EPISODE 1: SOFI AND MPARTICLE

TRANSCRIPT

A: Mike, you have a very interesting seat given you guys send data to all the platforms that marketers are using. Where are you starting to see people spend more money, spend more time and attention?

M: Yeah. Well, I think that there's obviously been this huge story around the explosion of data over the course of the past few years. There's also been this explosion of places to send that data to and opportunities. Leverage that data to do cool things. You start to throw privacy into the mix, and it complicates it. You throw some of the changes that have been thrust upon everybody by Apple and Google into the fold, and it goes from complicated in nature to somewhat complex. To your point, the walled gardens were once a place that represented a bit of an easy button where you could do user acquisition at scale, and it was pretty cost-effective. But the changes with iOS 14.5, that kind of pulled the rug out from underneath everybody. So now what you see is just a ton of testing across a wide variety of channels. The other walled gardens, whether that's DSPs using contextual targeting or using other signals to try to engage users. Or they're just also making that migration from dumping a bunch of money into user acquisition and doubling down on just building better customer experiences and retaining users and kind of putting the onus on them, to say, "You know, it's harder to acquire a user, so I'm gonna do more to keep them around." Yeah. And yeah, and I think that's been a big, big shift.

A: Yeah. And I've doubled down on that. I think the story, the headline that a lot of people hear is, you know, Apple made privacy changes, thus people are leaving Facebook. But I think actually underlying that, there's something different. There's a different sort of theme that's happening that we hear from a lot of customers, which is, the larger B2C brands, you know, whether it's in retail or more considered industries in finance, realized that by working through those walled gardens, they lost all connection with the customer. They had no ownership over the customer. The customer acquisition happened through the platform. The winback happened through the platform, cross-sell happened through Facebook. They didn't own the customer. You know, Amazon being the other great example where a lot of times you actually had no ownership over the customer. You know, all the orders came through Amazon. In retail, people first learned the risk of this when Amazon started having white label brands. And I think now people are starting to see actually that the same risk exists for other industries where, you know, it may not be Amazon, but if you're on Facebook and you're relying on that to have that customer relationship, Facebook can all of a sudden send all that traffic to somebody else, and you're out of luck. So I think brands are realizing that having a first-party relationship with that customer is transformational, whether it's you over SMS or voice or email or through the website, they're changing their spend and directing their investment so that it's not only through those walled gardens for sure.

M: You can't outsource your relationships with your customers to somebody else. You can't rent your results. The center of gravity has fundamentally changed. And so I think all these things are a kind of orbit around the same general themes, but they're all part of the same overarching.

A: How do you guys manage it? How are you able to tell what's working and what's not?

K: There's obviously tracking the user and owning that owner, owning that user relationship, you know, is hypercritical and capturing that information of your users, you know, and really building upon that is critical.

I think that measuring across channels, it's a very tricky one. Right. I think me and you have talked about this before where you know the typical marketer is , you wanna start with a way of measuring. So you start with every, every model is flawed. You start with the one that's easiest first, right?

Yeah. And that's some type of touch base type of attribution. And I think a lot of, you know, companies appear, a lot of people on this call are very familiar with that and probably is how we, they certainly do today. But over time you start to understand it, look, when you're making decisions based on this model, this touch base, you are naturally going to skew your decisions based on the wall of garden channels.

The channels that do require a lot of data on their users, so they know how to optimize for that one metric, right? And I think over time, what I saw at Uber especially is, you know, we were making decisions as a model and it works well in the beginning you know, when you're trying to scale. But over time, your users' behavior change, and the type of audience you're talking to now are not the first product adopters.

Their behaviors take more time to actually move them to the consideration phase, and you kind of have to rethink the way that you're measuring and, and it ties all to privacy as well, right?  that touch base attribution becomes more and more difficult with the privacy changes. Not alongside the iOS, you know, Google now enforcing their GAI being anonymized being anonymized as well.

Cookies are eventually going away and I. Naturally, you're moving more towards a higher type of media mix model type of environment as well, right. To truly understand the impact you really have to understand the incrementality of your media holistically. And what you sacrifice is the granularity of things.

And I think that's where the industry is evolving to in the long term. And that's kind of how we progress as well. When I was. Especially when I was at Uber, where we actually went to a solution where we had a mixed model for really measuring the impact of our marketing dollars. But you leverage something  touch base attribution for more granular optimization.

