Salesforce, Inc. (CRM) Earnings Call Transcript & Summary

Salesforce data pipeline, Salesforce native ETL to transform and sync your Salesforce data, and thank you all so much for joining us today. My name is Ariana, and I'm on the Corporate Marketing team here at Salesforce. And before we begin, I'd like to cover a few quick notes with you about our webinar platform. Today's webinar will be available on demand after we wrap up, and it will be accessible through the URL you are on now. Please note that slides will advance automatically throughout the presentation, and you can enlarge them by clicking on the Enlarge Slide button located in the top right-hand corner of your Slides widget. Should you need technical assistance, click on the Help widget located on the bottom left corner of your console. We've also added a few additional resources, which are available through the resources window to the right of the slide. There you can find some additional related content. And lastly, we encourage you to submit your questions at any time throughout the presentation today, using the Ask A Question widget at the bottom of your console. We'll do our best to answer as many questions as we can at the end of our presentation. And with that, I will be turning things over to Amy to get us started.

Thank you, Ariana. Hello, everyone, and welcome to the Salesforce data pipelines webinar. Thank you so much for spending this next 45 minutes to an hour with us. We really appreciate it and hope you find it very valuable use of your time. Let me kick us off here with some quick introductions. I'm Amy Smith. I am one of the Regional Vice Presidents here at Salesforce for sales. I'm responsible for the CRM analytics team. Paul Garvey on my team is one of our subject matter experts for both CRM analytics and Salesforce data pipelines. And then we have Terrence Tse, who is one of the directors in product management, responsible for Salesforce data pipelines and our CDP customer data platform. And one quick slide here before we kick it off to keep the attorneys happy, our forward-looking statement. Please just make all buying decisions based on products that are generally available in the market today, and we will be in good shape with the attorneys. Okay, based on that, I'm going to go ahead and turn it over to Terrence to get us kicked off.

