You may have heard about Snowflake, especially following the breaking news from Dreamforce. Salesforce Genie was declared the greatest Salesforce innovation in the company’s history, paving the way for highly personalized customer experiences, delivered in real-time.
Genie ingests and stores real-time data streams at massive scale, and combines it with Salesforce data, supported by Snowflake.
Snowflake, in simple terms, is a data lake – a place where huge amounts of data can be stored, and accessed when you need it. Imagine different data sets swimming around, available for you to “cast your net” (writing a query) to pull the specific data set you’re looking for.
Snowflake + Salesforce
Snowflake isn’t the only data lake out there. We’re spoiled for choice with major players such as Google BigQuery, Amazon Redshift, Oracle Database, among others. As a Salesforce customer, you should seriously be considering Snowflake for its growing alignment with the Salesforce platform.
Snowflake works with the Salesforce platform (Customer 360) in the following ways:
- Salesforce Genie (Salesforce CDP): With big-bang Salesforce Genie announcement, Snowflake provides secure real-time and open data sharing between the Snowflake data lake and Salesforce (and vice-versa).
- Marketing Cloud: A sensible bolt-on for Salesforce Marketing Cloud, where Marketing Cloud stores campaign data you use frequently, and the data lake stores contextual data from platforms that are not traditionally considered sources of data that can inform customer journey behavior.
- Tableau: A good match to enable someone to visualize data, to bring insights out of the depth of the data lake.
- Einstein Discovery: Run predictions from your deployed models with live, external data from Snowflake.
At the Connections ‘22 event, I spoke with a Snowflake Partner Solution Architect (i.e. an architect who solely focuses on designing partner integrations) to get some insight into how Salesforce and Snowflake work together. Little did I know where the partnership had been heading, and how core Snowflake would become to Salesforce’s Genie vision. Let’s start with the ways that Snowflake shines in the market.
“Customer 360 in the new world of data privacy relies on leveraging first-party data across your business systems. Making this available seamlessly, and in near real-time, drives better decision-making and faster business processes. We believe the new integration between Salesforce genie and Snowflake will enable our customers to realize that.” M.Chandramohan Nayak – Partner Solutions Architect, Snowflake.
What Does Snowflake Do?
Snowflake is a data lake – a place where huge amounts of data can be stored, and accessed when you need it. Data can be either structured or unstructured.
- Structured data “has been organized into a formatted repository, typically a database, so that its elements can be made addressable for more effective processing and analysis” (source: TechTarget).
- Unstructured data is the chaotic cousin (attends the same family gatherings, but is not relatable in conversation); in data terms, it’s not structured in a database format.
Back to the lake analogy. Different data sets are “swimming” around, available for you to “cast your net”. Writing a SQL query (i.e. I want data that matches X, Y, Z criteria) then pulls the specific data you’re looking for. It’s somewhat similar to the segmentation features that are part of marketing automation platforms, just on a larger scale (and accommodating unstructured data).
Snowflake Strengths
MarTech professionals have raved about Snowflake for a long time, and for multiple reasons. We can drill its popularity down to four core reasons:
- Easy to use: Anecdotally, people enjoy using the Snowflake interface. What’s more compelling are the positive sentiments from those who have experience using alternative platforms.
- Scales well: Performance isn’t compromised (i.e. query speed), regardless of how many more users you add. The performance that other platforms offer tends to slow down the more users that are added.
- Consumption-based pricing: You pay only when you run a query. This means that you’re not paying up-front costs per user seat (that may never get used).
- Data marketplace: “Access live, ready-to-query data to unlock new insights with just a few clicks”. Transact data with other organizations – both to acquire data and sell data. In the world where third-party data is shunned, this can be seen as controversial; however, it’s common practice that enables organizations to stay highly relevant. Example 1: You’re a large CPG company – you can sell specific data sets to a supermarket so that they know what to stock. Example 2: Public sector data, such as disease spread, can be critical for many organizations’ decision-making in disaster prevention.