A: I'll pick on, on one. Everyone now talks about SMS marketing. It's been very hot. I'll pick on SMS marketing a little bit. They're probably, I find, the most egregious offenders for taking credit for any user that happens to get an SMS from that channel. You know, anybody who's gotten an SMS, you know, and converts within seven days must be attributed to sms.

And I think the reality, you know, and that one as an example, is in many businesses you know, The user is getting a phone number and getting that SMS before anything else happens. So it's your point on , you know, sort of the funnel of that user and user behavior. So because the user is getting that text very high in the funnel,  if you would, you know, attribute anything that got a text to that channel, it's always gonna be wrong.

You know, it's, you're just gonna be taking conversions that really are happening through other channels or would've happened through other channels, if not, so a lot of what we work with clients on is how do you look at holdout groups to make. You know, you're not, you're not looking at it incorrectly and you're saying, Hey, what is the hold out group, you know, in terms of what is performance without SMS?

And then if you add in SMS, what is performance there in terms of lyft? Not in terms of  absolutely how many people saw an SMS. It's a crazy way to attribute the bookings. 

K: Sorry, just hopping on that, Alex, I think something that people always, you know, miss out on is really, what is your intention of the way you're measuring, right?  What is it really built for? You know, when it comes to your example, you give SMS taking credit, you know, the idea is that they want to know what's really driving what, right? So they have this attribution model that's flawed, that gives credit to SMS because of the interaction or another channel.

But when you really think about it,  what is this attribution model meant to change what decision you're meant to make for it? And most of the time, from my understanding, is also when it comes to budget allocation and maybe resource allocation. So when you really think about it, a lot of it's for budget allocation.

Would you really make a decision to allocate more budget to SMS and email? Would you allocate more to the channels that actually require a budget? As in , let's say if this, if this attribution said SMS and email was poor and didn't get anything, would you really stop emailing SMSing your, your team, your customers?

Yeah. And most of the time the answer is no.  Everyone that comes in, you still wanna build that relationship with you and engage with them. So you're not really making decisions based on that. And I think that's something that's really important when you decide what model worked for you.

Because the intention is what's really gonna make that decision. 

M: Right? Yeah, I agree. It's all connected too, right?  you can pull budget away from seemingly less performant channels, but then it has  a ripple effect too, right?  you, you can't get into this  a reductionist mindset because then all you'll be left with is retargeting and then  that will dry up eventually.

And it's , oh, well that's perfectly efficient. It doesn't work that way.  for sure. 

A: I mean, some of the channels I wanted to ask about explicitly  we've seen, so I remember when we were in the home services space, we saw the resurgence of direct mail, which, you know, for a while was considered dead.

Nobody wanted to use direct mail. And now all of a sudden you have Lob, you have Pebble Post, you have Share Local Media that with the shared mailers have become very popular. I think we've seen a lot of customers have success with this sort of more programmatic version. So to me it's  this funny trend now and obviously we're in the voice industry, which is bringing back voice.

So there's this funny trend of, what was dead is now coming back,  channels that, you know, became underpenetrated have become useful again. But what have you guys seen or what have you seen customers do in direct mail? Have you seen people use that in any interesting ways? 

K: I think the unique part of direct mail is how you layer on a lot of third party data sources to make your direct mail more effective. I think that the mailing mechanism, the distribution mechanism somewhat has evolved a little bit, but still very much is the same.

As far as I think that the unique part is now that bringing some of the learnings when it comes to incrementality experiment, how do you layer third party data? To enrich your DM experience, so you're targeting the right people. I think that those things have evolved and we're still continuing to see a lot more players coming in to help with that.

Right. And the beauty of direct mail is you do have, you know exactly who you're sending to. You can also do a holdout, right? I think that's the benefit of direct mail versus some prospecting channels you have, especially in the walled gardens. 

M: So I feel for some brands, direct mail never even really kind of went out.

It's still highly effective. Not for me as a consumer, because I'll just let my mail pile up and then I'll throw it out. Yeah, I'm totally that guy. I'm horrible, horrible with mail, but , you know, the majority of the country behaves very differently.

So I know for a lot of financial services brands, Direct Mail has been one of their most performant channels for a while. I wanna maybe step back though, because , I think one of the interesting phenomena that I think everybody is dealing with is that you have so many channels and so many partners and so many vendors to choose from.

