The Key To Customer Engagement: Consent and Iteration

Seth Familian: So I feel like the two hardest things when it comes to customer engagement are, one, be willing to pace yourself and be iterative instead of trying to boil the ocean and two, having all the right people in the room at the start to align on what the minimum viable data set looks like so everyone can be satisfied and then continue to align them as you move in to more advanced use cases more sophisticated profiles, etcetera.

Speaker 2: Hello and welcome to Good Data, Better Marketing, the ultimate guide to driving customer engagement. Today's episode features an interview with Twilio Segment's Matt Smidebush, RVP of Global Customer Success Programs and Seth Familian, Director of Global Advisory Services. But first, a word from our sponsors.

Speaker 3: This podcast is brought to you by Twilio Segment, looking for clean, reliable data that you can trust, Segment collects, cleans and allows you to activate your data in real-time across hundreds of applications and channels, learn about how Segment can help you personalize customer experiences by visiting Segment.com.

Kailey Raymond: We live in a hyper-connected world, technologies are rapidly evolving and often becoming extensions of how we live and work, this trend towards a more blended reality that crosses between physical and virtual, shows no signs of stopping, and with this shift the way that businesses and consumers interact is undergoing a massive change, consumers are demanding more privacy, more personalization, more omnichannel. Today, we're getting in the weeds and talking shop with two of Segment's own Seth Familian, Director of Global Advisory Services and Matt Smidebush, RVP of Global Customer Success Programs and learning about identity resolution, the power of iteration and building data champions.

Kailey Raymond: I'm really excited about the show today, and we often get asked about this from listeners and a lot of people and our customers as well is answering the question of how does Segment use data to influence customer engagement and customer experience. And what are some of the outcomes of their own customers, how are they reacting to a lot of the trends and customer engagement and customer experience, how are you supporting them? And so today, I have two experts in the room, both of which have been with Segment for nearly five years and supporting customers in different ways. I have Matt and Seth here with me today, but I'd love for each of you to introduce yourselves. So Matt, do you wanna kick us off? 

Matt Smidebush: Yeah, sure. So I think I bring us under the average around five, Seth has been here longer, but I lead a couple of teams within our customer success organization, and we focus on delivering programmatic experiences to our customers and with our main goal is to get customers to value as fast as possible, and of course, we use Segment to do this, so I remember when I first joined, I had a company, all hands raised. My hand, I was like, who own Segment here? I wanna talk to you, and that kind of kicked off how I thought about using Segment for the next several years.

Kailey Raymond: I love it, you quickly, I'm sure became the owner once you raised your hand for that question.

Matt Smidebush: Well, I think we'll get into that a little bit later. I would refer to it as a Data Council maybe.

Kailey Raymond: I love it, I love it. As data should be, it should be owned by a couple of different teams making it democratic. Seth how about you.

Seth Familian: So I've technically been with Segment for a little less than five years, but I've been implementing Segment for almost a decade at this point, and so the back story is that before I joined Segment, I was running my own data strategy consultancy, and for about five years in that 10-year period, I was doing growth hacking for my clients, and I discovered through one of my clients actually, Segment was a really, really easy way to get data flowing into downstream tools for both analytics and activation, so those were the services I would sell my clients where I would do an audit on their Google Analytics instance or whatever data they had flowing at that time, told him how to re-instrument in Segment, we got the data flowing, we analyzed what was happening, we have to be bid in new campaigns, we watched all of their KPIS move into the right direction, and after doing that for a couple of years, I realized, Wow, this platform is really amazing, I really should just go stop running my own thing and doing this Dev and just go work for them. And that is exactly what I did. I actually called up a couple of people who I had been collaborating with in that time that I had been doing instrumentation and I said, "Hey, can I think I'm a Solutions Architect?"

Seth Familian: And they were like, "Wait, you wanna be a solutions architect? That's great." So then I joined and interestingly, pretty quickly, I went from being a solutions architect to falling into much more complex engagements, I sort of just reverted back to a lot of the strategy work that I was doing when I was doing my own consulting work, and so sort of naturally, over a two-year period, I ended up doing these advisory projects that were a lot less of upfront implementation guidance and a lot more of stepping back, big picture, what does the architecture needs to look like, and after I did them a couple of times, like I and my boss, we're thinking we should build an advisor practice, and that's exactly what I did over the last three years.

