SPEAKERS

Toby, Steven, Intro, Spencer

Intro  00:00

I’m Steven and I’m Spencer. Welcome to Ragnarokast, your podcast for all things marketing and Martech. Hello, everyone. We’re the CO CEOs of Ragnarok.

Spencer  00:10

Beautiful intro. Every time gets me every time.

Steven  00:13

Oh my goodness. Tear to my eye. Tear to my eye. 

Spencer  00:17

Hi everyone. I’m Spencer. 

Steven  00:21

I’m Steven. Oh, sorry. I thought doing the intro again.

Spencer  00:25

Well, the intro, it’s, it’s, yeah, this time I get to say my name first. And then today we have Toby of well, depending on your region, it’s Jacquard or Jacquard. Welcome to

Steven  00:39

For me. It will always be Jacquard.

Spencer 00:45

Welcome, Toby, 

Toby

Hey guys, how’s it going? 

Spencer  00:51

Good. This is, this is what it’s like for our employees, by the way. 

Toby  00:55

So we’re just so I understand, do you guys as your intro play the YouTube video intro of a previous podcast, and is that the intro?

Steven  01:05

Should start doing that.

Toby  01:06

Is the editing budget that low. That’s how that work?

Spencer  01:12

Yes. Oh, we didn’t know how to do stop the screen share, but this is for your benefit and to get us, like, teed up. 

Toby

Gotcha okay. 

Spencer

It’s like, how when they record a music video, and sometimes you’ll notice that, like, so if you have kids, there’s Miss Rachel, and we noticed the other day that she actually has headphones in during one of the songs, that she was like, AirPods, and she’s listening to the song at the same time. Oh,

Toby  01:43

Because they’re not playing it like, yeah, okay, fine, okay, 

Spencer

Because you have to record her voice.

Toby

Yeah. It’s more like a realism thing for you guys. It’s, yeah,

Spencer  01:52

It’s getting us pumped up, you know,

Steven  01:55

Nothing that gets me pumped up, like, that five-second Ragnarokast intro, yeah,

Spencer  01:59

Yeah. Selin did a good job of like, you know, harmonizing us as much as possible. Anyway. Toby, welcome. Do you mind giving us your life story, yout Social Security? I mean, your your background and how you ended up at Jacquard, and then maybe a little bit about Jacquard?

Toby  02:18

Yeah, absolutely. So my name is Toby Coulthard. I’m the Chief Products Officer of Jacquard. You can pronounce it however you’d like. We’re not going to enforce a particular accent on it. I live in New York, despite my accent, I’m based in Brooklyn, working on my own Brooklyn accent, but it’s not quite there yet.

Steven  02:36

You’ll get it there one day. Don’t you worry

Toby  02:40

And then before, before Jacuard, for many years, about eight years at Braze, I was responsible there for things like solution, consulting, competitive strategy. And then before that, I was at marketing cloud, Salesforce, as well as a few other jobs in things like Formula One and and Electronic Engineering and IBM and a few other places. So yeah, good to be here now. Jacquard are responsible for product and we do high performance content generation for marketers, customer life cycle campaigns that enable marketers to increase productivity, efficiency, but most importantly, connect with their customers better and create highly engaging and personalized content, producing one of their users and customers.

Spencer  03:24

Excellent. Well, thanks for coming and actually question so you were at, so you were at Braze before, which we knew, but yeah, so you were at Salesforce before that.

Toby  03:33

Yeah, Salesforce Marketing Cloud back in 2015, or so,

Steven  03:37

Wow. So you really saw the light back then, huh? Yeah,

Toby  03:40

Yeah, I was based in the UK back then. And yeah, it was, it was, it was, surprisingly not, compared to how big that it was I would join right before exact target got acquired. So it was, it was right on that kind of threshold. I then left to go to a master’s degree, and then came back when it was Marketing Cloud, yeah, that was right at the threshold. And it was a very different world back then. I think that it is now.

Steven

Yes, it was.

Spencer  04:08

I’ve toyed with the idea of, like, going back to school at some point. What was, like, the, if you don’t mind sharing, like, what was kind of the, the onus of that.