Snowflake + Salesforce Genie
Salesforce Genie unifies versions of the same individual across applications (using defined identity rules), sifting through a wealth of data, reconciling individual people into individual profiles, and delivering customer experiences based on data sources beyond Salesforce.
“Isn’t this what Salesforce CDP does?” I hear you asking – yes, you’re right, Genie’s value-sells do describe what a CDP does. In fact, Genie is Salesforce CDP, just revamped in big ways, and going beyond the traditional definition of CDP.
Now for the meaty part. Snowflake provides Salesforce Genie with secure real-time and open data sharing. Following the water-based analogy, data streams can now flow seamlessly between Snowflake and Salesforce – in other words, Genie can directly access data stored in Snowflake (and vice-versa) without moving or duplicating data. This is also referred to as a zero-data copy architecture.
Salesforce said that they have forged partnerships with “who’s who” in the data space to expand the power of Genie – a huge vote of confidence for Snowflake 👍
Why Use Snowflake with Marketing Cloud?
Both Salesforce and Snowflake have dedicated technical resources (a significant investment) into a closer integration between Snowflake and Marketing Cloud. This joint venture is to save your time, money and hassle from developing integrations.
1. Not all data is supposed to be stored in Marketing Cloud
Beyond the campaign data you use frequently, additional data will end up “bloating” your account with data that’s only useful a fraction of the time.
While Marketing Cloud has data extensions (which extend the data model) these don’t always accommodate your data needs, for example, a limited number of joins you can form between data extensions.
So, instead of pumping third-party data into Marketing Cloud, direct it to Snowflake. Once an SQL query is run (which finds what data you actually want to work with), then you pair it with the data in your Marketing Cloud account.
- Marketing Cloud = Campaign data you use frequently.
- Data lake = Contextual data.
Snowflake is the repository of data, and Marketing Cloud automates how, and when, your brand communicates with contacts.
2. Compile data (even if it doesn’t gel well together)
Snowflake + Marketing Cloud’s main objective is to compile data from multiple sources, even if that data doesn’t necessarily connect together, initially.
See the data collected beyond customer journey points, from platforms that are not traditionally considered sources of data that can inform customer journey behavior, for example, payment platforms, reservation systems, or legacy financial technology.
3. Enhanced Marketing Cloud ← → Snowflake integration
The current focus is to enhance the Marketing Cloud ← → Snowflake connectors, but not in the ways you may initially think.
Snowflake has focused on their connection with Salesforce CDP (versus the Marketing Cloud baseline data model), as we saw in the “Salesforce Genie” section. This makes sense – Salesforce CDP is designed for marketers (and all other departments) to define a single identifier per individual in their database, the “source of truth” amongst data pumped in from left, right, and center.
If Snowflake is the lake, and Marketing Cloud is the village on the shore, then Salesforce CDP (Genie) is the market stall holder that coordinates produce from multiple sources (e.g. fish from the fishermen, vegetables from the farmers), while knowing every customer by any of their nicknames.
Snowflake + Tableau
Tableau and Snowflake are a good match to enable someone to visualize data, to bring insights out of the depth of the data lake. Bring the story your data is telling you, to life (cheesy, but it’s true!)
Salesforce and Snowflake are so confident in this technology marriage, that they are offering promotions for you to get started.
Snowflake + Einstein Discovery
Run predictions from your deployed models with live, external data from Snowflake (Winter ‘23 release). Genie uses the Salesforce metadata model, it can tap into Einstein for real-time predictions. The Salesforce Einstein engine is making 175 bill predictions per day.
Similar to how Salesforce CDP (Genie) and Snowflake have been brought closer together, predictions created in Einstein Discovery can leverage live data that exists outside of the Salesforce platform. Again, this is all possible without having to transport or recreate records in two systems.
Organizations with superb data quality (that goes beyond the CRM) will train the smartest machine learning models.
“If we can enrich and unify and deepen the data, then your AI can do more, and if your AI can do more, then your customer interactions are that much more tailored and personalized.”
David Schmaier, President & Chief Product Officer, Salesforce.