Seth Familian: So it went from me to being a team of four people with a portfolio of offerings that focuses on upfront architecture strategy, so helping our largest customers think big about how do we plan a multi-year deployment across many BUs, or many different apps, and then that strategy would lead to helping our other solutions architects have a nice plan where they could then go do implementation guidance for the first app or the second app.

Seth Familian: And then my team also does these things called architecture and advisory or schema audits, where once a customer is established, inevitably, their workspace becomes a little messy. Matt saw this probably in the instrumentation second itself, I've seen this over and over again. I like to call it entropy, has entered the workspace because that wonderful...

Kailey Raymond: That's a good word for it.

Seth Familian: Yeah. That platonic ideal of the five event tracking plan quickly spirals out of control as more and more people get excited about contributing events into the space. And so what we help our customers do is inventory everything that they've sent in over the past, say, 30 to 60 days de-do bit basically say, okay, we see the mess of what's there, let's build a new golden schema and then let's do a mapping exercise and reconcile and come up with the clean approach, and then talk about strategies for programmatically, how do we clean things up, and then practically down the road, how does your engineering team really fundamentally change things to make it better? 

Seth Familian: And then finally, my team more and more has been doing thought leadership, so stepping back and not just saying, how does the schema or the architecture need to change, but what is a framework that can actually inform how we talk about customer data maturity. And so that's been sort of the latest iteration of what it's been to run the advisory team, and I'm happy to talk more about that and the customer data maturity model as part of our conversation.

Kailey Raymond: I love it, yeah, I'm very excited to talk of it. Seth literally just wrote a book about this. Like no joke. And so very excited to get into that a little bit later. Actually Matt, I don't know your background, what's your journey into Segment before getting into the Customer Success area? What were you doing? 

Matt Smidebush: When I interviewed here, and then as I've told the story later on, I said all of these different pieces of my career were building blocks to thinking about the customer journey programmatically, I've done a bit of sales, I have worked for a data company previously in Customer Success, I've done product marketing, I was an analyst, and so all these things together, you have to learn how to tell the right narratives, you need to know what data to access, maybe be a little dangerous with SQL, and all of that together gives you a picture of, okay, what kind of experiences can I build along a customer journey, can I understand the customer journey, what are the most pivotal points in that customer journey to influence, and then what can I do to actually influence them at that point, and then measure it because all of this is about iteration.

Kailey Raymond: I love it. I always say careers are curvy, or at least they can be, and so being able to gain a lot of different experiences really does build that well-rounded empathy in so many ways for what your customers are going through, what your stakeholders are going through, and ultimately allows you to see the business a lot more clearly, so for both of you in your work, you're kind of at the very tip of the sphere interacting with customers every single day as we're starting to implement Segment to make sure that they're getting their data collected and cleaned and flowing into the right places, solving their needs and their use cases, I'm sure through that lens, you get to see a lot of trends coming in and out of customer engagement as a whole, and I'm sure that's changed over time, but I'm interested in this moment today, we're in a very interesting macro climate, especially in tech, I'm wondering if you're seeing any of the macro trends relating to customer experience today, what are those? What are our customers thinking about, which ones are you following closely? Walk me through that. Whoever wants to go first can go ahead and go.

Seth Familian: I'm happy to, Matt please, I would just like to say by the way that, Matt is one of the most empathic people I know. So in having the curvy career, building that empathy, I've got to experience it first hand. It's been awesome 'cause you are able to see things in a really cool way in terms of how our customers would experience them, etcetera.

Seth Familian: So there are some fascinating trends that I'm seeing, one of them is fear of this evaporating cookie. The world's going, cookie less. And what's fascinating is that, well, actually, 75% of cookies have already been eradicated from browsers, and so really this is just a last mile, and yet a lot of advertising models are hanging on to these old school ways of doing cookie based fingerprinting and stuff like that. And it's almost like there's this weird Armageddon moment that everyone is afraid is gonna happen, and yet it's already happened and we're sort of past it, and so it's now time to just get ahead of the curve and start measuring first party data the right way and start putting analytics.js or whatever your instrumentation Library is on to the page via your own subdomain instead of trying to load it sideways. But it's really this combination, I think of cookies going away, combined with an increased focus on privacy and consent, I think is the other big trend I'm seeing.