Toby  04:16

I was still, I mean, to be honest, I worked, I finished my bachelor’s, and I, like, worked for, like, the summer. And then I was like, you know, I want to continue. I was doing electronic engineering. My master’s was in nanotechnology and bioelectronics, and I just wanted to kind of go into that world. I worked for McLaren Formula One on the engineering side for a little bit, and I realized that it just wasn’t my vibe. And I so I went back to, I wanted to talk to people a little bit more than than being a full electronic engineer. And so, yeah, I went back to what then was Salesforce, and kind of went from that. And then I was, was only there for probably just over a year before Dan Head, who went over to App Boy, who was the CRO of App Boy asked me to join App Boy, and then subsequently he went from what Braze to Jacquard and bought me with him. So kind of went through the different businesses like that.

Spencer  05:12

So were you? Have you been with the company since previous to the the rebrand Braze?

Toby  05:20

Yeah, yeah. So I joined. I joined probably right at the beginning of 2016 we were app boy for a couple of years, and then, and then, I think we realized that we weren’t just for apps, and we weren’t all boys, and we decided to rebrand to Braze, which I think was a smart move. Yeah, that was, that must have been 2018 and that was Marissa Aydlett, who was the CMO of Braze at the time, is actually the CMO of Jacquard now. And that, and she was quite instrumental in the rebrand from what was Phrasee before and then Jacquard. So a lot of the DNA that went from App Boy to Braze is is part of the DNA from from Phrasee to Jacquard. 

Steven

Love that.

Spencer  05:57

That’s awesome. So you’ve got all that Braze growth energy, just

Toby  06:03

DNA. Yeah, exactly, yeah.

Steven  06:05

Surprise, she did swing back to Metallurgy after, after going through the rebrand,

Toby  06:11

Everyone’s always asking about the Jacquard name and why. And we can obviously get into it, but there’s a phenomenally few dot coms left. I will put it at that which are, which are actual words. And, I don’t think we wanted to commit to some concatenation of two unrelated words. And I didn’t, well, I didn’t, I don’t think the rest of them wanted to make up a word that didn’t really sound like anything. And so

Steven  06:35

You mean, like Ragnarok marketing.com, yeah, we 

Toby  06:39

Ragnarok is an existing word. It’s not like you came up with that. So I feel like you guys did a similar, similar thing, but

Spencer  06:46

We did a similar thing in that even back in 2012 Steven and I were sitting in a room for three days straight trying to come up with a normal-sounding name. We even tried coming up with some, like, made-up names, and it’s just like, none of it sounded good. Everything was taken we’re also, you know, a New York corporation. So we needed a business name that was not already taken. And there are just so many companies in New York, so, yeah. So we landed on Ragnarok. 

Toby

I like it won its best. 

Spencer

Yeah, thanks. So we completely empathize. We went through the same exercise.

Toby  07:16

I mean, the real meaning behind Jacquard, Jacquard was a French inventor from the 18th century who came up with what’s called the Jacquard loom, which is basically the precursor to the computer. To some extent, it was the first thing to take a punch card, to basically take a program and turn it into something. And we can go down the marketing spin as much as you want, but it, you know, it takes lots of input, it weaves a very complex output. And there’s plenty of parallels, I think we can, we can derive from, you know, something that’s both relevant to technology and that we take many data inputs and we make a very sophisticated output. But at the end of the day, there’s a lot of alchemy that goes into branding and rebranding and all that kind of stuff. And we just like the name at the end of the day.

Steven  08:06

Oh, yeah. That’s like, when Uber did their rebranding many years ago, and they turned their u to the side and made a C, and then they were like, it’s the bits in the atoms. Man, yeah, yeah. Spit on that one. 

Toby  08:19

Yeah. Yeah, exactly.

