Reverse ETLs have quickly become a popular tool for marketing, mostly because everyone is starting to take our advice. But really, centralizing your zero- and first-party data is a must in today’s modern data stack, and the warehouse-first architecture is quickly becoming the favored approach for data-driven organizations. So much so that Customer Data Platforms (CDPs) are starting to shift their focus to becoming Customer Data Infrastructure (CDIs) instead.
What is Reverse ETL?
In a nutshell, a Reverse ETL (Reverse Extract, Transform, and Load) is a data-integration tool to get customer data out of your data warehouse (e.g. Snowflake, BigQuery, Redshift) and copied into your favorite SaaS platform(s). When you combine a CDI and a Reverse ETL, you effectively have the building blocks for a homegrown CDP, but with more security and a reduced cost.
For marketers, think of Reverse ETL as the evolution of Unica or SAS into the cloud. You can see how we designed MessageGears Segment to replace your legacy segmentation software to enhance the modern data stack.
While Reverse ETLs aren’t limited to just marketing, sales and support are also great use cases, having a Reverse ETL capability and process will be crucial for companies leveraging the modern data stack. With the decline of third-party data as native advertising starts turning to the “Conversion APIs” model, using a Reverse ETL makes it very easy to get first-party data to advertising channels.
Limitations of Reverse ETL
We recently read a solution story where a Reverse ETL helped solve a customer engagement platform data-sync problem. In short, a marketing team had trouble syncing all 5 million of their customer records from Snowflake to Iterable every day, so they turned to RudderStack’s Reverse ETL functionality to help them batch their data to Iterable instead. While we applaud our friends at RudderStack for finding a solution to Iterable’s constraints, we still have some general concerns with the overall model.
Syncing data will always be a pain
While this is how most SaaS providers work, we don’t think it should be, especially for large brands and enterprise companies. Being forced to sync data from Snowflake to your SaaS marketing cloud provider means the following:
Limited Data Observability
The health and state of your customer data is invisible to you in an external platform.
Reduced Data Security
You introduce your customer data to further risk by letting providers keep it.
Duplicated Data Storage
You’re paying more for your provider to store your duplicated data.
Not to mention, it takes a lot of time and resources to manage SaaS application APIs, especially when you’re using different providers and aren’t taking advantage of a single cross-channel solution. While Reverse ETLs might make this easier for many, it will still struggle to perform for companies that have millions of customers and behavioral events, especially when you’re trying to sync from your data warehouse to a platform that limits the number of API calls.
Reverse ETLs are technical
Reverse ETLs solve a data problem, not a marketing one.
So far, they’ve been purpose-built for data and analytics engineers that are working within or beside marketing and sales teams. While some Reverse ETL providers are working to provide a low/no-code solution for non-technical users, they will need more time to mature. You’re not going to be able to create customer journeys, complex segmentation, or add additional labeling and grouping for personas.
They also may not supply you with what you need out of the box. Looking at this specific case, while RudderStack has Iterable as a destination, the API it used was causing the problems in the first place. Their final solution used a custom webhook integration that required Javascript transformations to hook into a different Iterable API. While many marketing teams have experts in HTML and CSS, coding Javascript will be a stretch for most teams.
The provider bottleneck
At the end of the day, even if you have a unicorn high-performing, no-code Reverse ETL, it’s only as good as the provider you’re syncing to. With each provider comes their limitations, from the number of API calls you can make per day to how the data should be transformed to fit the provider’s data model. Forcing your customer data model into something different for marketing campaigns is like trying to cut an orange with an apple slicer. It can work, but it’s really messy.
When to use Reverse ETL
Say you need to get your data to social channels (e.g. Facebook, TikTok), search (e.g. Google, Bing), support (e.g. Zendesk), or sales (e.g. SalesLoft), then you definitely need a Reverse ETL capability and process. A Reverse ETL should not be used for enterprise cross-channel messaging — it’s simply not sustainable.
At MessageGears, we believe that you shouldn’t be limited in how you use your customer data. We don’t believe in data syncs and understand that your customer data model has been built out to be what is best for your business — we’ll never try to change it. You’ve made investments in centralizing your data in data platforms like Snowflake, BigQuery, Databricks, or Redshift, so why not try using a marketing platform and an ecosystem that understands that?
MessageGears Message is a cross-channel marketing platform that connects directly to your data warehouse as a “connected application,” with no ETL (reverse or otherwise) necessary! MessageGears Segment is a Reverse ETL purpose-built for non-technical marketers to segment and activate your customer data to social and search channels. Interested in getting started? Let’s chat.