Seth Familian: And in combination with everything becoming less addressable, I have fewer email addresses, I have fewer anonymous users I can target. How do I do more with the data I have? I think that's the big challenge that a lot of our customers are facing, and it's been really fun to try and help them think through that challenge.

Kailey Raymond: I love that. Yeah, we hear about that all the time. Matt, what are you seeing on that side.

Matt Smidebush: I think that's totally right. I was gonna add on to that, that there's also challenges with how much money you can spend to acquire a customer. Everything is getting squeezed, right? And so when you have fewer users to target, there's pipeline pressure, if you're in B2B, there's attention challenges from a user perspective in B2C and user retention challenges, what data you have access to and how much money we can spend more efficient? So all these things are combining together, I think, in that pivotal moment.

Kailey Raymond: It's interesting and too, we talk about this often, is these trends... You just talked about privacy at the opposite end of that, consumers are also really demanding personalization, marketers in some ways think that there's an inherent conflict there, and I really, really don't think there is, as long as you're collecting the right data in the right ways, and to your point too, is like, yes, Matt, I think people are going now deeper into their base to be able to respond to it somewhat in some instances, a scary macro situation, trying to make sure that they are being as efficient as humanly possible. And personalization is a huge way to be able to do that, so... Yeah, those are the... So those are two big ones that we hear about all the time, anything else that you hear, especially, is there a difference with staff, self enterprise and other bands, or is that...

Seth Familian: Oh yeah. Well, yes and no, I think a lot of our customers, regardless of size, face the same problem of having nice, clean, consistent data resolved into nice holistic profiles that you can then activate in a cost efficient way. I think that's true across everything, and I do fully agree that the ROAS piece as Matt described it is all and more important today more than ever, the good news is as you have fewer targetable users your denominator drops which means just convert at the same rate, your conversion rates go up. How fun is that.

Kailey Raymond: That's a fun math problem to solve.

Seth Familian: Yeah, isn't that great? You know seriously though, in that particular piece, it becomes really interesting to think about when am I over-advertising to someone, or how do I manage the number of touches that I make so I'm not overwhelming someone with a message and then potentially an irrelevant message and then one more thought on personalization before I shift to the enterprise piece is, I actually think that personalization is... This is a terrible analogy. Is it a four-letter word or it's just like there are so many different types of personalization, and it's a fallacy which is just because you can't do predictive next product to buy modeling on top of the data you have to show someone the next thing they should be purchasing, doesn't mean you can't do highly relevant and impactful personalization on top of your data, that to me, I think is the issue where it's as simple as making use of high frequency data that you have about people and in the categories they engage in, and just showing them relevant content around those categories.

Seth Familian: It doesn't have to be hyper-complicated for it to be effective and therefore welcomed by consumers who are demanding more of it, and I think part of where the challenge is, is our customers might think initially, oh, well, if they're demanding more personalization, I have to go as deep as possible into every facet of everything they've ever done, and then that does feel like a privacy violation, but you don't even need to go that far, so long as you show you're listening and you reflect back, now you've cultivated a meaningful journey.

Seth Familian: So one of the things I'm seeing that more and more enterprise customers are doing is they are coming to us, and this is true in general, a Segment where one of our biggest competitors for many years has been home-grown solutions, and with our largest customers, I'm actually noticing a lot of them, and I've worked on a couple of big migrations over the last six months, they are migrating from their own home-grown CDPs, recognizing that they've cobbled things together or over a decade, and they really need in the name of operational efficiency, cost management and just to focus on their core competency, they shouldn't be managing the ingest of all their data, they shouldn't be doing all the Kafka pipe lining.

Seth Familian: And why I think it's especially interesting for our enterprise customers is you have deep expertise that's been developed within each of those customers around those things, and how do you change, manage the organization, so the people who built those things still feel relevant and valued because they are...

Kailey Raymond: That's an emotional conversation.