Spencer  08:20

I started this up a few episodes ago where I like to ask our guest to share a hot take, not a marketing example, but my, my hot take is that people that drink tea, you know, just aren’t badass enough to drink coffee, you know. And they just, they can only take. They can only take a quarter of the caffeine and so they, you know, they just can’t leave. It’s a half measure, or it’s a quarter measure, really,

Toby  08:47

I think it depend the motivation of why you’re drinking it in the first place, if it’s to get as messed up as possible, then, then people who drink coffee are not good enough to take caffeine pills, you know, like, so, like, it depends on the reason why you’re, you’re, you’re drinking it in the first place, if you’re doing it to get the caffeine, then, then, then, actually, I’d say that. Why stop at coffee? 

Spencer  09:10

It’s a nice blend of caffeine and flavor, yeah, whereas you know tea, I actually really like tea, by the way, so I’m just trolling. So what’s a marketing hot take that you have? 

Toby  09:23

My marketing hot take. Just before this call, I was talking to a C-suite of an engagement platform, not US-based, and I my hot take was I shared with them is I don’t understand the point of a CDP. I think CDPs are a temporary fix to something that ultimately needs to be solved or should be solved, and should have always been solved by either an engagement platform or anything where the data needs to be. And at the end of the day, the CDP is was a hot thing for a while, but I can’t imagine they’ll be around for a long time. That’s my hot take. Maybe it’s not a hot take. Maybe everyone reason

Spencer  09:58

I think it’s a hot take. 

Toby  10:01

If you’re an engagement platform, why would you want to, at the end of the day, get the data from the source and you want to deliver your content to where it needs to be? If it solves the problem of not having a real-time data stack, if it solves the problem of getting the data where it needs to be, then that should be done. It’s like a band-aid over the problem. And at the end of the day, engagement platforms should have SDKs that collect the data on the web and the app. It should be able to serve content to the web. Should be able to serve content to the app. It should be there shouldn’t be any middleman that makes it more difficult. And it should have real-time integrations to where the data needs to be directly and not have this thing that’s trying to, you know, people call it the piping within the house that connects all the things, but really they should be directly connected. And especially, think with things like zero-copy and Snowflake data sharing, I see less of a reason why CDPs need to exist personally. 

Spencer  10:54

What about a reverse ETL versus like a traditional CDP?

Toby  10:58

I think it’s different way of solving the same problem, and I’m happy to be wrong, but I just think, in my view, it should be a Braze or an Iterable or a Salesforce marketing cloud or an Adobe journey optimizer or a MoEngage, or an Insider or whoever it might be, to just get it to where it’s from and connect directly to your analytics platform. It should connect directly to what if you’re doing some kind of third-party segmentation thing, it should connect directly to Jacquard. It should and why do you need something getting in the way? Connect directly to data warehouse? 

Spencer  11:34

I think Steven can probably speak to this more technically. But I think one thing that comes up for our clients on that is that if the automation platform is the centerpiece problem is, if you rip it out, you rip out you have to rip out everything. And the CDP is kind of a way to hedge your bets a little bit. So if you need to swap out Iterable with Braze or whatever.

Toby  11:58

Although the ideal is that you don’t buy a platform that you have to rip out, or you that occurs to me to be a second order problem to the first problem is having the wrong engagement platform and needing to swap. Now that’s fair to say that maybe the needs of today is not the same as the needs of tomorrow, and you do need to swap, but paying X amount per year on another contract just to have the optionality of turning off another I can see that as a problem that needs to be solved, but as a secondary problem to having the wrong engagement platform. 

Steven  12:31

Yeah, yeah. my opinion, I think in the case of I have a CDP to connect to my engagement platform, and that’s my primary use case, it’s probably, not a good reason to have a CDP, because it should be doing different things than that. I mean, that problem has largely been solved at this point, and that’s I think most CDPs would be more interested in not solving that problem. For you, it’s a bonus to them. The primary use cases I see for CDPs today are cleaning data, centralizing it for machine learning, or for destinations that aren’t as broadly, I would say, accepting of data from like engagement platforms like you mentioned, Braze Iterable, these guys, they’re very, very good at taking data, because that’s the their business, right? Like, if they can’t take in data, then they’re not going to win. But if you are trying to pipe that data into your healthcare CRM or to your only, your own built microservices or CRMs, or to Salesforce. You know, there’s a lot of places that getting data in and out is painful, and cleaning it and giving it governance is painful. And so there’s a lot of things the CDP, I think still very much solves for and I think, yes, for engagement platforms. Yeah, probably been doing this for years. At this point, you’re totally

