Apr 18, 2019
Dan Head (CRO) and Robbie Matthews (Director of Integrations and Onboarding) join me in foggy Londontown to chat about the evolution of marketing, from old world email and DMPs to flexible architecture that is channel agnostic. Learn the nitty gritty of why new school tech is leaving legacy clouds in the dust. Dan also hints at a future of talking toothbrushes and Robbie shares his data dreams.
PJ Bruno: Hello again. Welcome back to Braze For Impact, your weekly tech industry discussed digest. This week, we're here in London, in our London office and I'm thrilled to have with me two close friends, Dan Head, Chief Revenue Officer at Braze. Hey Dan.
Dan Head: Hey PJ. It's good to see you.
PJ Bruno: And also, Robbie Matthews, Director of Integrations and Onboarding. Robbie, how you doing?
Robbie Matthews: Good, how are you?
PJ Bruno: I'm quite well. And I know it's a bit of a canned response, but I got to say, the weather's unbelievable here. I've been getting all this bad intel about bad weather and this is unreal.
Robbie Matthews: You got a British tan?
PJ Bruno: Yeah, exactly. I don't know how. I literally got sunburnt yesterday. It's ridiculous.
Robbie Matthews: It's all lies. Fake news.
PJ Bruno: Well, you know, coming into Easter weekend, how you guys feeling? Happy Easter, by the way. I don't know if you guys celebrate. Plans for the weekend? Any...
Robbie Matthews: We tend to celebrate confectionary throughout the year, anyway. So, we can eat chocolate eggs this weekend, too.
PJ Bruno: I notice in the lobby, there. You guys are loaded up. What we're going to talk about today, really, it's kind of like the evolution of marketing technology and more than that, we want to demystify a lot of the things. Because you hear a lot of snap terminology, a lot of jargon, but let's actually break it down and understand what we're talking about, using the facts. And, before we jump in, I love... you know, Robbie, I know you eat, sleep, and breathe data, or so says your girlfriend.
Robbie Matthews: I thought you may bring this up.
PJ Bruno: Yeah, can you just give us that anecdote really quick?
Robbie Matthews: So, Dan's already smiling at me.
Dan Head: I'm not familiar with this one.
Robbie Matthews: So, I was sleeping the other day and I woke up and my girlfriend was just staring at me, like a puzzled face. I was like, "Why are you looking at me whilst I'm sleeping? Firstly, that's a bit weird." And she was like, "Do you realize what you've been talking about for last two minutes?" And I was like, "What you talking about?" She said, "Oh, you've just been sleep talking about, like, piping data into various places and where you should send data to, and this is how you send data." And I was like, "I need to get some better dreams."
Dan Head: Need to cash in that brace therapy reimbursement.
Robbie Matthews: Or she needs to find a new boyfriend, is what she needs to do.
PJ Bruno: Well, that's her journey and that's for her to decide. But, anyway, I'm glad you were able to share that with us, thanks bud.
Robbie Matthews: Thanks, PJ.
PJ Bruno: Yeah. So, let's jump right into it. Best place to start, where we came from as far as marketing tech. You know, Dan, you're old. Why don't you tell us...
Robbie Matthews: Everybody says that.
PJ Bruno: But you still look so good. We were saying, "Tell me your moisturizer, because I need to get some of that."
Robbie Matthews: What's a face care regime?
Dan Head: I'm embalmed.
PJ Bruno: Yeah. I just let snails crawl over my face every night. It's a beautiful thing. So, how do we move from email to the concept of a stack, or an ecosystem. Do you think you can talk about this a little bit, Dan?