Seth Familian: Yeah, it's an emotional conversation, and then it's also critical that you don't lose the people who built those things because they're the ones who think creatively and wisely about how to leverage data, so how do you shift their context to say, ow we have a tool that could do a lot of the stuff that you built on your own, what do you wanna do with the tool? And to me, there's effectively this huge paradigm shift happening where enterprises are now finally embracing platforms like Segment full service CDPs, and then they have to change internally to catch up with the fact that they no longer need to build all this infrastructure on their own. So now what are we gonna do with it and how are we gonna innovate with it.

Kailey Raymond: Yeah, A buy versus build a classic scenario, yap absolutely.

Matt Smidebush: I love that, like how do we think about the people that are left that build these things and the value that they add I agree, I think they're incredibly valuable. I've seen this at Segment, using Segment internally, for this to be valuable, the data has to be used and in downstream tools, and basically the limit is the creativity that you have to how to use that data and ideally across channels, across teams for the more crossing use cases, the more valuable your data is. And those are the people that can enable that. I'm thinking about cultivating data champions across data organization.

Kailey Raymond: We've been talking around the idea of how challenging a lot of this stuff is, we haven't really identified it, but like what you're talking about is really, really hard for companies to do. We're reconciling vast amounts of data, stitching it together into profiles to say like, these are the activities that this one individual is doing that makes them unique, being able to target them in nuanced audiences by doing that. I imagine that in of itself, the silo might be the answer to this, but what I'm curious to hear from both of you are some of the biggest challenges that you're hearing from our customers around what they're doing for customer engagement, what's so hard about customer engagement for our customers? 

Matt Smidebush: I think from a Segment perspective, it does go back to that identity resolution piece. You have SDRs and AEs and marketing trying to engage with the top of funnel and prospects, you have support who's also woven into there, because we have users who are... Maybe as they are progressing through the sales funnel, signing up for an account and trying it out themselves, then they run it to my team in our scaled success world, delivering time to value, nurture content, and then hopefully they buy and then now they've flipped over into CES world and professional services, so all the different parties have different ways that they define users and define what moments are important, and that data and understanding could be siloed by different power users or in different systems, and getting all that stitched together is absolutely one of the most challenging things, because it's not just a data or a systems problem, it's also a human challenge.

Kailey Raymond: Yeah, the real human challenge of the departments and what their individual goals are and what they're working on, and how to make sure that everybody is really building the customer journey together. Yes, absolutely. What about you Seth? 

Seth Familian: Yeah, I think to Matt's point, so when you listen to what Matt just said, you're probably thinking, oh my gosh, there's so many sources of data that he's contending with, that's just like my world, I've got this and this and this and this and this, and I wanna try and bring it all together. And it's almost this deer in the headlights moment of, there is too much there, where do I even start? And then you start planning, and you wanna do all the things all at once, and so I feel like the two hardest things when it comes to customer engagement are one, being willing to pace yourself and be iterative instead of trying to boil the ocean, and two, having, Kailey, to your point, all the right people in the room at the start to align on what the minimum viable data set looks so everyone can be satisfied and then continue to align them as you move into more advanced use cases, more sophisticated profiles, etcetera. If you can get that notion of iterative improvement of data and profiles and engagement, and then come back to data and then go again to profiles and go to engagement, and then you get into this like habit. It's almost... When I originally thought of this, it was like a yoga practice, like when you do...

Kailey Raymond: You're doing your Sun Salutation...

Seth Familian: I'm doing your Sun Salutations and okay, I'm gonna get a little less of woohoo. I lived in the Bay Area for 12 years. You don't build heat with a single vinyasa. You don't. And...

Kailey Raymond: You don't. You don't build heat with a single vinyasa.

Seth Familian: You don't. You don't. And you can't...

Kailey Raymond: You've put a lot of slogans today.

Seth Familian: Yeah. A lot of slogans, I don't know.

Seth Familian: It's a lot for a T-shirt, maybe that's more for a mug, I don't know. [chuckle] Definitely a mug that contains hot beverages. Okay, but what it really means in practice is, you can't just try to create this really powerful system with one fierce overwhelming effort and you're done. It's just not gonna happen. It's the slow methodical, continuous improvement and making it work, and it's, okay, now we're cooking with gas a little bit, now we have some activation and, okay, the needles are moving and the KPIS are growing, and let's layer this event on top, let's build this audience. Let's do this engagement. And as you iterate more and more, I feel like you can align those stakeholders beautifully.