Thank you, Amy. Hi, everyone. Good morning, good evening, good afternoon from wherever you are, and thank you for joining us. So let me go into some slides to introduce you to what Salesforce data pipeline is, and then we will talk about some use cases and then a live demo, and we'll go into a Q&A session to answer any questions you may have after this presentation. So to start, let's take a step back and ask the question, how do you better understand your customers? Well, the answer is all within the data, and I think we all know that. Data-driven organizations better understand their customers and can provide them more value. And according to a study done by McKinsey, you can see some of the stats here on the screen, data-driven organizations provided an average of 15x more value to their customers. How would you like to provide 15x more value to your competitors? Well, I think we all would, right? And additional to providing more value, data-driven organizations are also 23x more likely to add new customers and 6.5x more likely to retain those customers. These are very amazing metrics that every single company can benefit from. But there lies a problem. To benefit from all that powerful data, you first need to bring it all together. Looking at this diagram, I think many of you may kind of feel like, "Oh, that's kind of the situation that we're in. We have data, but it's all over the place." So first, to bring it together, this can be very complex. It requires detailed tooling to integrate, prep and transform all of your data. And then you need to feed it somewhere to put it all together so that your users can get to it. This process can take time and a lot of computing power, which can be burdensome and can result in a lot less fresh data for your CRM users. So typically, what a lot of our customers do or a lot of existing companies do is they need to go bring all that [ satellite ] data together, but then they have to find these complex disconnected tools and implement those with the support of IT teams. So all of that is a very time-consuming process, as I mentioned. And by the time you get the data into a state that's ready for consumption by your CRM users, the data is probably already stale. And stale data means employees don't get the insights that they need in the moment that they need them. And what good is that, which is why we have introduced our product called Salesforce Data Pipelines. With Salesforce Data Pipelines, you can integrate, enrich and modify all your data natively within Salesforce, all with this one packaged product. And this does not require any more external systems or taking the data out of the trust boundary. With our point-and-click tooling and ML guidance, you can prep and transform your data faster on Salesforce because Salesforce Data Pipelines is native, and your admins can benefit from working in the Salesforce environment that they're familiar with, without needing to use any complex external infrastructure or tools that they need to learn and pick up new skills for. So all of this can be delivered to you -- sorry, to all your CRM users with just a few clicks and then provide a more comprehensive view of the relevant customer data, enabling a better data experience in Salesforce. As part of Salesforce Data Pipelines, we actually provide over 50 connectors out of the box, and that means you can virtually bring in any data into Salesforce, and make that available to your consumers or your -- sorry, your business users. And as part of that, it also makes administration simpler, such as providing scheduling, monitoring and [ configuration ] of these data syncs and transformation on your data. And we'll take a look at that later in the live demo. So after you feed the data into Salesforce, there are built-in functions in machine learning to help you clean and standardize your data. Machine learning capabilities enhance data quality by predicting the [ sync ] values in detecting anomalies and also enrich data using sentiment analysis and also recommend transforms to make that data cleaner and easier to consume. Then the clicks [ not quote ] experience is also very easy for anyone to pick up. You won't need any specialist data integration skills. Salesforce Data Pipelines is native to Salesforce CRM, so it uses and implements that admin-friendly architecture that all of our Salesforce customers are familiar with. And also to get things started quickly, we are also introducing something called data apps in this coming release. These are prebuilt templates that come out of the box to handle common data transformations on any data. To start, there will be one to help load contacts into Salesforce from external sources in order to calculate lifetime value, and one more to run sentiment analysis on unstructured data. Over time, additional data apps will also be added, and also new ones will be created by our AppExchange partners and offered through this gallery. So with that, let's quickly take a look at some sample use cases from some of our customers that have been using Salesforce Data Pipelines. So to start, we have a wealth management company who was importing and consolidating all of their wealth data across millions of their clients. And as part of that, they also wanted to pull in external investment accounts so that they have a full picture of what their clients' lifetime value is or kind of like what are all their assets. But with their existing or actually their previous solution, that whole process took over 70 hours to complete. So by the time a relationship manager gets that information in their CRM when they're talking to their customers during a consultation for instance, that data is already 3 days old. So it's not really that valuable. But since these implemented data pipelines that you can see here, this process was reduced to under 3 hours, which is kind of amazing. So in essence, this allows them to get updates to their data multiple times in 1 day. So the relationship managers are actually getting almost the latest information about their customers or their clients. So a lot of this was because using data pipelines reduce the in and out of the data out of the Salesforce, so you're not taking data out, processing it, waiting for that to finish, and then writing that data back into Salesforce, which typically takes a lot of time. So Salesforce Data Pipelines here helps a lot in improving some of those processes. Next here, we have a large manufacturing company who brings [ order ] data from ERP into Salesforce. So that could be the sales cloud or manufacturing cloud, if you're familiar with that. So that their sales teams can see current status of what products have shipped, customer's total order value and whether the payment has been received. But since orders often contain multiple products, which ship in complex, different schedules, this inbound data must be decomposed and matched against the Salesforce orders, accounts, opportunities and sales agreements objects that can actually give the sales and support teams the views that they need to be effective. So using Salesforce Data Pipelines, this can really help simplify this process. So without taking the Salesforce data out and then combining it, then putting it back in, we simply here just bring in the ERP data into Salesforce and do the joins and do the augmentation aggregations here within Salesforce and then present that back to the users right within CRM. So if a customer is using sales cloud or manufacturing cloud, where they click on a account object, they can actually have that information from the ERP right where they work, without needing to jump into different screens or going to another system and piecing that information together. And lastly, as a use case, we have a major brand that wanted to roll out a relatively new loyalty management cloud to their member service team, who actually have millions of loyalty customers, right? But then they realize before they can actually roll out loyalty management cloud, they needed to bring in a lot of their existing customer data from their existing external sources that contain information about their loyalty tier status, their activity history and their participation levels. What do they buy? How much have they interacted with the brand? And hearing all of that, it's actually not a really simple task, and it's also not a onetime data task, right? New data is generated on a daily basis, every time someone purchases something or does something with the brand, that new data needs to be brought back in taking accounted for to get these new and latest loyalty statuses of their customer. So in order to ensure a regular update of their loyalty's tier status and keeping customers moving through the program, they need a way to ensure [ regular ] imports of the activities and also rapid recalculation of their loyalty credits. So spinning up an existing system or external system to do this, then relating these results back to Salesforce can also, again, be a very overwhelming project. And of course, you probably have to involve IT and not just your Salesforce team to get this going. But again, with Salesforce Data Pipelines, because it's built on Salesforce, it also has the enterprise scale to handle all this large-scale data. We can easily connect to all these external systems and bring in this data and build out these pipelines or data pipelines so that these updates can be created and implemented very quickly. So with that, let me jump into a live demo. So I'll show kind of like how the product looks like and go through an example of recipe to build out those data pipelines. [Presentation]