Dan Head: Yeah. I think it goes back to the perspective that businesses have had on just these technologies from the ground up. Historically, I think companies have looked at providing communications on channels like email as being a utility. You know, you set up a business, I need seats and chairs and desks. I need water coming out the taps and electricity. And I need a way of sending email. And, when the marketers have wanted to have these tools to communicate to their customers, they've asked the IT teams to provide this technology, and the IT teams have gone out and procured that technology. So, over time, a stack of tech debt has built up, if you like. By acquiring these utilities, acquiring technologies for these channels and finding ways to stitch all of that together. Which is, if you like, a ground up, let's just go get the bricks and build the house, sort of thing. But, unfortunately, the world's different now, because every year, CES throws millions of new channels at us, you know? It's connected wind screens, it's connected forks and toothbrushes and all sorts of other crazy stuff. And, businesses just can't keep acquiring channel specific tech anymore. It needs to be data centric, it needs to be customer centric. Because, all of us as individuals have got our preferences about how we want to communicate with brands. And so, now we're moving away from a utility plumbing based approach, to a customer centric data oriented approach. And, that is how, I guess, the modern concept of a vertically integrated stack has been born.
PJ Bruno: Yeah. And, obviously there's these advances in technology, we're going way more mobile, right? But, there has to be those that are lifting up the torch and kind of pushing us forward. And putting that pressure on legacy, right? You called them the role models, yesterday.
Dan Head: That's right. Because there are businesses that have been born into this new environment. When you order your on demand taxi, or you order your on demand pizza take-away, whatever, when you're looking for a date, these are technologies and services that haven't existed previously. They've been born into the device. And, you look inside those companies, and the teams and the people and their KPI's and the technology, it's completely different from a legacy enterprise business. And, if those companies don't get customer orientation and data driven communication correct, it's an existential problem for them. They haven't got another business model to fall back on. And so they become, and this is why I see them as role models, they become role models for more classic enterprises. And, that would be my advice to any long standing enterprise, is to go look at those companies and see how they do it. It's a good place to start.
PJ Bruno: Do you think they're actually, these bigger legacy companies, do you think they're actually interested in completely reworking their tech? Or does it feel like stitch up solutions to stay relevant?
Dan Head: I think in big companies, big enterprises, I think there are a bunch of political reasons why it makes sense and it's safe to stay with the status quo, you know? If the CEO or CIO says we're going to do this one particular way and we've got our IT standards. And, I work in the CRM team and I know the tech doesn't work, well, I'm only going to have so much appetite to stand up against those folks because at the end of the day, I want to get paid.
PJ Bruno: Right, right.
Dan Head: Right? But, I think at some point, and I appreciate the political importance of standards and doing what the bosses say. But at some point, the political importance of customer experience and the political importance of driving outcomes with those customers, I think, is more important than the politics of IT standards, and so on. And businesses eventually figure that out. It's just part of their evolution. And one of the factors that we see, which I think is a good indicator is, where businesses have taken the CMO role and the CTO role and then they've sort of adapted it. And now you see this emergence of the CCO, or the chief customer officer role, where the customer officer has got a greater stake and influence on tech decisions. Because it is literally customer oriented. That tends to be a good indication of that evolution is fairly well progressed.
PJ Bruno: Yeah. You mentioned customer experience. It's funny that now it all kind of just comes back to that. It's like there's this strength in tech, but really it's ideating towards the customer. And so, my question, I guess, to you is, what is customer experience actually mean now?
Dan Head: Yeah, just use some examples. We all use these role models, these new technologies, and so on. But, I'm reminded of a quote from Keith Weed, the current, or perhaps former, because he's leaving, CMO of Unilever, he had this expression, "You know, our job as a brand is to get to the future first, and welcome our customers as they arrive." And most enterprises-
PJ Bruno: That's a great quote.
Dan Head: Yeah.
PJ Bruno: That's a strong one.
Dan Head: I dig that one up a lot. Yeah, I like that one. Most enterprises, I think, are just trying to catch up with the current world. Let alone the future world, so just as an example: If I'm driving down the highway and maybe I would fancy a coffee. And let's say, the folks at Costa Coffee, this business acquired by Coca Cola, they want to send me a notification to say, "You're coming up on a service area and there's a Costa and you can get a deal on a cup of coffee there." Well, I'm driving, there's no point in them sending that to me on my phone. But, as we saw at CES, there are technologies like connected windscreens and there's more connected computing in cars now, anyway. So, there are opportunities for Costa to send that message to me on these new devices. Which would be a more appropriate way to deliver it. But, brands just aren't ready for that. Because brands are channel centric, they're not data centric, they're not customer centric. They're not able to make the most of these technologies as and when they emerge. And, as a result, therefore, they can't be customer oriented.