Seth Familian: So then what becomes so challenging is that so often that process does not happen, you have a single stakeholder who shows up and said, I'm gonna implement this thing because our company needs it. And their intent is so in the right direction, but they end up in implementing only around their frame of reference, and then they don't consult these other stakeholders, and then those stakeholders start using or misusing the downstream data that they've made available and now people become highly misaligned very quickly, so then you have to step back and you have to reconcile everyone and say, "Okay, let's get on the same page, let's understand what we build, let's understand how it's useful." And then you start steering everyone back into that alignment and then you can get the process going again, but that to me is the hardest thing.

Kailey Raymond: Well, this is really interesting, I think what you're doing is you're also kind of teasing out the maturity curve that you talk about in your book, which is really talking about the iterative process, and to your point, in speaking to... I'm forgetting, I think it was an analyst last year, the fear of this all or nothing, of, once I start on this journey, there's a lot of stuff that I'm gonna have to do now, and it feels so daunting to have to dive in, 'cause you feel like you're going from maybe zero... Maybe you're at one, and then you wanna get to a real time personalization, that is unbelievably vomit inducing intimidating, I am sure.

Kailey Raymond: For most companies, but to talk about it in the way that you're talking about it, which is what your book does, I think is the way that makes it feel more approachable and realistic for people to actually be be able to achieve.

Seth Familian: And yeah, just to close the loop on the framework, what the framework really describes is three phases of maturity, foundational, advancing, adaptive, and we're just increasing levels of sophistication in three pillars, which I described earlier, data profiles and engagement. So foundational maturity is just getting the basics right, it's clean basic data resolving into some simple profiles, maybe across one or two sources with a couple of identifiers, with basic campaigns, and I mean basic, single step, single channel, the only fancy thing we talk about in engagement, foundational engagement is this notion of timed out audiences.

Seth Familian: So hey, if you send an audience downstream to Facebook for spending money around or Google Ads, yes, we will automatically suppress them once they convert, but you might wanna also remove someone from the audience after say four days, because do you really wanna spend on that person after that period of time, and could you do this directly inside Facebook? Sure, but what's nice is when you upstream that action in Segment, then you can federate that logic across any other destination that you send to at the same time. So that's the basics. And then you move into advancing and now we're getting more sophisticated with our data, we're auditing it, we're consolidating it, were governing it with active tracking plans and violation detection, we're layering more interesting traits on top of profiles like inferred preferences, oh, Seth's favorite color must be blue because the most frequent product color he's viewed in the past 30 days is definitely blue, and that's different than these explicit preferences of Seth told me his favorite color was green in the profile, so how trustworthy is the rest of his profiles? [chuckle]

Kailey Raymond: Don't believe the narrator.

Seth Familian: Don't. And yet it's so funny, 'cause in this basic world, you still wanna capture the profile, so you get the basic blocking tackling, tell me that you're a developer in the flow of onboarding you into the app, but then are you really just looking at all the marketer capabilities? Mmm. Maybe you're actually like a MarTech ops developer, and you know the engagement is more multi-channel, multi-step.

Seth Familian: Ultimately, when you're in this adaptive phase, it is just scale and speed and automation on governance and doing these really powerful things that you couldn't even imagine. The sort of nausea inducing things that Kailey you had mentioned earlier, and only then do you layer predictions and the notion of AI and ML on top because you've built this really strong foundation underneath. And so that's the framework, and yet what's so funny is we talk about it and ultimately, sure, it's like an analytical way of looking at things, it comes down to people, and it comes down to people choosing to either align with the process or do their own thing, and then you have to bring it back together, and I feel like, Matt, in a way that's a lot of what you've been talking about too, like you have to get people to instrument things in the right way, and then you have to get people who are using the data to come back to you and say, I need this different, I need this tweak made, and then all of a sudden you're able to make use of this amazing infrastructure you've created.