So how do you make sense of it? And I think what happens inside every organization that we see is the growth tactics start to fatigue. So they start to outgrow certain partners or certain channels. So new vendors get added, maybe some get removed, event tracking changes. All this happens across  different platforms, and you have  different development teams own the different platforms.

Then you have  all of the APIs provided by these vendors. They continue to change and  these changes  start to compound on top of each other. And more campaigns get run and more experiments get run and the cycle repeats and you kinda , you drive a bunch of growth, you plateau, you do it again, and then you just kinda have to keep throwing more stuff at it.

And then when you take into account the notion that identity is very dynamic, you take into account some of the stuff we were talking about before, the changes from Apple, and Google. You take into account new privacy regulations. Everything starts to break, right?  your data breaks, your data scheme breaks, your data, pipelines break, your APIs break.

everything is in this kind of constant state of entropy. I think to even get to being able to execute across these different channels, you have to think about how am I going to adapt to this?  this data chaos that happens within the organization because the best laid digital strategies won't matter if everything is in this state of chaos.

So it's  you can't, you can't separate the activation opportunity from the operational gain, it's all kind of  one and one and the same. Because those things are what inhibit or undermine your success. And so whether it's direct mail, whether it's voice, whether it's paid ads or video or whatever it may be.

Think about how you navigate all of that change? 

A: Oh yeah. My favorite thing is when, as a marketer, I'd go into whatever the system was and try to figure out, you know, what was the abandoned cart event? What was the conversion event? And you have eight different ones labeled conversion event A. I've definitely lived it. I mean, you know, maybe dig into that a little bit. I think one of the things you and I have talked about in the past is, you know, historically people did not have an event driven infrastructure at all. That was not how companies were set up. They had maybe a CRM, they had a schema of eight different fields. You know, maybe they stored those eight fields somewhere and they'd run an ad hoc segment and they'd blast everybody in a one T-shirt size fits all message. 

And I think you guys are one of those serenities that have really been out front, you know, converting people to this event-driven system and to a system where you don't have some of these data issues. Talk a little bit about where you think organizations are now on their tech stacks?

What do you think are the big pain points that are causing people to now move to doing things correctly?  How's that transition going in the market? 

M: Ultimately,  what it boils down to is data maturity, right? The fact is every organization will face data quality issues at some point, probably perpetually.

Every organization has to deal with governance and compliance and privacy, regulatory requirements. Every organization will want to get data from a bunch of different places, sources and systems. A bunch of different places, APIs, applications, destinations, you name it. And that chaos that I was alluding to kind of makes it hard, right? 

You cannot just create this nicely laid out, highly symmetrical, beautiful reference architecture and say “ah, I've got it all working.” Those things almost never really play out because of what happens inside every organization. 

So the path to data maturity is really about how well you can adapt to change. We talked to a lot of companies, big and small, domestic and international. You can be a big company or a small company and be very immature or moderately mature, or you can be more , more expert or, or advanced.

It's really about the culture and the people that you have. But the things that we try to uncover the tools are one aspect of it. Knowledge is one aspect of it. The process the culture is, is certainly an aspect of it. You start to get into the weeds a little bit and it's , do you have

a customer data strategy? Do you have a strategy around identity? Are you thinking about governance and protecting data quality? How many different places are you trying to get data from and to? Are you operationalizing any type of models?Is it fairly simplistic  customer LTV models or more elaborate machine learning models?

And ultimately the goal is to help people become more mature and, and certainly the event driven streaming architecture really helps with that because you have to be able to move with speed and, and agility and market in the moments that matter. But there's no data pipeline or data platform that will take any company all the way there.

It's also about change management and just being able to embrace a newer, better way of doing things. 

A: Kevin, you and I have talked about this before. You now run a very big organization,  how do you start managing, you know, across all these different vendors you have and all the different data sources?

K: There's always a cost associated with it. I think ultimately, it's all about priorities. And  realistically, migrations are expensive. I think that Mike mentioned a really good point is that it’s cultural too. 

What is, what is the priority and how is that, how, how is our perspective when it comes to having those different systems? Because there's also preference, right? When you're working on so many different functions, cross functions, there's also preference on top of it. And I think that, you know, it's really associating.

