Share post

Share post

Share post

Blog: Opinion

Explaining Salesforce & Snowflake’s new integrations from Summit 2023

Snowflake |

Salesforce

By: Chris Chandler

6 July 2023

Salesforce and Snowflake deliver new integrations for Summit 2023, however, they all require Salesforce’s CDP product which has been somewhat strangely renamed to Salesforce Data Cloud.

Salesforce Data Cloud is best described as an all-in-one integration layer for the Salesforce ecosystem. With pricing pitched at large enterprises, it is best suited to the most complex organisations who have gone all in on Salesforce. As such, adopting the Salesforce vision could be seen to run counter to a data warehouse centric architecture.


What is Salesforce Data Cloud?

Salesforce Data Cloud is core to all of the integration features announced at Snowflake Summit 23. Salesforce Data Cloud, not to be confused with the Snowflake Data Cloud, is the new name for its Genie product that was announced at Dreamforce 2022. In case that was too simple, Genie was previously known as Customer Data Platform, or CDP, and was also previously marketed as Customer Audience 360 (CA360).

Since the acronym of Salesforce Data Cloud and Snowflake Data Cloud are both “SDC”, and, SFDC was the old school acronym of Salesforce Dot Com. For the purposes of this article, we’ll refer to Salesforce Data Cloud by its previous name CDP.


What does it do?

Salesforce Data Cloud, or CDP, is basically a pre-canned integration layer for the Salesforce ecosystem. At its core, it is data infrastructure where you can ingest, model and activate data. It has a familiar UI and convenient location in the Salesforce App menu and mostly does what standalone CDP or integration middleware does. That said, it will enable a Salesforce Admin persona to create and manage integrations between the various parts of the Salesforce ecosystem, external systems and create unified data models.


Pricing

Pricing starts at USD 108k for the base platform, then you have to purchase credits to use the different features, called Data Services. I have not analysed the cost or output of a credit but judging by the target market, its pricing is likely to be higher than standalone CDPs. Activating data to other services like Ad platforms are also separate add-ons listed for an extra couple of thousand.


An enterprise target market

$108k before usage is quite a high barrier to entry.

The solution is pitched at the enterprise and likely only makes sense for those customers who have gone all in on the Salesforce stack with large complex deployments. For smaller organisations, with a more diverse best-of-breed stack, $100k of existing integration tools could deliver a similar solution for less. Similarly, customers who have gone for a data warehouse centric approach might be left scratching their heads.


Does it create a data silo?

In my opinion, yes. For those who believe in data-warehouse centricity, like us, adopting CDP would mean you actually take over work that should be done in Snowflake. CDP basically allows a Salesforce Admin to build data models from disparate systems in Salesforce. This is counter to the value that Snowflake adds as a central source of truth. I can imagine this is only valuable in organisations that are highly compartmentalised where a single data architecture has not worked.


The details the new integrations

The sessions and announcements highlighted the new integrations as Zero-ETL and real-time, using Snowflake data sharing. While it is true that CDP leverages real-time queries and data sharing between itself and Snowflake, the infrastructure of CDP remains separate to the other parts of the Salesforce ecosystem. There is still data syncing that happens between CDP and the core CRM, Marketing Cloud and others.

The core CRM, which is where most customers want to pull data from or activate it to, remains the most important target for integration. All in all, it still makes sense to integrate to these APIs directly.


Data going from Salesforce to Snowflake

CDP manages the fetching of data from the sources in the Salesforce ecosystem and then uses data sharing to deliver this to Snowflake customers. Under the hood, they use Iceberg tables in a Salesforce-owned Snowflake account. The data sharing portion is between this account and other Snowflake customers. While fully managed, CDP is an extra hop compared to a direct API integration.

It is real-time… kind of.

For certain standard CRM objects, Salesforce use platform events to send changes in real-time to CDP where they are stored as objects to be used by CDP or shared externally. For other objects, Salesforce use the usual API methods to sync data every few minutes.

On the other hand, today’s integration tools pull data directly from the Salesforce core CRM APIs. The biggest challenge is latency for changed records. Whether CDP solves this remains to be seen and Salesforce are yet to release documentation. We’ll have to wait and see what sort of latency exists between the CRM and CDP, and whether it’s worth the investment.