PJ Bruno: Right. And since it's all about the customer experience, right, you want to make sure that you're not abusing that use of data, right? So, we live in now, this world, it's a wash of data, right? It's all over the place. It can be collected. So, why is effective and responsible use of data important to what you do?
Dan Head: Yeah, I mean, of course we've had laws like GDPR and then there's going to be the California Data privacy laws coming in. And the world is realizing that this is... Brand connection with customers is about trust. And that comes through the responsible use of data. So, it's not that there are just new legislative frameworks, which make it law that we have to be responsible. But it's just the basis of good marketing, anyway. And so, the technologies that need to be adopted here, need to enable the marketer to use data effectively, at the right place, at the right time, to seek permission. And to act on that permission and behavior that is being granted and communicated by the customer. It's like a conversation. There have been studies, like the one Forrester released on the brand humanity index about the effect of, and the positive effect of talking to your customers in a human way. And what talking in a human way actually consists of. So, it's not just law and legislation that makes us a necessity, these are the ingredients of delivering positive outcomes.
PJ Bruno: Right. A human touch is super important. And, now also, more than ever, the stack and the technology is really important. And I remember, you said yesterday, the term you used is, "Architecturally enabling", right? And so, Robbie, you're a little more on the technical side. So, from your perspective, from these role models perspective, even, what are the fundamental engineering building blocks that are architecturally enabling versus those that are somewhat prohimitive.
Robbie Matthews: Yeah. If we start at looking at these traditional enterprise businesses, and typically they're built on relational databases as a foundation. And then we take an example of what Dan's mentioned of these role models, which are using document based storage as a foundation. We can look at what those are and then how they actually enable what you need to send to customers. So, firstly, just what is a relational database? It's essentially a storage of data that's organized into tables. And each of those tables has a very rigid, defined schema. And what you need to do with that, is be able index it, to be able to say, "Okay, I've got this data in one table, how do I join it to another data in another table." And Dan mentioned tech debt, earlier. And that's a big issue. So, if we take a client like Deliveroo, you can say, "Okay, what sort of data have they got?" They've got their user data, they've got all their purchase data. Maybe have reviews that they've left about you. Favorite restaurants. And that all exists in different tables. And as the landscape becomes more complex, you're going to only increase the amount of data you want about users and the environment they're operating in. And that's going to create tech debt. So, what does that actually mean for a brand, when you come to wanting to send messages and target and talk to your users in a human way? And, really, the first thing is agility. If you've got an ever increasing amount of data that you need to join together when you want to create a segment to target those users. That becomes a much, much harder piece to actually stitch those data tables together, access the data in a quick way and define the messages and the audience that you want to send to your users. So, what do brands actually want to do? They want to use that data. They want to define customer audiences. They want to send individualized messages based on any kind of behavior around a customer, like demographics, location data, contextual data. And as soon as you need to index all those tables, or predict what a customer is going to want to do with your platform, which is the tough thing. We can't predict what customers want to do, that changes every day, as new technologies get introduced. As breaking news happens, you're going to want a segment or an audience that we just can't predict. And as soon as you need to rely on the indexing and the schema to access that data, we lose the ability and the agility to do that, in such a rigid manner, which is what the databases in these relational databases require. So, how, I guess a MongoDB or document based storage actually solves that, is, firstly, it's a schema less database. So, we can define exactly how we want to structure data to support these customer requirements. Using documents, as opposed to a relational schema let's us model very complex objects. And then, due to the structuring of that data, we don't need to make those joins and stitch all the data together in this increasingly complex landscape. So it means customers can make a heap of customized segments, off the cuff, and really give them the agility to create these audiences, send out campaigns as they wish.
PJ Bruno: And that's all around like free flow of data, right? Because, relational is a lot more stringent. It needs to be set up in a way... I mean, that's kind of what I'm hearing, is...
Robbie Matthews: I guess there's two parts to that question. There's one is speed in the context of a database. And then, I think when you're talking about free flow of data, that's probably where we want to look at more around ETL processes and Kafka. So, maybe let's touch on the speed in relation to relational and document based storage and then we can come onto Kafka and ETL processes.