Kailey Raymond: Let's talk about a couple of those things, Matt, which is the stakeholders and the people that we need to be champions, and you run scale programs using Segment data on our own customer base, so adoption programs, retention programs, things like that. Can you share just a couple of those and some of the data inputs that you are looking at and how those programs run. Walk us through an example or two.

Matt Smidebush: Yeah, absolutely. So one of the things I am most proud of is what we call our quantitative health score, this takes a number of different dimensions, from breadth of instrumentation, depth of instrumentation, BI potential, how many users and what kinds of users are a part of the account or using Segment but this doesn't necessarily give us a clear equation to value, which of course that's what we'd love to be able to measure quantitatively, but what it does do is it gives us the basic building blocks of use cases, is a customer doing enough building blocks? Do they have enough Legos? And if they have enough Legos, they probably could put them together into a value statement, so that is completely powered by Segment, it takes intent data from Salesforce, which we collect in the sales cycle, it takes user data that we collect in the sales cycle as well as what is captured in our user onboarding clause in the product, it captures live product data, of course, and then usage data as well.

Matt Smidebush: And that's what we use to target various different intervention, so the score is out of, arbitrarily at a 55, if the customer is not at a score of 20 by three months, then we are very concerned about their path and their time to value, so we start triggering interventions there that could range from, hey, talk to your CSM who wants to run a strategy session, that could be a piece of content, it could be some documentation, it could be a webinar, and all of that has contributed to moving these customers materially long, and we're not measuring data maturity quantitatively yet. Seth, we're talking about this next week, but it's...

Seth Familian: We sure are.

Matt Smidebush: But we're moving customers along in acquiring more Legos so that they can be building valuable use cases. That's one concept. We also take a bunch of different data and surface that to customers regularly, we call this our monthly metrics campaign, I'm sure if anyone here is a Segment customer you've probably heard from Mason Schroder who runs this campaign from Mason Schroder at the bottom, and the goal there is we are highlighting where the customer is right now, some trends about their usage that we think are important, we offer what should you do next, what kind of content should you consume, maybe you haven't built an audience in a while, why don't you come back in and start building some more audiences, here are some examples for you, again, this is a very data-heavy, complex campaign, but it allows us to do a lot of personalization based on, again, user time and journey, what tools do they use already? So those are two examples. I can talk...

Kailey Raymond: I love those.

Matt Smidebush: You got as much time as you have, Kailey, we can talk through more. [laughter]

Kailey Raymond: I know. Matt's like, let me just be so proud about my team for a moment, and I love that. I am like, I am over here, maybe a little bit seething, but also so incredibly just proud of that and being able to work with you because I was trying to do that exact same thing at my last company and couldn't do it because I didn't have Segment. And I was trying to get us to purchase it, so I was trying to do exactly that. I was trying to do nudge-based behavioral emails to make sure we could get people on-boarding per persona, and I was trying also to do product adoption score. What parts of the product are gonna be the stickiest, making sure... We were focusing on the right areas. And those two programs that you've launched are the more mature, more robust versions of what I was trying to do, so it's so cool to watch.

Seth Familian: Fun fact, I will, from time to time, create a new workspace in Segment or create a brand new account. I actually had to create a brand new account the other day for some weird reason, and I think it reminded of all the super awesome stuff Matt's team does, 'cause then I get this amazing cascade of communications at the right time around things that I've set up and I pick a different persona and then it sends me a slightly different message, and it guides me towards how to set things up and get things going and how to unblock me and I'm like, damn, this is really good. [chuckle] We need more people to do this thing in our customers. So it's super cool.

Kailey Raymond: That's awesome. Yes, the onboarding is a critical, critical, critical first 90 days, so it's amazing to watch what you do. Seth, I wanna hear from the big enterprise if you have examples, customer side of the house, interesting use cases that you've helped solve.

