# The Composable CDP: Making Snowflake Your Marketing Platform

- Channel: [Hightouch](https://www.youtube.com/channel/UCmjKHswj5UPdKqkrHB8RhVw)
- Published: 2022-08-24
- Source video: [Watch on YouTube](https://youtube.com/watch?v=_6WUl-2yLC0)
- Read on ReadTube: https://www.read.tube/p/videos/_6WUl-2yLC0
- Language: Original (source language)

## Summary

**Composable CDP for Snowflake**

The video argues that Snowflake can serve as the backbone of a modern marketing stack by unifying customer data into a single source of truth and pairing it with Hightouch for activation. It contrasts this composable CDP approach with traditional CDPs, saying it reduces data silos, latency, governance risk, and time to value while supporting flexible modeling, identity resolution, and fast activation to tools like Salesforce, Google, and Facebook. The speakers also note that marketers can start with one use case, such as CRM syncing or offline conversions, and expand over time without storing data in Hightouch.

The video argues that Snowflake, combined with a data activation layer like Hightouch, can serve as a better marketing platform than a traditional CDP. The core idea is to keep all customer and enterprise data in Snowflake as a single source of truth, which reduces silos, lowers data latency, improves governance and privacy, and makes data more flexible for marketing use cases.

Traditional CDPs are presented as too rigid, slow to implement, and isolated from the rest of the data stack, often creating two sources of truth. In contrast, the composable CDP approach uses Snowflake for storage, modeling, and identity resolution, then activates that data into channels like CRM, ads, and email tools through Hightouch. The speakers say this gives teams faster time to value, broader use cases, and more control, while still supporting technical and nontechnical users.

## Article

### Opening the LinkedIn Live Event and the Two-Part Series

All right, hello everyone. Welcome to the LinkedIn Live event, **Snowflake the Marketing Platform**, a two-part series exploring why and how Snowflake is your most powerful marketing tool.

Today is part one: **the business benefits of making Snowflake your marketing platform**.

Before we get started, just a few housekeeping items. We have a packed agenda for a 30-minute broadcast, but we're going to do our best to save some time at the end for questions, so please feel free to submit any on the stream as we go. This session will also be recorded and shared on YouTube, so don't worry if you or your teammates have to catch up at a later date.

Lastly, please keep in mind that this is only part one of the series, with part two's technical deep dive happening this Tuesday, August 23rd, at 10 a.m. Pacific Time, same place, where the team will dive into hands-on demos of everything that we discussed today.

All right, all that said, let's jump in.

### Introductions: Hightouch, Snowflake, and the Marketing Persona

My name is **Alek Aze**, I'm a product evangelist at **Hightouch**. I oversee product marketing as well as product management for Hightouch's marketing personas. Before joining Hightouch, I was a director of data products at Red Ventures, where I worked across their various brands implementing martech tools like CDPs and ESPs. Funny enough, I was actually a customer of Hightouch for about a year before joining, before jumping over to join the team.

Now enough about me. I'm joined today by our data cloud expert **Lorenzo Mello**, who will help us better understand how and why modern organizations are investing more and more into making the Snowflake data cloud their most important marketing tool.

Lorenzo is a product marketer at Snowflake. He is currently leading the marketing efforts for the marketing persona and has over six years of experience in enterprise cloud products.

Lorenzo, thanks for joining us today.

> Hey Alek, thanks for having me and I'm really looking forward to this conversation with you all.

### Why Modern Marketing Teams Need Unified Customer Data

It's no secret that the best marketing teams have an ability to fully understand and leverage their customer data.

Modern marketing teams share the desire and the ultimate goal of collecting and aggregating their customer data to create a single view of consumers across their many browsers and devices. They know that once available, that single customer view provides the deep, accurate insights needed to validate each decision that they make, from the channels they spend ad budgets on to the experiences that they build.

Beyond insights, these teams have an ability to take action on the single customer view, deploying audiences to ad platforms or sending profile information to email tools. This helps to create the personalized experiences desired, if not expected, by the modern consumer.

Historically, the best solution to collecting, aggregating, and actioning on customer data was a customer data platform, or a CDP. While cloud CDPs offer an off-the-shelf solution to collecting, cleaning, and activating customer data, adopting organizations have long since struggled with their rigid data models, long onboarding times, and data redundancy across analytics and marketing tools. The CDP Institute's latest survey found that only **58 percent** of companies with a deployed CDP say it delivers significant value, leaving much to be desired from these packaged CDP solutions.

