited scalability for your first party data while unlocking access and options for second and third-party data. This platform should bring the workloads to the data so you can run seamless pipelines between the different workloads that power your understanding of your fans/ customers giving you the competitive advantage.
Data Engineering (Fan/ Customer 360 — Foundation Layer)
Snowflake Differentiation: Creating the data foundation layer is the first and most important step to intelligently understanding our fans/ customers. We need to bring in fan/ customer data from a wide variety of different source systems like Sales, CRM, Social, Marketing etc. This data needs to go through several transformations and pipelines to ensure quality and consistency. From here we make available in a secure and governed way a consistent 360 view of our fans/ customers to the data science teams to help power AI and ML models like lead scoring and next best action. The same copy of data is also made available to the data apps and operational reporting teams for deeper insights and action.
How did Data Engineering support this use case:
Snowflake took the base understanding of the fan/ customer (name, address, age etc.) and augmented it with over 17,000 demographic, psychographic, and behavioral characteristics from Neustar to develop the foundational layer for this use case.
From here we could also see an operational view of the customer operations analyzing sales, customer count, and other touch points throughout the fan/ customer journey. We could also get down to the fan/ customer level of detail to see their lifetime value and activity.
Collaboration
Snowflake Differentiation: You can discover data, services and apps from 310+ providers across 18+ categories (as of July 31, 2022) to power your most critical data demanding workloads (see Figure 6). Snowflake enables you to access the most current data available from providers while reducing data integration costs with direct access to live, ready-to-query data with no ETL required.
How did Collaboration support use case:
Third party data from Neustar (demographics) and ShareThis (behavioural) were leveraged directly from the Snowflake Marketplace to augment our understanding of the fan/ customer.
Neustar’s “ElementOne (E1) Market Analytics and Segmentation” solution helps marketers identify their most valuable audiences with unparalleled precision, understand their current and potential customers, and ultimately sharpen their marketing strategies for improved ROI.
With ElementOne, you have on-demand access to a robust dataset of over 17,000 consumer demographic, psychographic, and behavioral attributes, sourced from multiple syndicated research providers. It includes attributes describing: Demographics, Psychographics, Geography, Attitudes, Needs, Purchase Behaviors, Buying Preferences, and Media & Channel Preferences.
Additionally, Neustar’s authoritative identity methodology enables you to integrate any of your internal data against this vast consumer dataset, to build an accurate 360° view of your market. The result is a predictive data set aligned to your unique business.
ShareThis represents another data asset adding value to Snowflake initiatives. ShareThis data is powered by consumer behaviour on more than three million global domains, representing 40 billion events per month. ShareThis leverages AI and advanced decision science to synthesize and analyze context, keywords, concepts, geolocation and more, providing a comprehensive, global dataset, offered as fast as real-time.
ShareThis online behavioural data is uniquely suited for enriching 1st party data, by supplementing purchase data with both interest and intent data, derived from online behaviours. The data enables rich insights which can be used for initiatives including, but not limited to, customer segmentation, personalization, and to predict future online activity. The data asset impacts every aspect of marketing, offering solutions influencing planning and insights, targeting and activation, as well as measurement.
Data Science & ML
Snowflake Differentiation: Snowflake gives data scientists the choice of framework, language and tool. Leveraging the Snowflake platform, data scientists can run scalable and secure ML inference with models running inside Snowflake as UDFs. Model results can be made available in Snowflake for other workloads and teams to take advantage of and act on ML-driven insights, all on the same copy of data.
How did Data Science & ML support this use case:
Leveraging the foundational understanding of the fan/ customer enables us to provide the input needed for our data scientists to leverage Snowpark to determine the lead score and next best action. Understanding which fan/ customer are likely to attend upcoming events, allows targeted outreach to fill seats at events and increase new fan/ customer attendance.
The model uses fan/ customer activity — including purchases, website and system activity, and outreach -, Neustar demographic data from the Snowflake Marketplace, fan/ customer history at each event, and venue demographics. The model is deployed in Snowflake and scheduled to give a daily updated score for each fan/ customer.
Our data scientists leveraged SHAP values (see Figure 7) to get an indication of how features are contributing to the model prediction. SHAP values are at the individual level but can be grouped across multiple fans/ customers. A positive SHAP value indicates that a feature value is increasing the likelihood prediction of attending an event, while a negative SHAP value indicates that a feature value is decreasing the likelihood prediction of attending an event.
Enterprise Analytics — Insight to Action
Snowflake Differentiation: Business intelligence (BI) tools enable analyzing, discovering, and reporting on data to help executives and managers make more informed business decisions. A key component of any BI tool is the ability to deliver data visualization through dashboards, charts, and other graphical output. Snowflake supports a wide range of BI tools through generic or native connectors.
How did Enterprise Analytics support this use case:
The data engineering and data science output was surfaced in a consumable format for our sales and marketing stakeholders. They were able to see high level KPIs on the revenue from each event, average order volume, and drill down to the individual fan/ customer level of detail. Figure 8 shows an example of the information available on a fan/ customer where you can see the lead score that would be ranked across the entire fanbase and then the next best action to determine how our sales and marketing teams should communicate with the fan/ customer.
Applications
Snowflake Differentiation: Leveraging Snowflakes native connectors it’s easy to power applications like Salesforce to take advantage of and operationalize advanced analytics and ML models and near-unlimited scalability and concurrency.
How did Applications support this use case:
Leveraging Streamlit we are able to simulate a CRM system and showcase how Sales and marketing teams could leverage Snowflake powered data apps as another medium for surfacing the output from our data engineer and data scientists. Figure 9 shows this fan/ customer view that could be embedded inside your CRM to inform sales team members on which fans/ customers should be contacted by phone or email for each event.
Summary:
Figure 10 showcases that the foundation of understanding your fan/ customer is imperative to opening up several other value added uses with this data. We started by establishing the fan/ customer 360 which enabled us to determine the lead score and next best action.
As a next step this sports league could look at leveraging a composable CDP (as seen in Figure 11) for the cross channel advertising activation where organizations can engage with fans/ customers across multiple channels in a consistent manner. In addition, we could then look at cross channel marketing and ad measurement which looks to evaluate each marketing touchpoint across all channels and revealing which combinations of media perform best. All of this could also be done in a privacy compliant manner with the data clean room.
Join the One platform that powers the data cloud to execute your most critical workloads on top of Snowflake’s multi-cluster shared data architecture in a fully managed platform that capitalizes on the near-infinite resources of the cloud simplifying your ability to execute on complex use-cases like optimizing customer experiences.
Solution designed and developed by team frostbyte (Kaila Chen, Marie Coolsaet, Swathi Jasti, Jacob Kranzler, Shriya Rai, Adithya Nanduri, Alex Woodcock, Vernon Tan, Brendan Tisseur, and Robert Guglietti)
