How to transform GCP BQ into The Perfect CDP for Retailers

What are the Retailer’s needs?

B2C and D2C eCommerce Retailers are looking to transform their business using a Customer Data Platform (CDP) in three main ways:

1. Customer acquisition and upsell

To create a real-time Customer profile by joining Customer data from their onsite and offsite channels, and using this to power seamless omnichannel marketing, product recommendations, and onsite personalization campaigns.

2. Journey Optimization

To leverage the rich Customer data for Journey Analytics, creating actionable insights for improving Customer experience onsite.

3. Machine Learning and activation

To mine the Customer data to identify data patterns common across successful/desired outcomes, creating advanced segmentation using machine learning, and activating these segments in rapid A/B testing to identify most promising “topical” strategies.

What is Google Cloud BigQuery’s sweet spot?

Keeping the cloud transformation conversation aside for now (why and how a retail business should migrate to the GCP cloud), here are the TWO most significant reasons why Google Cloud Platform BigQuery (GCP BQ) should be the engine to power the business-critical function of a CDP:

1. Google BQ is one of the most successful Data warehouse applications in the market, thanks to its ease of use and cost-efficiency 
2. Google BQ has native data connectivity to Google’s Ads ecosystem (search, display, video …) that most retailers use to reach their end-Customers when they’re not on their properties. This makes GCP BQ a de-facto choice when it comes to storing omnichannel interaction data in one place.

How does Syntasa transform BQ into a CDP powerhouse?

Syntasa is bolt-on software application for BQ, that can be downloaded from Google Cloud Platform Marketplace by GCP Customers, to transform their own GCP BigQuery into a CDP engine.

The five steps in which Syntasa transforms Google Big Query into a CDP engine giving maximum control to the business in achieving its goals are:

1. Data Ingestion

Almost all Digital data sources (web analytics, marketing clouds, ecommerce platforms, email systems, enterprise systems, and so on) can land their raw data into GCP Cloud Storage without requiring additional effort.

Syntasa provides data connectors for the popular Digital systems that not only can extract and load the data into BQ inside neatly structured data model, but also has intelligence built into it to cleanse the data, fire alerts when files go missing, trace the lineage of the data, produce DQ reports, manage complex jobs and task dependencies.

2. ID Unification

Once all the source data start flowing into BQ, Syntasa ID Graph module plugs on top of BQ to unify the customer data under one unique identifier, including first party data like analytics cookie, GDPR consent, encrypted email address and phone numbers, orders, and products purchased, multiple device IDs, and merges/throws out any duplicates.

Syntasa stores the resulting cleansed and merged data back into BQ and Big Table, where it can be utilized by the rest of the organisation without worrying about any of the complex plumbing. 

3. Segmentation

Various native GCP data science and ML tools can be used to mine cleansed and matched data in BQ, like AutoML TensorFlow.

Syntasa provides a low-code, no-code add-on for Customer Segmentation, that includes templates for many of the common ML challenges like an ML model template to find out customer’s propensity to churn, or affinity to a brand or category, in-market, high ARPU, and more. These templates take away most of the burden of plumbing and productionisation, leaving valuable time for finessing model hyper params.

4. Journey Analytics

Google Cloud has native general-purpose tools to create dashboards and visualization on top of the data curated in BQ by Syntasa.

Syntasa provides a more advanced and special-purpose Customer Journey interrogation dashboard that plugs into BQ dataset and the Syntasa ID Graph.

5. Activation

Syntasa makes Google Big Query the central activation and orchestration hub.

Syntasa activation connectors for many major Marketing, Onsite Personalisation, eCommerce and enterprise systems, can be deployed in drag and drop manner on top of BQ. Syntasa software manages pipelines and data lineage, and keeps the connectors up to date with any necessary changes..

Google Big Query – A Data Model for the CDP

Figure: a simplified visualisation of the most common digital IDs across digital sales and
marketing channels, and offsite systems for transactions, CRM, and customer care

Traditionally, the Customer Data management challenge in the Digital world was primarily about mapping web analytics with the AdTech eco-system so that individuals could be tracked and targeted across the Internet.

With the tightening of regulations and ethics around cross-site tracking and targeting, the focus of the Customer Data management needed to shift more towards curating an enterprise-grade first-party identity mapping across online and offline datasets.

ID mapping is only half the challenge. After all, IDs are complex time-series and not just simple mapping between the two systems. Without the focus on time-series aspect of an ID, a mapping will only provide irrelevant results and poor match rates.