Toby  13:47

You’re totally right. You’re totally right about all of that. I just would rather have. I’d rather the CDP do all of that. You know, you’re kind of doing some you’re you’re refashioning data, and you’re refactoring it in ways that it should be able to be accepted between two platforms that should be able to be integrated. And maybe this is a perfectionist world, but an unrealistic one, but ideally, if you need to connect your POS with your whatever HR data with your Salesforce instead, whatever it might be, that should be a problem that is solved by those platforms, because that is not a weird use case that is very much a standard thing that you should be able to do, and those platforms should facilitate that. But I agree.

Steven  14:27

And now your CEP is now focusing all their energy on getting more data pipes opened up and you’re a 

Toby  14:35

CDP should be a priority of any of these platforms. I think that should be the same. 

Spencer  14:40

It’s basically there should each your point Toby is that, or your take is that there should be a CDP within each of these platforms. Essentially, it should have its own capability. 

Toby  14:49

All these things are being like, we are a CDP plus CDP. No, just a CDP that you’re able to integrate effectively and like, that’s right. So, so yes, but. Put it simply, yeah, I suppose port data portability should be, should be important to every platform, and they should standardize methodologies for doing, for doing so, you know, it’s like having a, you know, an adapter for different connectors and having, you know, like, I wish the world would just use one plug socket, right? Yeah, we didn’t all have adapters so we can connect things together. And, you know, it just seems like there’s a, there’s a better way, you know, but, yeah, we’re digressing. 

Spencer  15:29

I mean, you’re the product guy it makes, you’re the one that sees it, and we’re, that’s why you’re here. That’s why you are, yeah, you got to dream it up before it can happen. Okay, that’s awesome. That was, like, one of the, probably the one of the best hot takes we’ve had so far. Exactly the spirit of what I was looking for.

Toby  15:47

My partnerships team, your partnership team is probably gonna call me up and be like, What the hell are you doing?

Steven  15:55

You’re product person. Like, you gotta rock the boat, you know?

Toby  15:59

Yeah, I do. I find that when I we speak to, you know, we have a PR team as well, and I do have hot takes that I feel like, when you’re speaking to a publication, you’ve got to have a hot take, otherwise they won’t print it. So you’ve got, you got to poke your oppos out a little bit. 

Spencer  16:14

You can’t just toe the line. Got to push forward. 

Toby  16:15

You’ll be a bit controversial. 

Spencer  16:17

So we’ve talked a lot about AI on a podcast recently, in different lights, but I think there was a when we were prepping for this, you brought up that there’s some confusion around, like, generative versus other types of AI tools. And so I think starting out, you know, with the state of AI and marketing right now, kind of breaking that down for us, yeah, would be really great way to start.