Data going from Snowflake to Salesforce

There were two ‘demos’ of Snowflake to Salesforce functionality during the Summit sessions. A keen eye 😏 will notice they switch to a Figma prototype suggesting that this part of the product is not quite ready yet.

The speaker even says, “ Enough slides, let’s get into some software”… (he) doth protest too much, methinks.

Nevertheless, you will be able to create what’s called a Data Stream that will query data directly from Snowflake with a JDBC connector. While this direct query method will indeed be real-time and free for ingestion limits to CDP. To get data all the way to the core CRM is another step.

In the Summit session “CO213: Salesforce and Snowflake Data Sharing: Bring Real-Time Data to Power Your Customer 360” , they show a feature called Data Cloud Enrichments. This is the feature that will sync individual fields from Snowflake table in CDP to CRM Objects.

The use case demonstrated was a LTV calculation that appeared on a Contact object, tackling the most common Reverse-ETL use case. They did not specify how often this could update, but since CDP is replicating data to the CRM, the usual data sync limits would apply here. In this case, CDP is performing the same function as a Reverse-ETL or other middleware tool. The biggest benefit here is that it is directly in the realm of Salesforce Admins to configure and manage.


Many integration options now

The Summit presentations used the below diagram to demonstrate the new architecture, however, I think this slightly confusing. It means that the data actually flows like this (I drew the arrow):

Weird right?


So I made a better diagram

Salesforce Data Cloud is a somewhat better integration layer for the Salesforce ecosystem, however, the core Salesforce CRM can be as easily integrated directly via API.

Indeed, lots of options. So, where does Omnata fit into this?


Omnata Connect - what we solve

In many cases, customers have already chosen to centralize enterprise data in Snowflake. It becomes the single source of truth for the organisation. In these cases, Salesforce is a downstream customer of Snowflake. As such, there is no need to create new data models or use Salesforce as a hub for distributing data to other applications.

Within the Salesforce ecosystem, the core CRM is the primary platform, but it should not be used as a datastore for anything other than what is generated by its use.

Omnata allows you to live-query data from Snowflake directly to CRM objects. It’s a lighter weight and lower cost solution to the core integration problem, without needing CDP.


Live-query Snowflake directly to CRM objects

Omnata, paired with Salesforce Connect, enables the Core CRM to live-query Snowflake tables and views and return the results to special objects called External Objects. External Objects are virtual objects, which return the data without replicating it. This means you:

  • Don’t create a data silo

  • Don’t need to deal with loading and reloading data

  • Save on Salesforce development and storage costs

Plus, Omnata Connect is a fully native app that runs in your Salesforce deployment. This means we:

  • Have no data-handling infrastructure making us great for PII and sensitive data

  • Don’t store your credentials or metadata for easier cyber approvals

  • Deliver cost savings from buying and developing with SaaS middleware

Omnata is used by companies like Vonage, Payroc, HBF and more to integrate large and real-time datasets from their enterprise data warehouse Snowflake to their Salesforce core CRM.

If you'd like to try it, visit the AppExchange or get in touch.


OK… one more meme

Salesforce and Snowflake deliver new integrations for Summit 2023, however, they all require Salesforce’s CDP product which has been somewhat strangely renamed to Salesforce Data Cloud.

Salesforce Data Cloud is best described as an all-in-one integration layer for the Salesforce ecosystem. With pricing pitched at large enterprises, it is best suited to the most complex organisations who have gone all in on Salesforce. As such, adopting the Salesforce vision could be seen to run counter to a data warehouse centric architecture.


What is Salesforce Data Cloud?

Salesforce Data Cloud is core to all of the integration features announced at Snowflake Summit 23. Salesforce Data Cloud, not to be confused with the Snowflake Data Cloud, is the new name for its Genie product that was announced at Dreamforce 2022. In case that was too simple, Genie was previously known as Customer Data Platform, or CDP, and was also previously marketed as Customer Audience 360 (CA360).

Since the acronym of Salesforce Data Cloud and Snowflake Data Cloud are both “SDC”, and, SFDC was the old school acronym of Salesforce Dot Com. For the purposes of this article, we’ll refer to Salesforce Data Cloud by its previous name CDP.


What does it do?

Salesforce Data Cloud, or CDP, is basically a pre-canned integration layer for the Salesforce ecosystem. At its core, it is data infrastructure where you can ingest, model and activate data. It has a familiar UI and convenient location in the Salesforce App menu and mostly does what standalone CDP or integration middleware does. That said, it will enable a Salesforce Admin persona to create and manage integrations between the various parts of the Salesforce ecosystem, external systems and create unified data models.