The benefit that we would gain from this, what is the cost, which is the migration of it, you know, and how quickly are you growing? Right. I think that those are all contributing factors because I'll say that at Uber, I experienced this as well. We went through a lot of data overhaul, and realistically, it's because when you're growing  a hockey stick, your tech infrastructure cannot keep pace, right? 

And in order to satisfy some of this, you are adding an incremental solution to it that, you know, is probably gonna make things more difficult in the future, but it's what's needed at that moment. Right? And it's more about prioritizing, , hey, do we want to take a stop? Do we want to take that effort to migrate?

Or is it actually something that, hey, we need to prioritize our growth right now. We need to be able to solve this problem. And this takes precedence over that, right? Yeah. And those are the decisions that are happening perpetually, and it will happen perpetually. 

A: Yeah, it's hard for sure. Yeah.  I said, I remember, I mean, every organization I've been in has issues with data quality.

I think it's just a perpetual thing and very difficult to overcome. I wonder about your perspective, Mike, in your market, as you think of the role of the CDP. SoFi uses mParticle, Regal and Braze to move data around into various channels. 

I see Segment now extending beyond just being CDP now with what they're calling Twilio Engage, they're actually starting to send SMS and send email.  

The Iterable and Braces of the world are starting to look more and more like a CDP in some regards.  they obviously have not built all of the integrations yet. They certainly have the infrastructure.  how do you think that market's gonna go?  are there gonna be standalone point solutions, point being a relative word that are sort of CDPs versus the customer engagement software?

Or do you think that you'll start extending across actually engaging the customers well through channels or other people in the CDP space? 

M: It's a great question and something we have to address quite regularly. So as you can imagine, I have some views on this one.

If we take it back to one of the things that we were talking about before, which is all of the change that's happened in this space. Especially the need for brands to be able to own the relationship with the customer. Well, what does that require? It requires building a really strong foundation rooted in first party data.

So having a first party data strategy and just commanding ownership over their first party data. What you've seen is  CDPs have emerged as the leader in helping companies think holistically and in some ways objectively about leveraging their first party data.

And you'll typically hear them talk about three parts of it. It's  ingestion, segmentation and activation. And the value is connectivity across  partners and channels. It kind of feels like everybody's a CDP these days.

If you listen to the garbage from the modern data stack, it's  everybody's CDP killer these days. 

Let's talk about the job that we are ultimately hired to do. It's to help unify data, protect data quality, and instantiate a layer of governance so that you're not out of bounds from the perspective of privacy and compliance. And then be able to connect that data really anywhere.

And when you start bundling together campaign execution capabilities, it eliminates that neutrality. And then you start seeing channel conflict. Inherently,  a platform can't be a feature of an application, right?  the world doesn't work that way. We think about ourselves as infrastructure.

I think that us versus Segment,  those days are pretty much over because they've been so intertwined with the Sendgrid team. And it makes total sense for Twilio, but does it make sense for their customers and for the market? I mean, we've had a ton of customers switch over from Segment mParticle over the course of the past couple years because of that.

Yeah, so, you know, this is, this is just my view, but , even, even the part around activation and being neutral. I go back to one of the things I was, I was talking about before, which is , how, how do you focus or  how do you, how do you help people address the need for greater adaptability and resilience and dealing with  all the stuff that's going to break so that you can do the activation part. 

And the campaign execution tools don't solve for that. The reverse ETL tools don't solve that. Most application CDPs that just focus on audience insights activation – they don't solve for that. But you can't just pretend that everything is  this steady state of calmness and serenity.

Things break inside every single organization. Every single day. And if you're not dealing with that, you're not actually addressing the problems. 

A: Kevin, you've been on the other side of this as a buyer, how do you think about building out your tech stack? Do you want one unified application that does everything? Or do you have multiple point solutions where email marketing, phone marketing and CDP or in different pieces?

K: It's clear the lines between these different technologies are blurring more and more, right? Because companies recognize that there's so much synergies and efficiencies to be gained for having some certain consolidated systems. But in reality, when you're looking at all these different systems, there are certain systems that do things better than others, no matter what. Because they have the focus right in that area, and they're focusing on those integrations. They're focused on those features and having a system that does everything. 

In an ideal world, I'll say yes. But in the real world, do I believe one can do it well in every area? No. And because of that, I feel  you'll always be in a world where you have some solutions that can do multiple things, but then you'll still need some point solution systems.