Thanks to the rise of the Snowflake data cloud and data activation platforms like Hightouch, the **composable CDP** has emerged as a category-leading solution for modern organizations to manage their first-party data strategies.

In today's conversation, we're going to discuss how marketers today leverage the data cloud, the rise of the composable CDP, and hopefully help you all understand how the data cloud can be your most powerful marketing tool.

To start it off, it'll be helpful for everyone to get a better understanding of the state of the data cloud today and how organizations are using it to produce a single source of truth for all of their customer data. Lorenzo, do you mind providing some insight here?

### The Data Cloud as a Single Source of Truth

> Yeah sure Alek. So even if we take a step back, and before getting into specifics about things around marketing tech stacks that you and I are passionate about, I think it's easy to look at where we are as consumers, which is in its most simplest form I think the way to understand how the market has drastically changed over the last decade or so.
>
> I'm referring to the absolute avalanche of content platforms, ads, digital properties that we have access to and that we interact with on a daily basis. So all of that is actually leaving behind a massive data trail that comes in all different shapes and sizes, right? Ideally you would come very organized and structured, unfortunately that's not the case.
>
> So when we think about that, I think as consumers we can kind of empathize with how challenging it is for marketers today to achieve what they're really trying to achieve, which is obviously centered on providing the most relevant and impactful experience for all of us. And that can only be accomplished with a data-driven approach, right? Or a marketing strategy that's founded on unified and integrated data.
>
> And the reality is if you're operating with silos and you're not embracing this concept of a single source of truth for all your data, then I believe, I think we believe, that you're setting yourself up for unnecessary barriers and ultimately poor results.
>
> And so Snowflake, the data cloud, fundamentally is that component in your stack that breaks down those silos, right?
>
> So all of you listening out there and wondering what this actually means, let me dissect it a little bit for you. Fundamentally it means that you're working with a single data platform for all your enterprise data.
>
> So as a marketer that means you can ingest data from any source, whether it's internal and external data. You could be talking about first-party data, second-party or third-party data via Snowflake Marketplace. You can take in any format of data across structured, semi-structured or unstructured data, and really any method of data as well across batch, streamed, and obviously shared data as well.
>
> So I say that all of that to say the moment that you can't take in one of these types of data that I just listed out, then you're forced to have a silo. Snowflake doesn't have that issue, right? You can put it all in Snowflake and not worry about it at all.
>
> What you're looking at in this graph, you're seeing event data ingestion on the left, so partners that come to mind that, you know, certainly Hightouch and Snowflake work with like Snowplow, for example. We work very closely with batch data ingestion like Fivetran. It all flows into the data cloud where organizations can then model their data directly, for example with a tool like dbt. So that's kind of what you're seeing here.

### Why Data Latency and Governance Matter for Marketers

One of the things that we sometimes see as marketers that really hinders experiences that I think Snowflake can address is really regarding **data latency**. So what can you tell us about that?

> Yeah, that's another component of the data cloud that is a really important callout, and that's obviously critical in the marketing context.
>
> So many of us out there, and maybe you as well Alek, we've been in situations where we're getting ads for a product that we've already bought, or just simply the wrong ads, right? Typically that can be attributed to things like data latency.
>
> In latency, you might have a couple silos of your data and they're not communicating with each other. One of them is updated with fresh data, the other one is not. And so with the data cloud, marketers work with that fresh data that is updated in near real time. That even includes all the shared data that organizations work with throughout their ecosystem.
>
> So the simple concept of not having multiple data copies, but also working with a high performance data platform like the data cloud, it drastically reduces that data latency and that issue that we just laid out.
>
> And then lastly, I didn't mention this but I'd be remiss not to, regarding governance and privacy, this is obviously top of mind for every marketer out there as we deal with managing and treating with all care in the world customer PII.
>
> And the data cloud mitigates those privacy risks by limiting the number of data copies and the movement of those data copies. So by definition if you have less copies you're actually reducing that risk, right? And this obviously provides a really solid foundation for marketers, but alone it doesn't get you across the finish line of what you're trying to achieve because as marketers you also obviously need to take action on your data. You put it all into Snowflake, but then you need to act it, right? And Alek, I'll hand it back off to you because you're certainly the right person to walk us through that part of it.