Toby  16:40

Yeah, everybody. It’s one of those things where AI, everyone says that AI, and the no one’s AI, and everyone tries to, you know, this is big thing right at the moment about AI washing. And, you know, the big thing with the SEC around companies saying that they now do AI, because it’s this, you know, it’s like the.com bubble, right? Everyone was a.com someone was telling me the other day that there, there was like, a guy who put, I can’t remember what it was. He was like a plumber, and he put.com on the side of his van. He didn’t have a website, but they put, you know, this.com because it makes you sound, you know, back in 2001 made you sound like this was your big deal. And I think AI has become that to some extent, where everything is AI because it’s supposed to be so that’s one problem where AI is kind of meaningless, because it can mean predictive AI can be used in the application of neural networks. It can mean machine learning. It can mean using LLMs. It can be agentic, where you’ve got LLMs making decisions. I kind of categorize it in a number of ways. There’s one is AI, which is just using it for no reason. And I’ll better put that aside for a second. And then there’s kind of two other platforms, two other types of platform, one which I would describe as predictive AI, which is propensity scoring, which is predictive segmentation, which is predictive churn, which is finding out which asset is the right one to use, what offer is the right one to use, product recommendations. And I would put things like Movable ink DaVinci in there. I would put OfferFit in there. I would put the Amp in there. I put, you know, CDP based stuff like Braze catalyst in there. And then you have the generative AI tools. And so you’ve got the foundation stuff. So obviously you’ve got Open AI and Gemini and all that kind of stuff, but you’ve got these kind of more SaaS Enterprise tools, such as Jasper.ai, writer, Copy.ai and all these other ones which are using generative AI to come up with kind of enterprise-level content or foundational model services across your business. So it could be not only for marketing, but legal and HR, whatever it might be that implements AI in your business, there’s a big directive and all of these large brands to businesses to implement AI everywhere and gain the productivity and efficiency gains that it brings. And so you have this generative AI world as well. I think where we sit, we try and get the best of both worlds when it comes to predictive AI and generative AI, meaning, when you have predictive AI, especially what I would describe as these content selectors or AI decisioning, where, you know, you might select from a fixed number of assets, select from a fixed number of you know, hero images, banners, might select a different subject line, a different body, whatever is that creates the need for huge amounts of content. If you’ve got 20-30 assets, and you’re swapping them in and out for different users and trying to make a personalized experience, you run out within days, right? And on a generative AI space, you can create all this content, but you’ve got no idea what to do with it. You know, you’ve got, you know, you can create different assets, and you can come up with tone of voice Based Message bodies and all this kind of stuff. But you there’s a human that has to then take it and put it in the right place on the CDP, or wherever it might be. And so what Jacquard, what we do is we combine the generative side so we can generate huge volume of content, and we have predictive sides that we can select who should receive the content. Where should they receive it and plug it into the right place and distribute that content to where it needs to be? So that’s where I kind of see it as see. I was at CES in Vegas a few weeks ago. And these businesses are very much kind of self-sorting into this kind of generative AI or predictive AI space. And I think that there’s still very much a need, or there is an increasing need, to combine these two things into a single platform. 

Spencer  20:22

Do you have a term for it yet? Are you working on it?

Toby  20:24

That’s a good question. I should, probably should, shouldn’t I? 

Steven

What about Jacquard AI? 

Toby

To be honest, maybe this is my second hot take of the episode where I think AI is going to turn into a bit of a dirty word, in the same way that no one calls themselves a .com anymore, like if the hypothesis is true, that AI is so great that is going to be pervasive in everything, it goes back to this. If AI is everything, then AI is nothing. And really we should be talking about performance. We are a performant content generation platform. AI is going to be in everything. So I when people say, Oh, we a we do this, AI, we do that, AI, or brand it as AI, I think there’s going to be a bit of a when this bubble inevitably pops, people are going to say, we’re not, we’re not that kind of AI where we know we’re natively, we’ve always done it this way, and, you know, that kind of. 

Steven

We are scalable AI.

Toby

Yeah, everything is like I say, so, so I think about it more about we utilize generative AI. We utilize predictive AI. But at the end of the day, we solve a real-world problem. And I think the a good way of thinking about it is, why do people open a message? Why do people, if you receive an email, push notification, SMS, whatever, why do you open it? It’s typically the who, so who sent it? So it’s the brand, it’s the where and the when, so the time it’s sent, it’s the channel that it’s sent on. It’s the what, which is the product or offer that’s in the message. And no one yet from this, who I can tell, thinks about the how you say it and so, and then the how, how it’s articulated to you in the words that are used. And so I think about us as the how, like we solve for the how you already do, the what and the where with your CDP, you know, send time optimization and intelligent channel, you’ve got your predictive AI to work out the product, you know, product recommendation of the offer, but nothing is has told you actually, don’t say, you know, I saw this demo the other day, and it was like these three different platforms of predictive AI and deciding the right offer and the right products, and, you know, the right time and the right channel. And the language was like, great deals for hikers, or it was like great deals for runners, great deals for climbers, great deals. And I’m like, you spent a million bucks on these propensity, segments and these predictive this, and you lost it at the last mile where you could have articulated that in the way that that person could have benefited. So thinking about the how you say something, that’s what we think about, and it means you can, you can add value, no matter what your stack looks like, if you don’t have the predictive recommendations, and you don’t have the predicted offers, and you don’t have your send time optimization, you still need to know how to say it. You still need to know how to articulate a generic message in a way that’s at least for the bell curve great for that campaign, even if you’re not doing it on a personalized level. So that’s what we think about, is, how can we say, how can we use all these decisions that you’ve made with all your data platforms, you know, predictive AI, how can we use that to make your brand sound engaging and resonant with whoever sees it,