Seth Familian: So there's one that we worked on recently that I thought was really fun, a lot of the enterprise stuff is just a high volume or super high complexity. So a high complexity one we did a while ago was... And this actually, this case study's in the book. The book has a bunch of different... And I wish I could say I wrote the whole book... I did write a lot of the book myself, but it would not have been possible without our marketing team and our design team, which is incredible, and they actually found a ton of these use cases, they also made sure my grammar and language was actually good, but [chuckle] beyond that, they found these amazing examples and use cases from our own use case libraries in our customer testimonials, and this was the end of December, it was a new Segment customer, I won't tell you their name, but I will say they're a massive website in the travel space, we ingested half a billion anonymous profiles in 30 days, and then cookie-matched all of them with Google Ads and then enable them to build audiences where they could in near real time sync down with DV360 and the Google Ad Manager. So they could re-target those anonymous users on their own site for selling inventory.

Seth Familian: But that wasn't even the magical part. What we also did is we were able to ingest all of the data, the behavioral data from another one of their website in the past 30 days using reverse ETL because they didn't have a streaming pipe yet, and all of the member data associated with logged in users for both of those websites, and by doing complex identity resolution, we were able to uncover a 2 million person Segment that overlaps between the members of one website who were anonymous visitors on the other site and members of the other website that were anonymous visitors on the first site, and it blew their minds because now they have the ability for the first time to leverage this known first party data on one side of the house with a non-otherwise completely anonymous data, it was linked back with a member ID and a shared email, and that's how I was able to stitch together, and now they had a whole new way to market and ultimately monetize their traffic in a way that was in accordance with their privacy policies, they weren't doing anything dodgy, they were just making the most of the consent that people had given them to say, "Sure, market to me." And sure enough they were doing it, and now they have a whole new way to monetize and a whole new way to target them, so that's the most recent one that I think was really, really fun, and it was a insane scale.

Kailey Raymond: That's awesome, yeah, I think identity resolution is really one that it will just blow your mind, and the first time that you think about it, if you think about it too hard for too long it might just continue to blow your mind. Yeah, it is an unbelievable problem solver, especially at enterprise scale, it's just 2 million in a couple of seconds to be able to resolve those. Unbelievable. Crazy.

Seth Familian: Yeah, and also when it gets really interesting, I love the fact that our platform doesn't make assumptions, so I had an experience this past week where I was helping this customer build their own identity resolution on top of multiple email addresses and using the Segment ID, and we ended up sending down to an analytics tool, and that analytics tool made assumptions around the identity graph, and it decided to do us a favor and start merging one set of profiles against another set of profiles with no way to easily undo it or set the configuration. And thankfully, I worked directly with the analytics company and we were able to figure out a work-around, but it reminded me that it is very dangerous to make assumptions around how identity should be stitched together, and yet it takes a whole lot of maturity for a platform to put together a system where you can set your identifier order and cardinality and do it on a per engaged base basis, so you can have different graphs for different ways of looking at customers, and the ultimate result is things merge the way you're expecting them to merch, and complex mergers happen in a way that is predictable and it's deterministic, and it just gave me new appreciation for what we had, and that flexibility and that flexibility is not necessarily a given, so it's fun, it's mind melting, problem-solving on the one hand, from the other hand, it's really fun and incredible flexibility.

Kailey Raymond: I do feel like we're melting a couple of brains there, and I think that's probably a good thing.

Kailey Raymond: We're really in the weeds and we're stoked about, and I think it's cool.

Kailey Raymond: So I'm gonna shift it a little bit and we're gonna talk about some other folks, some people that you may find inspiration from. So are there companies or folks that you look to in the space that you think are doing it really well? Customer engagement, customer experience, those moments when you're like, wow, they got that right.

Seth Familian: I think there is a very large thing that is doing it right in a way that is mind blowing, and they are ingesting in real time all of the page for you data on their website, they're making websites and they are identity resolving all that data around people who are logged in and they are flagging people who look at fraud-related pages, like I need to file a complaint, I need to file a dispute, a charge, and they set that as a computer tree on the user's profile, and then when that person calls in to the company's call center, they look up the profile against that phone number and they see that fraud visit is true or interested in dispute is true, and they re-order the options in the menu provided when you call to make if this is about fraud, press one, and to me, that is doing it right? 

Seth Familian: Nice.

Seth Familian: Because that is, it's understanding, you deal with the super important pain points, first the things where people are scared, the things where people have urgency and you recognize they don't care about opening a new account at that point, they better know that you're a customer and that you are calling in and you are concerned, they're meeting you where you are as a result of reducing the amount of call time because they forward you directly to the right agent, they're being very insightful and thoughtful with how they're using data, so to me, they're doing it right.