Pricing

Pricing starts at USD 108k for the base platform, then you have to purchase credits to use the different features, called Data Services. I have not analysed the cost or output of a credit but judging by the target market, its pricing is likely to be higher than standalone CDPs. Activating data to other services like Ad platforms are also separate add-ons listed for an extra couple of thousand.


An enterprise target market

$108k before usage is quite a high barrier to entry.

The solution is pitched at the enterprise and likely only makes sense for those customers who have gone all in on the Salesforce stack with large complex deployments. For smaller organisations, with a more diverse best-of-breed stack, $100k of existing integration tools could deliver a similar solution for less. Similarly, customers who have gone for a data warehouse centric approach might be left scratching their heads.


Does it create a data silo?

In my opinion, yes. For those who believe in data-warehouse centricity, like us, adopting CDP would mean you actually take over work that should be done in Snowflake. CDP basically allows a Salesforce Admin to build data models from disparate systems in Salesforce. This is counter to the value that Snowflake adds as a central source of truth. I can imagine this is only valuable in organisations that are highly compartmentalised where a single data architecture has not worked.


The details the new integrations

The sessions and announcements highlighted the new integrations as Zero-ETL and real-time, using Snowflake data sharing. While it is true that CDP leverages real-time queries and data sharing between itself and Snowflake, the infrastructure of CDP remains separate to the other parts of the Salesforce ecosystem. There is still data syncing that happens between CDP and the core CRM, Marketing Cloud and others.

The core CRM, which is where most customers want to pull data from or activate it to, remains the most important target for integration. All in all, it still makes sense to integrate to these APIs directly.


Data going from Salesforce to Snowflake

CDP manages the fetching of data from the sources in the Salesforce ecosystem and then uses data sharing to deliver this to Snowflake customers. Under the hood, they use Iceberg tables in a Salesforce-owned Snowflake account. The data sharing portion is between this account and other Snowflake customers. While fully managed, CDP is an extra hop compared to a direct API integration.

It is real-time… kind of.

For certain standard CRM objects, Salesforce use platform events to send changes in real-time to CDP where they are stored as objects to be used by CDP or shared externally. For other objects, Salesforce use the usual API methods to sync data every few minutes.

On the other hand, today’s integration tools pull data directly from the Salesforce core CRM APIs. The biggest challenge is latency for changed records. Whether CDP solves this remains to be seen and Salesforce are yet to release documentation. We’ll have to wait and see what sort of latency exists between the CRM and CDP, and whether it’s worth the investment.


Data going from Snowflake to Salesforce

There were two ‘demos’ of Snowflake to Salesforce functionality during the Summit sessions. A keen eye 😏 will notice they switch to a Figma prototype suggesting that this part of the product is not quite ready yet.

The speaker even says, “ Enough slides, let’s get into some software”… (he) doth protest too much, methinks.

Nevertheless, you will be able to create what’s called a Data Stream that will query data directly from Snowflake with a JDBC connector. While this direct query method will indeed be real-time and free for ingestion limits to CDP. To get data all the way to the core CRM is another step.

In the Summit session “CO213: Salesforce and Snowflake Data Sharing: Bring Real-Time Data to Power Your Customer 360” , they show a feature called Data Cloud Enrichments. This is the feature that will sync individual fields from Snowflake table in CDP to CRM Objects.

The use case demonstrated was a LTV calculation that appeared on a Contact object, tackling the most common Reverse-ETL use case. They did not specify how often this could update, but since CDP is replicating data to the CRM, the usual data sync limits would apply here. In this case, CDP is performing the same function as a Reverse-ETL or other middleware tool. The biggest benefit here is that it is directly in the realm of Salesforce Admins to configure and manage.


Many integration options now

The Summit presentations used the below diagram to demonstrate the new architecture, however, I think this slightly confusing. It means that the data actually flows like this (I drew the arrow):

Weird right?


So I made a better diagram

Salesforce Data Cloud is a somewhat better integration layer for the Salesforce ecosystem, however, the core Salesforce CRM can be as easily integrated directly via API.

Indeed, lots of options. So, where does Omnata fit into this?