PJ Bruno: Okay.
Robbie Matthews: So, in speed... As you say, customers just demand customer engagement platforms to run fast. Customers demand it. If you need to send out a break in push, ABC wants you do that as fast a possible. And, really, that's about read and write throughput on your database. So, let's take an example where you have to, like in a relational database, where you need to make a number of individual queries, join different tables, and that could take days to produce the results. Whereas, a no seek with MongoDB style database allows you to run queries in parallel, break up your dataset, run higher number of queries, and get that data back almost instantly. So you get those audiences that you want to target much, much quicker. Yeah, so that's the speed element on the relational database versus MongoDB. The second part of your question about ETL versus Kafka and the streaming is, again, like, "How do I actually access that data, or upload my, data into my data warehouse?" Think previously, Dan mentioned utility. Like they way that brands store data has just changed. Various data from different DB's could be loaded into a master data warehouse, once or twice a day. And that's fine. However, let's take the example of Deliveroo. I don't care to admit how many Deliveroo's I order a day. But, they need to know if I've made two this morning, or one this afternoon. And they don't want that upload of data to happen 24 hours later. And to get told that I actually made three orders in the last day. They want to know about that as soon as it happens, be sent to them as that event takes place.
PJ Bruno: Right, but you don't want to know that you've ordered three times in one day?
Robbie Matthews: My waistline doesn't want to know that I've eaten three times in a day, but that's a different story.
PJ Bruno: It's for another podcast.
Robbie Matthews: So, ETL versus parallel stream processes. And, what that parallel stream processing does, is essentially sets up a queue of events as and when they happen. Then you can listen to those events, publish those events and then essentially subscribe and act upon those events as and when they happen. So, let's take an example of a train booking company, a UK train ticketing company. Where, if I make a purchase for a journey later today, in the ETL world, that upload may happen tomorrow. In which case, I've already missed my train. I've missed all of the communications around it. It doesn't help. In the Kafka world, as soon as that event purchase takes place, you, as a business can act on that purchase immediately, send information about when the train is, any delays about that train. And make sure that user has a good experience with your brand. But, like I said, in the ETL world, I've probably already missed my train because it left the station 30 minutes late. And then the data was only sent to you as a business, to act on the next day.
PJ Bruno: Yeah.
Dan Head: This term, real time, I think is sort of used and abused and misunderstood. Like the examples that Robbie just used about travel, timing, of course it has to be real time. It has to be in the moment, otherwise, you're going to miss your train. You're going to miss your taxi, whatever. And, those examples apply to all other businesses, as well, right? Because again, this interaction between a brand and a customer, in order for that to be human, it has to be current. It has to be up to date, based on information that is relevant right now. So, as an example, let's say, I'm on the train and I'm on the way to work and I'm browsing for some sneakers on my mobile web, on my phone. And I add some things to my cart, but I don't transact and then I get to work. And I nip out for lunch and I actually go an buy those sneakers in store. But, maybe there was an abandoned cart email scheduled for 2:00 PM. And that abandoned cart email is still going to come out at 2:00 PM, it's not going to be cognizant of the fact that I went into the store and I bought those sneakers at lunchtime. So, that is a bad customer experience.
PJ Bruno: Right.
Dan Head: Or, just to continue that example, if I'm one of those companies that does sneaker drops, and there's a limited amount of inventory at a particular location. The communications can and should be based on that inventory dropping, you know? Create that exclusivity and that need and that excitement, which is what the sneaker drop is all about. And, I guess what Robbie's talking about is, unless you've got a database structure that can make and broadcast, or make available data in real time elsewhere in your technology ecosystem. Unless that data's being distributed in real time, you can't enable those use cases and you can't satisfy that need for a good, premium customer experience.
PJ Bruno: And the value of that is just less and less as time goes on, right?
Dan Head: Yeah, exactly. So, I don't really want to know tomorrow that the sneaker drop inventory has run out, you know?
PJ Bruno: Just creating the opposite of good experiences?
Dan Head: Right.