So let me jump into an account record within Salesforce. So a very typical use case is to simply just do some aggregations on some metrics on your accounts or your customers. And normally, before Salesforce Data Pipelines existed, you would either have to do this using APEX or Formula Fields, but there were limitations of how much data you can aggregate and how many in terms of objects you can actually join together to get those results. And if you go hit those limits, then you would actually need to take this data out and do with an external tool and then put it back into Salesforce. So let me show you here how we can do that with Salesforce Data Pipelines without moving that data in and out. So on this account record, I've added 2 custom fields that we would want to calculate. So down here, you'll see the latest sentiment, which is now blank, and also the lifetime spending, which is also blank. So we're going to fill these 2 fields with the information using Salesforce Data Pipelines. So if I jump back to the Data Manager and open this recipe that I precreated, and we will be able to see kind of what are those transformations that we do on the data that we've brought into Salesforce Data Pipelines. So if we look at the top stream here, you'll see that we are bringing in both the opportunity in line item, object and the opportunity object here. So first of all here, what we do is we join these 2 objects together. So you can see here we support multiple types of joins. But since here, we kind of want to match our line items with our opportunities by their operating ID. So here, we specify the join key, then you'll see a preview here on the right with all those records that match between these 2 objects. So this is a sample of the data that you're joining, so it gives you a good idea of what's happening as you work through this pipeline and make those transforms and joins on that data. So next into the stream, we want to take into account only the opportunities that are 1. So we create a filter here where 1 is equal to true, and now we will resolve in all the records that are only flagged as 1. And then next here, we do an aggregate. So what we want to do is for each account, we want to aggregate their total price, so the total amount that they have spent on the deals that we've sold to them. So then all of this output, again, you can preview and see that now with that for every single account. I know these account IDs look kind of similar, but they are different, that we get a total price for each of those accounts. Then if we move down to the bottom stream here, we have a source -- data source here called customer comments. So this data was stored in our Amazon S3 bucket, and we've brought this in. And as you can see here, it contains an ID of who the customer is, but also contains a free text field of comments of their reviews or their feedback to our products. But this is not very useful, right? If it's just a bunch of free text, you can't really analyze it, there's no good idea to kind of sniff through what the sentiment is. So built into SDP, we have a couple of what we call Smart Transforms, which means that they're powered by ML. So the one we're using here is called Detect Sentiment. What it does here is it passes through all the free text that you have for each record, and it will classify that free text as being negative, positive or neutral. So then now you have a more definite value that you can evaluate when you're looking at these results. In addition to that, we have something called predict missing values. So if you have data coming into the system and then some of the fields are blank, what this one would do is actually goes through all of your records in that data and tries to find correlations and try to identify what the possible value could be for that missing value. And then we also have a time series forecasting, which is very popular with our customers that can help you forecast different values, like predicting revenue or sales numbers et cetera. And lastly, we're also have something to do with clustering. So looking at your data, it will decide and figure out what those clusters should be and group that data together so they could be analyzed or looked at from a more simple point of view. So that's a quick overview on some of the analog transformations that we provide. But with that, now we have the total -- aggregated total price for each account, we also have the latest sentiment for each of our accounts. Now what we want to do is let's put those together, so for each account, we now have the sum of total price and their latest sentiment. And once we verify, that's good, we want to write that data back to CRM. So if I expand this here, you can take a look at that. What we're doing is we're writing this out as an output into our SFDC, so Salesforce local incidents or your Salesforce side. And then we want to choose that we want to put this data back into the account object. And what we want to do is update existing records, so we don't want to create any new records, we don't want to upsert any records, but we want to just update the existing records. And then -- I can get this to go away. We then map this back to our ID field. And then, since these fields are similar, the auto map, what those values should be -- sorry, which fields they should be, that needs to be updated. So I'm now going to quick save on this because I made changes. But if we jump back to the data manager where our job was running, we can go to the job monitor. Hopefully, you can see that here, that job has now succeeded in running. And then if we jump back to our record here -- we can do a quick refresh. Back to the Details tab, you now see that this data that we've calculated for this specific account has the latest sentiment of a neutral sentiment. And also, their lifetime spending is $131,000. So you can simply put this data back into a field within the object or you can take that data and you can build up some lightening web components and display it in a more visual way so that our users can actually see that data has been updated and is available to them as they need it. So one thing I realized that I glanced over is going back to the data manager, I forgot to talk about the connections. So here, you can see that we have a default connection created to our current Salesforce instance. So here, you can see that we have all these different objects coming in from Salesforce, and these are running to sync the data so that we can operate on those using Salesforce Data Pipelines. So in addition to bringing data from Salesforce, as I mentioned before, we have bunch of other connectors to bring data in from any source. So again, virtually you bring in any data and combine that with your Salesforce data. And of course, you also write this data back out to a couple of destinations for consumption. [Technical Difficulty].