Omnata Connect - what we solve

In many cases, customers have already chosen to centralize enterprise data in Snowflake. It becomes the single source of truth for the organisation. In these cases, Salesforce is a downstream customer of Snowflake. As such, there is no need to create new data models or use Salesforce as a hub for distributing data to other applications.

Within the Salesforce ecosystem, the core CRM is the primary platform, but it should not be used as a datastore for anything other than what is generated by its use.

Omnata allows you to live-query data from Snowflake directly to CRM objects. It’s a lighter weight and lower cost solution to the core integration problem, without needing CDP.


Live-query Snowflake directly to CRM objects

Omnata, paired with Salesforce Connect, enables the Core CRM to live-query Snowflake tables and views and return the results to special objects called External Objects. External Objects are virtual objects, which return the data without replicating it. This means you:

  • Don’t create a data silo

  • Don’t need to deal with loading and reloading data

  • Save on Salesforce development and storage costs

Plus, Omnata Connect is a fully native app that runs in your Salesforce deployment. This means we:

  • Have no data-handling infrastructure making us great for PII and sensitive data

  • Don’t store your credentials or metadata for easier cyber approvals

  • Deliver cost savings from buying and developing with SaaS middleware

Omnata is used by companies like Vonage, Payroc, HBF and more to integrate large and real-time datasets from their enterprise data warehouse Snowflake to their Salesforce core CRM.

If you'd like to try it, visit the AppExchange or get in touch.


OK… one more meme

Salesforce and Snowflake deliver new integrations for Summit 2023, however, they all require Salesforce’s CDP product which has been somewhat strangely renamed to Salesforce Data Cloud.

Salesforce Data Cloud is best described as an all-in-one integration layer for the Salesforce ecosystem. With pricing pitched at large enterprises, it is best suited to the most complex organisations who have gone all in on Salesforce. As such, adopting the Salesforce vision could be seen to run counter to a data warehouse centric architecture.


What is Salesforce Data Cloud?

Salesforce Data Cloud is core to all of the integration features announced at Snowflake Summit 23. Salesforce Data Cloud, not to be confused with the Snowflake Data Cloud, is the new name for its Genie product that was announced at Dreamforce 2022. In case that was too simple, Genie was previously known as Customer Data Platform, or CDP, and was also previously marketed as Customer Audience 360 (CA360).

Since the acronym of Salesforce Data Cloud and Snowflake Data Cloud are both “SDC”, and, SFDC was the old school acronym of Salesforce Dot Com. For the purposes of this article, we’ll refer to Salesforce Data Cloud by its previous name CDP.


What does it do?

Salesforce Data Cloud, or CDP, is basically a pre-canned integration layer for the Salesforce ecosystem. At its core, it is data infrastructure where you can ingest, model and activate data. It has a familiar UI and convenient location in the Salesforce App menu and mostly does what standalone CDP or integration middleware does. That said, it will enable a Salesforce Admin persona to create and manage integrations between the various parts of the Salesforce ecosystem, external systems and create unified data models.


Pricing

Pricing starts at USD 108k for the base platform, then you have to purchase credits to use the different features, called Data Services. I have not analysed the cost or output of a credit but judging by the target market, its pricing is likely to be higher than standalone CDPs. Activating data to other services like Ad platforms are also separate add-ons listed for an extra couple of thousand.


An enterprise target market

$108k before usage is quite a high barrier to entry.

The solution is pitched at the enterprise and likely only makes sense for those customers who have gone all in on the Salesforce stack with large complex deployments. For smaller organisations, with a more diverse best-of-breed stack, $100k of existing integration tools could deliver a similar solution for less. Similarly, customers who have gone for a data warehouse centric approach might be left scratching their heads.


Does it create a data silo?

In my opinion, yes. For those who believe in data-warehouse centricity, like us, adopting CDP would mean you actually take over work that should be done in Snowflake. CDP basically allows a Salesforce Admin to build data models from disparate systems in Salesforce. This is counter to the value that Snowflake adds as a central source of truth. I can imagine this is only valuable in organisations that are highly compartmentalised where a single data architecture has not worked.


The details the new integrations

The sessions and announcements highlighted the new integrations as Zero-ETL and real-time, using Snowflake data sharing. While it is true that CDP leverages real-time queries and data sharing between itself and Snowflake, the infrastructure of CDP remains separate to the other parts of the Salesforce ecosystem. There is still data syncing that happens between CDP and the core CRM, Marketing Cloud and others.