PJ Bruno: It's just frustration. I'd rather just get nothing, right?
Dan Head: Right.
PJ Bruno: I'd rather not even see it. All right, cool. Why don't we do something fun and round up with, I mean, you were talking a little bit, Dan, about toothbrushes that talk to you. I want my toothbrush to talk to me so badly. I have to just talk to myself at the time. But, what's the future hold? What kind of last words do we want to put out about, maybe predictions we might have, or a word to the wise about a good strategy, or just some advice to keep in mind moving forward.
Dan Head: I think there's two angles, to look at it from a customer's perspective. Speaking about myself, I mean, there are always new devices and new technologies and some of us are going to be quicker on the adoption than others. But, as a population now, we are more tech savvy than ever before. More willing to try technologies more than ever before. We have an expectation of being delighted more than ever before. We're willing to kind of experiment and try stuff. And, from a brand's perspective, how on earth are you going to deal with that? So, brands just need to future proof themselves, they just need to be data centric. They just need to have the engineering horse power. They just need to have the developed muscle groups, the ways of working, the team structures based around what's possible with real time data. And then, whether it's that toothbrush talking to you, or the fork telling me that I'm eating too quickly, you know? Whatever it is, the brand can deal with that. And, I think ultimately, where brands are trying to get to, is there should be this ability to be creative and playful with technology. We shouldn't just be worried about the plumbing, "Oh no, I've gotta send a message on this channel now, rather than that channel because of this new device's come out." If we can use any channels, anywhere at anytime, based on current data, then we can really use that as a creative canvas for just amazing, delightful, surprising marketing. And, that's the kind of thing that we, as a population, kind of enjoy. We don't enjoy stuff that's irrelevant, we don't enjoy stuff that we haven't opted in to. But we are willing to be surprised and we are willing to be engaged if we think the brand is making an effort to do so.
PJ Bruno: I love it. Robbie, you got anything?
Robbie Matthews: I think that, I'll bring it back to the architecture, to Dan's point around what brands want to do. You want you customer engagement platform at the point you're ready to use sort of surprise and delight strategies that Dan's talking about in reaching customers. You want your engagement platform to be ready to allow you to do that. And not still be developing itself. And, the way its architected, and the way it's able to meet the marketplace's changes is going to be crucial. So, I mean, WWDC happens, Google IO happens, there's a heap of changes to how you can send messages to users. And your engagement platform has to keep up to date with that. So, how they're architected, how they're able to be flexible to meet the new abilities of these OS's, is crucial. And when you look at, again, going back to the sort of, if you want to make joins, create new tables, that's tough to do. And then you have to look back at your current architecture and work out how you stitch that onto the end. Whereas, we take this example of something like a push carousel, where you send a push and you can swipe through a number of images to that. That came out last year. And just the way we were architecting meant you're able to immediately respond to what is new in the market, and give the brands the ability to go out and use that, as and when they're ready.
Dan Head: Yeah, and we've talked about food quite a bit here. Maybe it's because it's coming up to lunchtime and it's all on our minds, right? But, an example that just brings all this together for me, I think some of us have seen it, which was recently Burger King ran a campaign where they hijacked McDonald's branches and if you got within 600 yards of McDonald's, you would get a message on your Burger King app saying, "You can now redeem a Whopper for one cent." And it would direct you to the Burger King. And it was this cheeky, creative, just technology embracing campaign that got so much attention, they got to the top of the App Store. I mean, how crazy is that? That a burger app can be higher than Instagram and YouTube on the App Store. It's just-
PJ Bruno: Well, burgers are future proof, Dan.
Dan Head: They are. And, particular popular where you're from, huh, PJ? But what a great example of creativity with technology. That campaign wouldn't have been possible unless you had the technology in place that Robbie was talking about. And then, once you've created that surprise and delight, now you've got so many more people that are engaging with your brand, that have downloaded the app. And now you can really start to go and have fun with them.
PJ Bruno: Totally. Well, Robbie, Dan, thank you guys so much for being here. And thank you all for joining us. Just remember Shift is happening. And Legacy Clouds, in the words of Dan and Robbie, don't miss the train.