And I think that there are certain things that you need to recognize that should be managed in one place that can then speak to these other systems and some of the things that aren't required. I think an example of this is managing your customer journey. You can manage your customer journey in your ESP or Braze or some of the system.

You can then manage a different customer journey in another system. Regal is an example of that. You can have a specific journey there as well, but if those systems don't talk to those journeys together, that becomes a challenge. You're not really building a holistic lifecycle journey. 

Versus having something like a CDP  mParticle that could be the master of your journeys, you know, that speak to the different systems. Long way to answer your question, but ideal world, yes, but realistically, I don't see that happening anytime soon. 

M: Well, Kevin brings up a great point too, which is that no individual application will have all of the data to map to the entire customer.

So the entire customer journey goes beyond just  the marketing channels. There's on-site or in-app personalization. There's the required measurement. There's the customer support tools,  Zendesk for example. You have to get data into things like your analytics and BI stack. 

And so no one individual application is going to have all of the data that can do it holistically. , and that's, that's part of the value that you get from CDPs. And it's also why we've been so steadfast in saying that we are not going to get into the execution layer. We are not going to become a campaign activation system because while we would have every opportunity to do it, if we were going to do it, we probably would've already done it by now and maybe it would make some people's lives easier, primarily  marketers. But customer data infrastructure is for the whole business, including marketing, but not limited to marketers.

A: Let's pick on Twilio a little bit more as, as they're not here. But one of the questions I always sort of find interesting, you know, when you're running these big marketing orgs is for certain channels, do you use off the shelf products or do you go to a provider  Twilio and build it yourself?

So as an example, do you go and pay for a SendGrid and build capabilities on top of SendGrid, or do you use an Iterable or Braze and use that off the shelf tool? I'd say by and large, I've found that the more successful orgs end up using off the shelf tools so they don't have to invest engineering time in building a journey builder and connecting with these other systems. 

It feels to me  sometimes,  organizations, they wanna believe they're unique. They wanna believe that they should build it themselves. I guess you see it firsthand, obviously from mParticle because you know, I'm sure you have customers, some that have built their own custom systems and some that are using off the shelf ones.

How do you help guide customers as they're thinking about which way to go? 

‍M: Yeah, I mean, build versus buy is kind of the, you know, it's the, it's the decision that I think every team has to, has to make. Probably continually, and, and one thing I would say is what we see inside a lot of organizations is  this, this natural trajectory where in the early days of a startup, they'll start with  a bunch of third party tools and it'll work. 

And then eventually they start to outgrow those tools. They're , “screw this. We can build this. We have engineers and it's , all right, no more third party tools. We're gonna own everything in-house”. And for some reason, they go down that path, but then they hire a new head of product or a new executive.

And they come in and they're, “Well, why are we spending millions or tens of millions of dollars building systems that we can just buy off the shelf?” And it becomes a CapEx versus Opex conversation. And then they start to slowly re-embrace third party tools and it's this kind of weird S-curve.

And what I've found is everybody kind of has to go through that journey on their own, at their own pace. I can provide them with historical context and we can provide them the pros and cons, but our job is not necessarily to make those decisions for them. But ultimately it's about getting data from wherever they want to whatever they want.

And most of the time there is going to be internally built tools and third party tools because they'll never want to build mobile measurement. Because they have to get access to certain APIs, which just aren't readily available. I think in most cases though, especially when you're looking to connect data to lots of different systems, it's not just  a builder or conversation. 

It's a build and maintain versus buy. You think it's a great idea to go out and you throw a handful of engineers, but then everything starts changing and everything starts breaking. All the stuff that I've been talking about. And the next thing you know, you have  a small division just to keep up with  your internally built stack, which will never move at the pace of innovation that you'll see from third parties who are focused exclusively on that and nothing else. 

A: I have a couple more questions, but just for the audience, if you have any questions, put them into the chat and, you know, I can ask those afterwards so, you know, feel free also to direct them at specific people if you .

We have, you know, approximately, you know, 25 minutes left. So, if you have questions please, please add them. So in this industry, both of you guys have been very successful, you know, obviously on, on different sides of the quantum one, on the vendor side, on in the marketing industry. And one thing I wanted to ask is , what do you think has made you guys successful in this marketing industry?

Where were there certain people who really sort of pushed you, where there are certain sort of ways of approaching these problems that you sort of think you use differently or led to success? What made you guys successful ultimately in this marketing industry? Kevin, if  you wanna go first.