### Activating Snowflake Data with Hightouch

Yeah, and this is really where Hightouch and data activation comes in, so going beyond analytics to activate that single source of truth that's in the Snowflake data cloud.

At its core, Hightouch is powered by a technology called **reverse ETL**. To put it really simply, we take data from the data cloud and we pass it down to any of our over **100 pre-built destinations**.

We started out by focusing on empowering data teams who define queries, which are called data models, using SQL. Hightouch then runs these queries against your data source and sends any updates or diffs out to any of your selected destinations, so you can see some on here: Facebook, Google, Salesforce, Braze, etc.

Now we handle all the scheduling and field mapping to make data activation seamless for any modern data team. And the beauty of this approach is that it's warehouse centric, meaning only SQL is required to activate data into the various destinations. There are no new databases to set up and maintain, and teams really get the confidence that they're working from their single source of truth, the Snowflake data cloud.

> Yeah, and that's great. One of the things that I think people out there might be thinking about and really wanting to get your thoughts on this, Alek, is marketers generally struggle with some of the more technical concepts like SQL, right, and query language. So how does Hightouch think about making a UI more friendly for marketers and having them kind of deal with that modeling piece in a little bit more of a friendly way?

Yeah, and we actually have a little bit of a recording here, and we'll go through the demo a bit more in the next session on Tuesday next week. But as the product has evolved, we realized that there was an opportunity to support all users regardless of their technical skill set, right?

Through the development of **Hightouch Audiences**, marketers are now able to come in, use this marketer-first no-code UI, create granular customer segments that can then be activated to their many business and marketing tools like Salesforce, Google, Facebook, etc.

So this allows modern organizations to manage activation of all of their customer data centrally while giving their data and marketing teams the tools that they need to do their jobs more efficiently, again all pulling from one single source of truth being the Snowflake data cloud.

### Why Traditional CDP Architecture Breaks Down

> Yeah, and I think, you know, when we talk about the traditional CDP architecture, even taking a step back, the idea behind taking all your customer data, putting it in a single place, cleaning it, and then activating it, it makes perfect sense. It made perfect sense back then when the emergence of the category came about. It still makes sense today, right? That's kind of where the logic behind the emergence of the CDP category came to be.
>
> The challenges, and really where poorly constructed CDPs, you know, I think fell short, is that CDPs became a data repository for customer data but it was isolated from the rest of your data, which traditionally lived in the data warehouse.
>
> So when you're looking at this graphic, which I think paints it well and kind of shows you the complexity of the architecture, you can clearly see that there's a source of truth for the data which is the CDP, right in the middle, and then there's another one that is the data warehouse.
>
> So the moment you have multiple data repositories, you're back to square one dealing with data silos again, which, you know, ironically were part of the value proposition that CDPs intended to break down in the onset of the category.
>
> And by the way, this creates also again challenges related to governance and security, right? Because you're having to update multiple copies and you first of all have multiple copies of data, so you're really creating some challenges for yourself if you're having an architecture like this.
>
> But I've touched a little bit on single source of truth, we've mentioned security and governance, we previously mentioned a little bit about the data latency issue, but there are other issues with this architecture that you've seen out. Do you want to speak to those?

> Yeah, so I think this architecture really speaks for itself, right? So as you said when early CDPs were created, the data cloud was nowhere near as powerful, right? Let alone actionable. So it made sense why we wanted to house everything in one bundled solution.
>
> As the data cloud adoption rose, however, more and more teams wanted their data syncing from CDPs down into the cloud, leaving us with two sources of truth for customer data, right? Your data and analytics teams were running off of Snowflake, your marketing teams were running out of their CDPs, right?
>
> I think we've touched on that enough though and counted enough sources of truth here, so let's highlight some of the other big pain points. Let's talk about things like time to value.
>
> Anyone who's gone through an end-to-end CDP evaluation knows it's nowhere near as easy as it sounds, right? So from the time of kickoff to when you can actually get value, you're looking at close to a year. You need teams of engineers and data stakeholders required before you can even think about activation.
>
> This is in part because you're locked into meeting their rigid data collection and modeling requirements. You can't just use your existing data and you need to abide by their rules across your entire stack.
>
> We believe in the core concept though that your activation tool should conform directly to your own data and not the other way around. We bring the activation layer to the data, right?
>
> And so it should start to be clear here, and given the rise of Snowflake in the data cloud in combination with data activation, why and how teams can overcome the challenges associated with traditional CDPs.

