Warehouse-native Analytics for Growth Marketing
Following the earlier blog “Unveiling the Truth Behind Warehouse-native Product Analytics”, let’s cover how the growth marketing team for digital products can effectively explore customer 360 with integrity with this new approach in analytics.
What is Growth Marketing?
There are many definitions out there. We’d like to think of Growth Marketing as an approach to attract, engage and retain customers through campaigns and experiments focusing on the everchanging motives and preferences of their customers. In practice, growth marketers build, deliver and optimize highly tailored and individualized messaging aligned with their customers needs through multiple channels. They are a cross functional team between Product, Marketing, Customer and Data. Product analytics play a significant role for this job with the focus on self service customer insights.
From customers’ lifecycle perspective, there can be several stages:
Acquisition
At the top of the funnel, customer acquisition is all about the strategy to target potential customers with tailored content through multiple channels with highest efficiency and fastest but most accurate measurements. The campaigns can be executed as ads, paid search, call-to-actions, free offers or discount coupons on various third-party channels. Often, there is a significant amount of budget allocated with these campaigns, which are also super dynamic.
Often the term “attribution” is used, which means to attribute every customer to the proper channel they come from in order to measure and find the most effective one. It requires constant monitoring, A/B testing and tuning to optimize acquisition channels on the fly in order to adapt to the market dynamics and get the best ROI.
Engagement
Once a new customer comes in, the focus now is to drive their engagement and collect more data to help build better experiences by enhancing their customer journey. Typically there are critical stages or funnels for a digital product like onboarding (tutorial), sign up, engaging with the core loop (e.g. watch a video, invite a friend, add to cart), and checkout. The goal of the engagement is to prompt customers with the most relevant and attractive content and push them through the desired sequence in order to keep them in the application.
Besides optimizing the user flow by improving design and usability, growth marketers typically rely on incentivized offers (first order discount, free trial), social/viral loop (invite a friend, refer someone) and loyalty programs to keep their customers engaging. All of these efforts require a deep understanding of customers’ journey (e.g. funnel, conversion, drop off) through product analytics in order to make the right decisions.
Reactivation
There will always be customers who become dormant or churn completely. In order to get them back into the application and retain, growth marketers utilize every possible communication channel at their disposal: email, push, SMS or even targeted ads to get their attention and get them back. Often some third party tools like Braze, a Customer Engagement Platform, will be utilized to deliver these messages. Though, product analytics will be the driver for these campaigns to identify different cohorts, target them and measure the ultimate results, which is not only about impression and open rate, but also the long term impact inside the application: e.g. retention, subscription attach, LTV (lifetime value).
Problems with Siloed Product Analytics
Those last generation product analytics platforms worked out for growth marketing needs at the time when they needed to run fast, but with some high cost.
Super Complex Data Flows
Since there are always different vendors for each stage of a customer’s journey, the product analytics must have a copy of precious customer data from each and every vendor in order to deliver the insights. That is the root cause of all the headaches and issues.
There are criss-cross connections to be established through various approaches (e.g. ETL, API and storage sharing) and conduct heavy duty data copying. More often to anyone’s like, these connections can be broken, require maintenance, or even worse need to restate the historical data because of mistakes made. Just imagine the impact on the critical campaigns which require almost real-time insights.
Identity Matching Becomes a Nightmare
When privacy concerns like GDPR arise, there are more and more limitations on what kind of customer identifiers can be shared with and between vendors themselves. Often the growth marketers get stuck in the middle of the battle between data engineering and security personnels. Eventually some aerobatic maneuvers have to be done on the data pipeline, which makes everything further more complicated and fragile.
No Single Source of Truth
This one is really simple: with copies of data lying outside of the enterprises, how can anyone trust the insights where the analytics platform is a blackbox and there is zero transparency to understand how the insights are generated. Needless to say, there is no reconcilability whatsoever. It would really take some vote for confidence to rely on these findings to make growth marketing decisions, which often involves millions of dollars of budget.
Limited View on Customer 360
For growth marketers, often just having the impression, conversion, CPI/CPM data is not enough at all. The deeper the insights into customer behavior, the better. For example, just measuring the open rate of a push campaign only scratched the surface, it is often desirable to understand what kind of content did the customer engage with, how long did they stay in the application, did they come back the week after, or if they converted to subscriber and if/when did they churn again.
In order to get this complete view of customer 360, operational data are required (e.g. Items and Subscriptions), but often it is almost impossible for traditional product analytics platforms to get these data because they are usually not part of the clickstream (behavior) data and will require very complicated ETL integration to send a copy out.
The Warehouse-native Way
All of these problems can be solved with the warehouse-native approach when the enterprise is committed to have full control of their data within a cloud data warehouse. By bringing all of the clickstream, campaign, impression, conversion data from the vendors together and making their own data warehouse the Single Source of Truth, new generation of warehouse-native analytics platforms can connect directly to the custom data model with effortless integration and ensuring both the integrity and self service perspectives required by growth marketers.
Simplest Data Integration
For the enterprise, they just need to collect their own customer data (including clickstream/behavior and operational data) and all vendors’ data into a central data warehouse which is under their full control. Often, access to vendors’ data can be achieved through Data Sharing protocols (available in most cloud data warehouses) instead of duplication with ETL or API.
There is no complex graph of data flow outside of the enterprise, especially between vendors. When there are data issues, only one place needs to be fixed and it is easily verifiable instead of coordinating with several third parties to pray that they will do the right thing since there is no visibility into their black boxes. There is no data backfilling, scrubbing or restating required.
Customizable Identify Resolution
Because all the data now goes to the enterprises where the customers are from, all available customer identifiers can be explicitly stated and used for analytics internally without the need for hashing and complicated matching (often guessing) algorithms.
Even better, enterprises can experiment different identity resolution or attribution strategies on the fly, without the need to engage with vendors or reprocess any data. The ability of asking and validating “what if” questions before commitment gives complete confidence and flexibility.
Moreover, sensitive identity data can also be hidden or dynamically masked for warehouse-native analytics platform’s access since they don’t need to see the individual data as long as the underlying join works.
Exploration with Integrity
With One Single Source of Truth, and the ability to provide the SQLs behind every insight, growth marketers can now explore customer insights in a self service manner while maintaining the highest level of integrity. The transparency delivered by warehouse-native analytics makes it complimentary to any other BI, AI or machine learning tools, where they can not only reconcile the insights but also build on top of them.
Full Customer 360 View
With all the data about customers’ complete lifecycle stored in one place. warehouse-native analytics can easily bring in any operational data (e.g. Items, Subscriptions or LTV) into the analyses. Best yet, because all insights are generated dynamically, there are no ETL jobs to develop, no data to backfill when new data is required. That means that growth marketers don’t have to wait weeks or months for some data model changes required for specific vendors. Live customer insights with thorough depth is not a dream any more.
Summary
The days of data silos are long gone. With the convenience and advantages, warehouse-native analytics for growth marketing is an undeniable trend for enterprises with customer focused digital products. Besides exploring customer 360, getting reliable, trustworthy insights from a Single Source of Truth should be on the top of the mind for every serious growth marketer.