K: There's no one,  one size fits all type of way to be successful here. I think that it's more from your approach to your marketing efforts. I think that the one thing that has stuck with me is just being very customer obsessed, and living by those principles. 

You want to provide the best experience possible for whoever you're reaching, wherever you're reaching them.  It's all fueled by that idea that personalized experience, great experiences, great storytelling.

Great customer success is what's going to get you your business. And then everything will fall in line. And you'll apply it differently depending on the business that you're working on, on the audience and the people you're trying to reach. But knowing that at the end of the day you want to drive the best experience possible that'll lead you to the decisions that you're gonna make when it comes to technology, messaging and everything else. 

A: Was there anybody in your career that you really credit and you say  that, you know, working for that person really , taught me something got me there?

‍K:  I'll credit everyone that I've worked with, and especially in the Uber organization where I spent a lot of my time. The leadership there, everything from the general managers there to the early folks leading that growth marketing team, has really enabled me to spread my wings and experiment and learn by trial. 

A: Assume people are earlier in their career as a marketer, what advice would you say to them? 

K: Most performance marketers start with one particular channel. They become experts in that channel and they expand beyond there. But then you really build your habits, your tendencies on yourchannel being very organized, making decisions based on some type of attribution model. Then you expand with that knowledge everywhere else. But what happens is you box yourself in.

Because you're actually very comfortable. Because guess what? I can take this action, I can tie it directly to some data that shows how well I'm doing. And you kind of get into a comfort zone when you're doing that, because you can prove what you’re doing every time. But then that means you're not creative, you're not thinking outside of the box.

It's great to understand the data model, but don't rely on it. Don't build bad tendencies and really try to always think outside the box, no matter what the data says.

What are the customers really saying? What do you think makes sense to the customer for your product? Right? And really think about that first versus the data, and you can then look for data to maybe help inform your decision, but don't box yourself .That's another thing that I was lucky enough to experience early on at Netflix.

I'll say that when I joined Netflix, I was there at a pivotal time where we were deciding to become an entertainment company, right? Yeah. And guess what? Every model that existed at Netflix at that time told you not to do it right. It told you not to change your market. You know, it would tell you to continue showing these red banners with a yellow button that says “start your free trial”. 

And if you made decisions based off of that, Netflix would never become where it is today. So that was my first learning experience. Understand data, but don't don't rely on it, really think about your customers, think about their business, think about your product, you know, and really make sure you're able to do things outside the box. 

A: Mike, you've built a career sort of on the flip side of the coin where, for 10, 15 years, you've now served marketers and helped them achieve their goals. What advice would you give? 

M: Pretty similar to Kevin to be honest. Surround yourself with great people and optimize to be able to learn.  

Whether that's more in the, in the micro, if you're running a campaign or, or an experiment, as long as you have the structural capacity to be able to learn from it. I know everybody wants to, to move up the ranks and everybody wants to get promoted and bigger titles and more money.

All that stuff comes if you have a growth mindset and if you're just focused on continuing to learn. The most successful people that I've ever seen, they just have an insatiable intellectual curiosity that is never ending.

And so become a learn it all. If you do that, there will be no shortage of opportunities for growth and advancement. And over time it'll definitely be lucrative. 

A: We can be futurist for a minute as to what's gonna happen, and I always think it's interesting to think through what the next five years will look like as we start engaging with customers. 

If I had to make my prediction, five years from now you know, I'm a little bit negative on brands that are just using existing platforms to reach customers.

If you're only using Shopify to reach customers and you're only differentiation as a brand, , I'm negative on that. 

I think the brands that really survive will be the ones who learn how to build a differentiated experience by customizing the flows or build differentiated service experience because they do something special where they're, they're differentiating beyond what exists.

Banking is a great example to me. There's now so many startups that allow you to build a bank, and people are only differentiating on brand, and I think the great ones will have to do other things.

So I look for who these companies are that are doing something unusual outside of just the brand or the channel. 

We can start with you Kevin, if you think forward five years. What do you think is gonna be different five years from now, than where we're at today?

K: Yeah, I think tied to some of the things that you just said, I'm aligned with some of the things you mentioned, Alex. I think the only thing, I'll take a step further is I think some of it is happening today where, I think that in the future people are recognizing that the brand and the product experience and everything is the same thing.