### The Composable CDP: Snowflake Plus Hightouch

Exactly, and this really brings me to the composable CDP. The composable CDP is kind of a combination of those first two slides that we walked through, right? Where you have Snowflake taking in all the data, you have a single source of truth, and now you're putting Hightouch on the back end of that, connecting it to all your downstream destinations and activating that data.

So it's a best-in-breed cloud CDP alternative that harnesses the power of the data cloud and data activation to empower these modern marketing teams.

Just like a traditional CDP, teams can address data collection, this time though it's through behavioral event collectors and ELT solutions; storage, modeling, and ID resolution, this time though it's in the Snowflake data cloud; and finally that activation layer with Hightouch as that final mile.

Aside from a much cleaner architecture, teams get some massive benefits here.

#### Flexible data modeling with SQL and related models

Flexible data modeling is the first one and by far my favorite to talk about. There's no longer a need to abide by the traditional users or events limitations of CDPs, as you have the power of SQL and the data cloud to handle any possible data modeling that you need.

While ID resolution is definitely more flexible and you can set whatever rules you want, the biggest standout in my eyes is related models.

My favorite use case and best way to explain related models is actually one of our largest customers. A national pet store wanted to run birthday email campaigns in Salesforce Marketing Cloud. The birthdays though weren't for the customers themselves, right? The birthdays were actually for the customers' pets.

Having to tie dates from an unlimited number of pet objects back to a user profile would have been pretty much impossible with a traditional CDP, and apologies out there to anyone who has ever tried to do this, you know the pain of doing so.

In the composable CDP however, using Hightouch it's actually made easy and solved in a matter of minutes thanks to related models, where you simply can define a relationship between a user and an object, right? You're using all of the data that you have and defining whatever relationships that you need. You don't have to abide by these rigid rules anymore.

#### Privacy, security, and fewer data copies

Privacy and security is then another really big callout. Your data is stored centrally in one place, Hightouch simply queries from your data cloud, meaning less data in third-party systems. I know today is focusing on the business use cases and value props here. The next session will actually dive in a bit more to that topic specifically.

#### Time to value and launching in minutes

Time to value is then really the final big pro here for marketers. Launching only requires hooking up Snowflake to Hightouch, which can be done in a few clicks thanks to Snowflake Partner Connect.

Marketing use cases like offline conversions to Google or Facebook can literally be stood up in minutes end to end. You just simply connect Snowflake directly from Snowflake to Hightouch, connect Facebook, connect Google, and set up your first sync. Super easy.

### Additional Benefits of the Composable Approach

> Lorenzo, any other big callouts that you would want to add here that I may have missed?

> I think you covered it well. You brought a great list. I think I'd say a couple of things as well.
>
> One of them is, you know, you touched on it briefly Alek, but the fact that you're working with best-of-breed tools across your stack I think is a huge plus. And we hear with customers that do this successfully all the time, that best-of-breed part is really appreciated by customers. You're working with a tool that's an expert in what you're trying to accomplish and it seamlessly integrates with the stack that you're trying to build, so that's one thing I wouldn't overlook.
>
> And I think also the second thing I'd call out: even when you think about it from a use case perspective, if you're siloing your data by implementing a CDP that isn't connected to the data cloud or your organization's whatever that may be central data repository, you're actually limiting the use cases that can be successfully delivered because you're limiting the data that marketers have access to, right?
>
> So one example for this is connecting marketing data to supply chain data, for example, can allow marketers to push excess inventory, right? So if you have something that you want to push out, tie your ad campaigns to that.
>
> Alternatively, what we see is if you don't have inventory of a specific product then you shouldn't be pushing ads for that, right? So we actually saw this I think during the beginning of the pandemic with, you know, toilet paper being pushed out in a bunch of advertising campaigns, people were getting to the store, there wasn't any, right?
>
> So just the fact that you can tie that kind of data together and put it all in one place in your data warehouse, you know, ultimately kind of improves or increases the breadth of use cases that you're able to address.
>
> And for marketers, you know, it improves ROI and return on ad spend, for example. So that's just another example of why having the data cloud as kind of what we're calling the backbone, right, of your marketing data stack and really this composable CDP approach can be so impactful.
>
> And, you know, obviously there are things that naysayers will say about this approach though, Alek, right? Like the detractors of the composable CDP will mention some of the issues with this. For instance, what would you tell someone you know that mentions the shortcomings of this approach related to, for example, real-time use cases?