Kailey Raymond: That's awesome. I love examples that are on the border of customer experience in the call center form, because that's when probably all of us, I don't know about y'all, but maybe you've gotten a little bit heated, maybe you've gotten just a little bit frustrated before. And you're talking to a robot and it doesn't feel like anybody understand you...

Kailey Raymond: But like that's the moment when you need a human, and that's when thing you need them to know all of the things that you've done on all of their apps in their website, and they should know that about you. So to know that a bank is doing this well, it gives me a lot of hope. That's awesome.

Seth Familian: And that also is what we really call omni-channel, I know that is a buzz word that's thrown around a lot and the more...

Kailey Raymond: That's true.

Seth Familian: I got into writing this customer data inventory thing, the more I realized, What is omni-channel? Oh, it's bringing to your next interaction every single bit of context that you have on that customer to make that next interaction relevant and meaningful and effortless. And that's exactly what you've just described.

Kailey Raymond: It's realizing that technology is an extension humans and channel of choice could be in person, it could be a text message, it could be a phone call, and you need to know that.

Seth Familian: Yeah. Yeah, I love that.

Kailey Raymond: Matt, you got anybody that you think is doing it right? 

Matt Smidebush: Seth, as you were speaking about your example I was like, are we talking about the same company that... And actually not, because this is an insurance company, but it's similarly doing...

Seth Familian: Really.

Matt Smidebush: The same kind of call center optimization and storing everything on the profile, leveraging the profile API to pull up all of the context of that user and reducing call times within their agents. This is more, I don't know if they're doing the automated system optimization, but in terms of their actual agent to do customer interactions, that was personalized via profile API. I think maybe we go to the same place of as a consumer where I experienced the most frustration is in call centers... [laughter] so anyway, I think they're absolutely a company that is doing it right.

Seth Familian: I love it.

Kailey Raymond: My last question for both of you is, if you were to give somebody advice right now about the steps or recommendations that you would have for them, if they were going to start to think about up-leveling their customer engagement tactics and strategies, what would that be? 

Seth Familian: I've got two that come to mind really fast. Capture consent immediately. Do not pass go unless you have captured consent, because if you don't get it at ingest, you can't use it elsewhere, and you are just going to be fighting an uphill battle. And infer. Be willing to use computations to infer meaningful things and then test hypotheses on top of those inferences, whether it's inferring someone's favorite UTM source over the past 30 days, or their favorite color or whatever it might be, it's a surprisingly easy thing to do. I'll give you one really quick example, which is there's a customer that's a digital marketplace for goods, including downloading music, and they've been in instrumental on Segment for many years, and one of their best personalization use cases and case studies was they had generic emails that they sent out to people saying, download some more music, most of them were generic. They changed it to... They measured the first category that someone ever downloaded and the last category that someone ever downloaded, they emailed out from blank to blank, first to last category, we have everything that you need, download it now, and then they carry that out throughout every single call to action. Their click rates went up by 30%? This was not like boiling the ocean and doing it, it was just a simple set of computations that were relevant to that human nature of communication, so infer and capture consent are my two biggies.

Kailey Raymond: I love it. Matt, what about you? 

Matt Smidebush: My biggest one, we've already talked about, it's iteration, it's really... You don't have to start big, in fact, I guarantee everyone listening to this has an idea that they could start with. Just do that. Do something very simple. Do... Optimize one email campaign, optimize the one advertising campaign, think about one additional interaction that you could change and start there, it'll be a lot easier to generate buy-in for what you're trying to do, once you have a couple of iterations to show, "Hey, we actually get a lot of value out of doing things this way, why don't we apply this methodology to X other thing?"

Kailey Raymond: I think that's probably the theme of the show today is iteration, starting small, being able to test and gaining buy-in from multiple different stakeholders, building champions, but it all has to do with just starting. So Seth and Matt, I really appreciate your expertise today, thanks for being here. It was fun.

Seth Familian: It was so fun. Thank you.

Matt Smidebush: Thanks, Kailey.