There's no single team that owns the brand. You know, marketing doesn't own the brand. You may have a brand team and they definitely are putting a lot of effort into it, and they do influence the brand. But guess what? When you come into a horrible experience, that impacts your brand. You come in, you're taking an action and what you expect to happen doesn't happen.

That impacts your perception of that brand. Right? So I think we're moving towards that direction where everyone's recognizing that, everything needs to be more integrated because everything influences the brand. The product and brand goes hand in hand. I don't believe there will be a one point solution that can solve all that together.

I think that people marketers are being more cautious around how these systems are talking together? How do we make sure that this doesn't break the experience for the user?  

I think solutions  mParticle, you know, being more of the heart of all those systems, you know, are some of the things that are gonna help us get there. 

Especially when it comes to identity with web three and decentralized identity, how is that going to impact things as well?

M: I think one thing that will forever be true is this kind of back and forth rotation between bundling and, and unbundling. Aggregation and, and, and disaggregation. It’s tough for me to say exactly how things will play out in the world of marketing.

I can give everybody a general sense of where we're going. I've talked a lot about helping teams build resiliency through data quality and through solving data governance requirements and using that as the foundation for activation.

We've been in this, in this world that's , that's become highly fragmented, right? So  the digital experience was once single threaded. And now I think  the, one of the latest, most interesting stats I saw a number of months ago was  the average house has something  a hundred different connected devices.

So think about just  all of the data that's being transmitted, even, even passively in, in our daily lives. So there's this need to aggregate and, and, and wrangle all of this data and kind of shape it together and make it into this nice kind of pretty picture. But it's all a setup for what are you gonna do next?

How do you actually help teams?  Making decisions based on that data drives better personalization. Use that to make predictions and, and, and automate outcomes ultimately. So, for us and our customers, we try to provide as much flexibility as possible, but it requires a whole bunch of manual configuration.

How do we start to automate a bunch of that stuff? How do we surface insights and anomalies and opportunities proactively? So people can get ahead and stay ahead of opportunities or issues as they may arise? 

So you get into things , certainly as you connect a whole bunch of different channels,  next best action or optimizing towards  the next best outcome, whatever it may be.

I think that's the natural evolution that I think we've, we've long envisioned, and I think this year is the year where we start to really put our foot forward and march towards that opportunity. 

A: Next best action. What channel do you prefer? Now it's Monday to Friday during the workday.

Look at what channel's. Or now it's the evening, or now it's the weekend.  how do you actually know? 

M: For me, it's not direct mail.

K: Totally agree on, you know, the data side, the machine learning, AI driven side on how we're targeting and how we're reaching people with the next big action. I think in addition to that, in the next five years, I think the whole idea of incrementality applied to that as well.

Because you're seeing that shift right now, right? Propensity modeling targeting based on propensity model has, it is not that it was formed before, but has really picked it up the last couple years of people trying to do more and more so, but I don't think the incrementality of that is something that's thought about as well, because a lot of people are associating with, “Hey, based on their actions, high propensity, you should target them here, et cetera.”

I would actually challenge that those people are the ones that are actually less incremental because a lot of their actions they're taking is showing their hyper propensity and you should actually be teaching the people that have lower propensity and how do you drive their actions and convince them.

M: Right. Amen. 

A: You guys also started mentioning a little bit of AI and machine learning, maybe I have a minority viewpoint in this. Where I've seen AI and machine learning be very successful is with companies  ASAPP in the customer service space where they go work for the largest  customer service centers in the world.

JetBlue has 20, 30, 40,000 customer service representatives and there's enough data that they can start automating some of the responses and doing the next best actions. And where actually the company's savings are big enough that they can afford all of the cost of machine learning. 

Because actually,  the badly kept secret of machine learning, it's a services business, not a software business, right? The margins are not software margins. It's actually very expensive to keep training these models and updating them constantly. I sort of have this opinion. The same thing will happen within marketing.

Where the largest organizations are able to access these very expensive tools to optimize marketing and do this, but it's gonna be very hard for small organizations to do it. So there's gonna be an exaggeration of the data moat as this becomes possible for the largest organizations. 

But you're going to be at a disadvantage if you're not Netflix. If you're the next one coming up, you're gonna be at a huge disadvantage because you don't have the quantity of good data or the resources to be able to go and spend on machine learning.