### Questions and Objections from the Audience

#### What does real time really mean?

Yeah, that's a good callout and I'd say that's definitely one of the biggest pushbacks we get. But one of the first things that we respond with is we ask what real time really means, right?

The phrase real time is widely interpreted and in many cases data that's fresh within even **15 to 30 minutes** is totally appropriate. So most of the time we actually realize that you don't mean real time.

When they do, API calls can be made to power things like transactional emails, right? Like when you send that welcome to your first customer you want that as soon as possible.

Now all that said, we're also actively exploring things like streaming sources to help make this even easier in the future.

#### Identity resolution in the data cloud

Now another big pushback that we typically get that's an issue with the composable CDP is related to ID resolution, right? You have to do it yourself. Off-the-shelf CDPs offer that as part of their solution.

Lorenzo though, you're definitely better equipped to speak through this one given it'll live in Snowflake.

> Yeah, right. And we hear this one quite a bit when the composable CDP is kind of mentioned as an approach.
>
> And I think the data cloud is really uniquely positioned as the right data platform to address this particular issue.
>
> So besides working with nearly all of the leading identity resolution providers that marketers typically work with, seven of those top providers today are natively built on the data cloud, probably with more coming.
>
> So just to list a few examples for those of you out there who work with these partners: **Axiom, Experian, Infutor, Merkle, LiveRamp, Neustar, and Identified**. All of these are natively integrated into data cloud.
>
> And you're probably thinking, well what does that actually help me with? The impact for marketers there, besides having top-tier identity resolution available in this approach, the real step change is that it used to be that you had to extract your data, copy it, take it to the identity resolution provider, have it be resolved, and then ultimately bring it back to your data platform.
>
> So used to pull your data and take it to your application of choice. Now it's actually the opposite, right? The application of choice actually comes directly to your data and you're able to resolve identity right there.
>
> So that's a really, really powerful step change of having it natively built in the data cloud.

#### Data modeling is already happening in the warehouse

> Yeah, I know I would say this is actually really closely aligned to kind of that final objection that we hear as well of data modeling, right? You have to do it yourself.
>
> We hear people pushing back on that need to model their own data directly in the data cloud, but most of the time it's being done anyway, right? Like if you go back to that graph and you have two sources of truth, your data team is not just sitting there twiddling their thumbs. They're doing data modeling in the warehouse already, right?
>
> So it's just adding this layer of ID resolution and then using their existing models, which by the way goes back to unlocking massive opportunities where you can move faster off of your single source of truth that's more flexible and more secure than what any traditional CDP could offer you.

### Closing Q&A: Where to Start, Common Use Cases, and Privacy

I know we covered a lot here. I did want to leave time for some questions, so feel free to throw any in the chat. I'll keep an eye on this as well. It does look like we have a couple, so Lorenzo maybe you take the first one.

> What's the first step that marketers should take when beginning to think about making this vision come to life?

> I think the first step, I'd say there's two first steps. One of them kind of technologically speaking, which we've been covering a little bit today, and then another one that's more around alignment.
>
> The first step, and we see this sometimes with customers, is don't try to boil the ocean. Focus on this particular problem that you're trying to address and start there, right? And start with internal alignment across your business teams, your marketing teams, and your technical teams, and everybody get alignment into one or two things that you're really trying to address concretely. Start there and then expand, right?
>
> And I think the technological piece is, and Alek, like we've talked about this before, you're putting all your data in one place and then suddenly you have a reverse ETL, leading reverse CTL through Hightouch, that you can activate it. Putting all that data in one place is step one, right? You need to have that single source of truth of data to truly kind of see the potential and the impact that's available through this composable CDP approach.
>
> So I'd say two things: one of them is aligning internally on specific use cases you want to tackle first and then eventually expand, and then the second thing is put all your data into Snowflake. Make sure that all your data is in that single source of truth.

Exactly. And I think one of the really interesting things that we've seen customers do is they find initial use cases for this and they'll just launch from day one, right? So let's say that you have offline conversions that you want to send up to Facebook. They want to build toward a composable CDP, but guess what? Just hooking up Hightouch and Snowflake, you can stand up that use case, and again in a matter of minutes, while still then building toward the broader, bigger composable CDP.

So you get this first initial use case as you're building, where in a CDP you really have to do all the legwork upfront before you can even think about that activation.

Another question here: what are the most common activation use cases?