The core CRM, which is where most customers want to pull data from or activate it to, remains the most important target for integration. All in all, it still makes sense to integrate to these APIs directly.


Data going from Salesforce to Snowflake

CDP manages the fetching of data from the sources in the Salesforce ecosystem and then uses data sharing to deliver this to Snowflake customers. Under the hood, they use Iceberg tables in a Salesforce-owned Snowflake account. The data sharing portion is between this account and other Snowflake customers. While fully managed, CDP is an extra hop compared to a direct API integration.

It is real-time… kind of.

For certain standard CRM objects, Salesforce use platform events to send changes in real-time to CDP where they are stored as objects to be used by CDP or shared externally. For other objects, Salesforce use the usual API methods to sync data every few minutes.

On the other hand, today’s integration tools pull data directly from the Salesforce core CRM APIs. The biggest challenge is latency for changed records. Whether CDP solves this remains to be seen and Salesforce are yet to release documentation. We’ll have to wait and see what sort of latency exists between the CRM and CDP, and whether it’s worth the investment.


Data going from Snowflake to Salesforce

There were two ‘demos’ of Snowflake to Salesforce functionality during the Summit sessions. A keen eye 😏 will notice they switch to a Figma prototype suggesting that this part of the product is not quite ready yet.

The speaker even says, “ Enough slides, let’s get into some software”… (he) doth protest too much, methinks.

Nevertheless, you will be able to create what’s called a Data Stream that will query data directly from Snowflake with a JDBC connector. While this direct query method will indeed be real-time and free for ingestion limits to CDP. To get data all the way to the core CRM is another step.

In the Summit session “CO213: Salesforce and Snowflake Data Sharing: Bring Real-Time Data to Power Your Customer 360” , they show a feature called Data Cloud Enrichments. This is the feature that will sync individual fields from Snowflake table in CDP to CRM Objects.

The use case demonstrated was a LTV calculation that appeared on a Contact object, tackling the most common Reverse-ETL use case. They did not specify how often this could update, but since CDP is replicating data to the CRM, the usual data sync limits would apply here. In this case, CDP is performing the same function as a Reverse-ETL or other middleware tool. The biggest benefit here is that it is directly in the realm of Salesforce Admins to configure and manage.


Many integration options now

The Summit presentations used the below diagram to demonstrate the new architecture, however, I think this slightly confusing. It means that the data actually flows like this (I drew the arrow):

Weird right?


So I made a better diagram

Salesforce Data Cloud is a somewhat better integration layer for the Salesforce ecosystem, however, the core Salesforce CRM can be as easily integrated directly via API.

Indeed, lots of options. So, where does Omnata fit into this?


Omnata Connect - what we solve

In many cases, customers have already chosen to centralize enterprise data in Snowflake. It becomes the single source of truth for the organisation. In these cases, Salesforce is a downstream customer of Snowflake. As such, there is no need to create new data models or use Salesforce as a hub for distributing data to other applications.

Within the Salesforce ecosystem, the core CRM is the primary platform, but it should not be used as a datastore for anything other than what is generated by its use.

Omnata allows you to live-query data from Snowflake directly to CRM objects. It’s a lighter weight and lower cost solution to the core integration problem, without needing CDP.


Live-query Snowflake directly to CRM objects

Omnata, paired with Salesforce Connect, enables the Core CRM to live-query Snowflake tables and views and return the results to special objects called External Objects. External Objects are virtual objects, which return the data without replicating it. This means you:

  • Don’t create a data silo

  • Don’t need to deal with loading and reloading data

  • Save on Salesforce development and storage costs

Plus, Omnata Connect is a fully native app that runs in your Salesforce deployment. This means we:

  • Have no data-handling infrastructure making us great for PII and sensitive data

  • Don’t store your credentials or metadata for easier cyber approvals

  • Deliver cost savings from buying and developing with SaaS middleware

Omnata is used by companies like Vonage, Payroc, HBF and more to integrate large and real-time datasets from their enterprise data warehouse Snowflake to their Salesforce core CRM.

If you'd like to try it, visit the AppExchange or get in touch.


OK… one more meme

subscribe

Deliver the goods to your inbox

subscribe

Deliver the goods to your inbox

subscribe

Deliver the goods to your inbox