I think a lot of people talk about machine learning, machine learning, machine learning. I think it's gonna be hard for a lot of organizations to really go and do it. And even if  Braze has machine learning, it's not gonna work for you unless you have enough data within Braze to go and do something with it.

M: I guess, yeah, there's, there's ways I think to, to get around that. I think the thing that may hinder some of those opportunities may be the privacy and regulatory issues. 

A: The model that you'd wanna have is, Braze should be able to say, let's look at profile data across all of our, all of the customers that use Braze and do machine learning across all of them. 

The trick is you can do it in an anonymous way, but if you wanted to do it in an attached way, Actually, you're technically under the new rule around selling customer data. It's not actually selling data, but you're now sharing it across organizations. Every single customer of Braze would have to ask permission of their customer to enter them into this machine learning model so that they could be learning cross customer.

So I think there's data privacy, even in the United States challenge and in Europe. Forget about it. I don't think it's ever gonna be possible. 

M: Yeah. Yeah. I mean, you, you could be right. There are, there are ways to, to do it from from the standpoint of having it at  at, at some sort of  abstraction, but even still, you know, you run the risk of kind of being or crossing the line from being potentially a processor to a controller. 

Everybody points to CCPA. But there's actually 17 states right now that have either passed consumer privacy acts or are in the process of passing consumer privacy acts. So I think within the next five years we're probably going to see one mostly from, from every state, at least every state maybe except for  Florida and Texas. So 48 outta 50 states. 

A: Well, when we think about the future, my co-founder Rebecca and I have made a big bet on  talking to their customers.

We think that's the channel of the future. If today everybody talks about SMS marketing, now, five years from now everyone will be talking about talking live with their customers. If you guys were making a bet, what is every marketer gonna be talking about in terms of the channel to engage with customers?

M: What becomes old becomes new again. What becomes new, becomes old again. Everything is cyclical. I actually see there being more channels, not just  one individual single channel. Connected TV still has a long way to go. It's still very much in its infancy.

I think that if there is something that's really unique and disruptive, whether in the next half decade or decade. I think just broadly speaking, I'd probably be willing to bet that it's a new channel that looks maybe nothing  what's available today.

K: Mike mentioned the old channels become new, new channels become old. I think that one idea that is going to stick around will be owning your original content. And I'm not saying just original content from the entertainment side  Netflix, but user generated content. 

Your own content is really providing value for your customer base, but also the expanse of your customer base, your future total addressable market. Because you're providing value for them, in which case you would actually generate a lot of volume and people are coming to you because of the value you're providing from that content and now give you a really true, sustainable, organic growth.

Things that drive programmatic SEO. Those types of things. I think a lot of companies are gonna start investing more and more in that, knowing that that's how you're going to continue to drive sustainable, organic growth? Versus right now you're investing a lot of paid dollars.

Everyone reaches that point where they're diminishing returns and people are trying to find ways to kind of go beyond that. 

A: Yeah, I agree. I mean, part of this is yeah, because Google keeps getting more expensive every year. People are searching other places. I'll steal an idea from my co-founder.

One thing she thinks is fascinating that could happen in the future is you can start having existing customers talking to new customers. So similar user generating content, but go one step further. They're not just generating content. 

Patagonia is a brand that I love. So would I as a Patagonia customer be willing to talk to a new Patagonia customer about my experience and what relationship do I need to have with Patagonia in order for that to be possible?

Is it, is it  I'm a member or is it I have access to certain content? Is it financial? Do I have discounts? I don't know. Maybe I just get something out of the fact that I'm sharing my experience with other people. And , that's a community I want to connect with. So I haven't seen it done successfully in too many places. It's how you use your existing advocates to talk to your new customers.

K: You can do it in a lot of creative ways, as you mentioned. You know, even when you look at a lot of productivity tools out there. There’s things that support you from building certain documentations to things that help you build templates. You can leverage your existing user base and say, “Hey, can you publish this to our existing user base?” 

Publish this publicly, and build a system and now your users, as they're using your system, they're able to actually generate that content for you. In which case you can grow in a more relevant way. 

A: Yeah, agreed. Well, I think we're running up against time, so I want to first thank both of you for taking the time to do this. It's wonderful to have you all in the same room for once. And hopefully we can do it again a year from now or two years from now and see where everybody is.

Thank you so much for your time today.