I can definitely tackle this one. I would say CRM syncing is definitely one of the largest, so making sure that Snowflake, all the data from Snowflake, is aligned in Salesforce or HubSpot, right? Your CRMs, making sure that that profile has always up-to-date information about a given user for many use cases, right? B2B, any type of lead scoring model, syncing that out, etc.

Another one that we see is related to all the ad tech platforms. We could do a full webinar on the demise of third-party cookies, first-party data strategies, right, so I won't go too deep there. But basically Facebook, Google, Snapchat, Twitter, LinkedIn, all of those platforms, all those walled gardens, need first-party data to match to given user profiles when you're sending conversions or you're sending audiences, right?

So one of the easiest use cases to get stood up is enrich conversion events from the warehouse from the data cloud, where you're sending conversions tied with all the other user identifiers that you have, right? So you're enriching all of your conversion signals and making sure that those platforms get the best signal from there.

Then you have audiences as well, where you build out and you can sync audiences across Facebook, Google, Iterable, all those different platforms, right? So I'd say those are definitely big too.

Looking at the diagram here, you can see transactional DBs. Another use case that's come up a lot is people wanting to personalize experiences, right? So we can sync things out of Snowflake to a transactional database that's then used in personalizing an actual on-site experience or application.

Another question here: **How can working with a composable CDP help address issues around privacy?**

> Yeah, I think it's never too much to keep hitting this point, right? Because privacy is so top of mind for everybody that I think is listening and all of us marketers.
>
> But I think, Alek, we mentioned two of the things. I'll say the first one is the moment you have a single source of truth, you know, this is very basic but very true, right? You actually have more control and more governance of that data and therefore you can increase the privacy of it, right? The moment you have multiple copies of that PII, now suddenly you're more vulnerable. So single source of truth in this composable CDP approach is one basic but very real benefit of working with this approach.
>
> The second piece of it, we touched on it a little bit with the identity providers of having it natively built in Snowflake. The fact that you're not having to copy your data and move it somewhere else again kind of for the same logic increases the privacy and ultimately supports this goal that every marketer has of protecting their customers' data.
>
> And then lastly, you know, I think there are very specific types of use cases that normally involve matching personally identifiable information with other partners or data providers where a data clean room is a good option, right? And Snowflake has made tremendous progress with clean rooms and continues to make progress by having leading media publishers like Disney Ad Sales or Roku or NBCUniversal build their clean rooms and their advertising platforms directly on Snowflake.
>
> So as a marketer you can actually join in their clean room and securely match data without ever exposing PII and vice versa. So I would say those three things are ones where this approach, you know, working with Snowflake as kind of that centerpiece of all your data, can really help from a privacy standpoint.

Totally makes sense. I think one thing to even hit on there, to that clean room piece, right, like when you think about the composable CDP it's an activation layer on top of a data source, right? We don't necessarily care what that data source is, whether that's your first-party customer data or a data clean room behind the scenes, right?

So we can start to solve all of these issues across publisher use cases or B2B use cases, B2C use cases, right? And it really doesn't matter what that data is as long as there's a concept of data that you need to activate.

Another question here: **Doesn't Hightouch store data like a CDP?**

> No, that going back to that Hightouch overview slide, we're just simply querying off of Snowflake and looking for diffs and sending the diffs up. That audience UI is doing the same thing right behind the scenes. We're just generating SQL, granted complex SQL for all of those different models, but it's really just querying off of Snowflake, looking for diffs, and sending those up. So that's for adding users and removing users as well.
>
> One of the things that we do now though is we will write back to Snowflake to really create this end-to-end loop. So let's say that you want to write your audiences back into Snowflake or you want to do, it's called sync logs, to see type of like QA, right? So if you're just syncing back logs on why certain rows updated, you're able to actually look in QA end to end.
>
> Or if you want to look at analysis on, hey, when this user was in this audience, what did they do? You can actually do more analytics off of this audiences.
>
> So it's a loop, but again we're not storing the data by any means.

### Closing the Session

Awesome. I think we are just about at time, so anyone else feel free to reach out to either of us. Again, we have a session next week Tuesday, same time, same place. Feel free to join then as well, as Andrew and the team will help give a hands-on demo of everything we talked through today.

Lorenzo, thank you so much for joining. It was a blast, man.

> Thanks a lot, Alek. Look forward to doing it again sometime.

Thanks everyone for joining you.