Can you still hear us, Terrence?

Sorry about that technical glitch. So I will be right back. So okay, good. So that was the live demo. So I guess the question that you may want is like, how do I get this, and how much is it going to cost? Well, Salesforce Data Pipelines, our list price is $2,500 per month, so that makes it less than $30,000 a year. And what this comes with or include is 30 hours of processing time per month and 10 gigabytes of data written back into CRM. So of course, there's other limits, but I want to point those out because those are the main differences. So for customers that are -- that have a CRMA or CRM analytics, this is where Salesforce Data Pipelines differs a little on how we provide kind of the limits here. SDP is more of a usage base, so you can buy more as you need more. So these processing hours can [ stack ] as you need to buy more Salesforce Data Pipelines licenses. And of course, this is available now, and it is an add-on to any of our customers to have sales service or any of our platform clouds and industry clouds here. All right, so that does it with my presentation, and I know Paul, you've been monitoring our Q&A there. Let's jump over to some questions we can answer.

Sure, Terrence. We've got about 5 or 6 questions for you. First is, how does this compare to MuleSoft or Datorama?

Great question. So in the Salesforce, of course, there's a lot of different types of ETL integration type tools, even more out there in the market. But the main difference here is MuleSoft is typically more focused on more event-based API types of integrations, so real-time events coming in from your apps or your mobile devices and such that are out there. And Datorama also provides data but specifically for marketing use cases, right? They connect you to your social feeds, connect you to your ad sources like Google Ads or Facebook and provides analysis on that data. But SDP is more a batch-based processing, so you can bring in a large volume of data from various different sources. So it's agnostic to what kind of data that you want to analyze or look at or add to your system. And of course, you can have very fine control of kind of how you want to transform that data, how you want to build those data pipelines and where that data should go, be it to your CRM or to an external source, for consumption.

Great. Thanks, Terrence. Another question here is you mentioned CRM analytics and Einstein, what's the minimum module necessary for using Salesforce Data Pipeline?

So you can simply buy and start with one. And so when we say one, it's one license or we use a metric called DPUs or data processing units. And as I mentioned earlier, that provides you with 30 hours of processing time. So a good way to kind of look at it at it if your CRMA user or a customer, you can take a look at how long your existing recipes or data pools are running. That will give you a good idea of how many DPUs you would need to add on to your existing -- sorry, your existing processing power. So just to go down in a little bit more detail for customer -- sorry, for CRMA analytics customers or EA customers, regardless of how many license that you buy, there is a fixed limit of how many recipes you can run at the same time and how much you can process within a fixed time frame. Many customers, especially larger customers who have a lot of data, they may run into these limits, where what is allocated to them through, which is the CRMA licenses, may not necessarily enough. So looking into SDP, that's how we can give you more capacity to help you process more data for your business.

I guess that question, look at it in reverse, if you're not a CRMA customer, so I don't own any CRMA analytics and I don't own any Einstein, I assume that you just need a Salesforce sales and service cloud seat to actually invest in the Salesforce Data Pipeline. Is that correct?

That is correct, that is correct. So you don't need to be a CRMA or Einstein analytics customer. You can just buy SDP, if you're a sales cloud customer or Service Cloud customer to help you handle those ETL processes that you need on your data. So no analytics needed here.

Okay, and another question was, is data pipeline included in CRM analytics, if I'm already a CRM analytics customer?

To a certain extent, yes. So you get the functionalities that I've shown you. So you get the recipes, you get the connectors, you get monitoring, the scheduling as part of CRMA. What you don't get is that [ stackable ] or expandable processing time that SDP provides. So as I mentioned before, with CRMA, and I think there's the help documentation that is available there in one of the resource links, you can take a look at how those compare between the two. But there is a limit of how much you can do with CRMA, whereas with SDP, for each additional access you buy, those limits get increased to help you.

Okay. And a question about filters, and I think you actually demonstrated this, but just to clarify, can you add filters on a subset of your data used by an access by Salesforce Data Pipeline?

So if I get that question correctly, is that we want to filter on a subset of data before we process it, or is it more of like as part of the process, we want to filter the data so that we use only a certain part of that data as an output, if that even makes a difference.