Spencer  23:34

Your Jacquard, your contextualized personalization engine? Yeah, there you go. 

Steven  23:41

Well, I think I’ve heard that term. Well, that’s what we call hundreds of companies. 

Toby  23:45

We call it, well, we debated between calling it contextual relevance and personalization. Problem is personalization, to some extent, is this, it’s kind of like, AI. Have you ever written an email where you use someone’s first name in the subject line, like it’s, it’s no, I mean, no, it’s like, whenever I see an email that says Toby, you know, and I’m just like, that’s a brand, and they’re trying to sell me something, and it’s like, then they and they’re going, their marketing teams are going, we personalize that. Like, Nice job guys, well done. They use a merge tag. And we’re like, Finally, we’re personalized, and we’ve got we use the first name, and actually that has become distinctly unpersonalized. Now, when someone uses my first name, I’m like, You’re not really. You’re just kind of, there’s a gratuitous personalization when really, I want them to speak to me like I want to be spoken to. I want it to sound like someone sending me an email. 

Steven  24:41

You know, what I wish I could do is like, those databases that collect all that information about you that they’re basically reselling to these brands who are then grabbing that first name personalization. I wish I could, like, change my name in there. So when they do all the, you know, email address mapping or IP mapping where they’re getting that info, I would know that it was from a brand that I’ve actually. Actually never interacted with or didn’t give a legitimate name to.

Toby  25:02

So here’s a here’s a good tip. So I don’t know if you’ve seen my LinkedIn, done your research, but my on my LinkedIn, I have a world emoji just before my name, and I do it because when I receive LinkedIn messages, if the emoji is in the message, then it’s a merge tag. And if it’s not in the message, they’ve written it because no one would put the emoji. They wouldn’t go find the emoji, because they know it’s not part of my name. So I can now tell if this is an automated message or it’s someone personally reaching out. So you can do that.

Steven

I like that. Hot takes and hot tips there. Yeah, you go.

Spencer  25:42

We’ve established what Jacquard is and is not, the state of AI, and how AI and personalization and .com may all become, well already .com but kind of personal. It’s like how everyone was saying multi-channel or omni-channel at one point.

Toby  25:58

It just becomes table stakes. You know, I think at the end of the day, it becomes table stakes. And then when you start talking about it, you’re like, start talking about it, you’re like, Well, yeah, well, obviously, you know, I think AI is gonna become that where it’s like, oh, you’re an AI company. Like, oh, that’s what else it’s like being you’re an internet company. What does that even mean, you know? Like, we solve for the how you know. And I think that’s, that’s what we are as a business, yeah.

Spencer  26:17

So we’re talking about, uh, contextualizing it. We have a segment here where you where we’d like you to share three keys to effective AI, again, in this case, it’s the predictive slash generative together. Jacquard, way of doing things. But would you mind sharing these tips?

Toby  26:34

My sales team’s gonna get mad at me. There’s not one way of doing it. You know, like I said, there is plenty of value in a product recommendation tool, and there’s plenty of value. And as you guys know, there’s plenty of value it’s and a lot of time there’s a lot of a lot of value in doing something very simple. You know, the kind of the Pareto law of, you know, 20% of the work will be 80% of the value. I think there’s, if you’re not doing anything, doing something is going to give you big results. I think the way that we think about it is, obviously that we add value regardless of what you’re doing in the rest of your stack. So that’s the way we think. I’m not saying that that way is better than this way, but if you’ve got, if you’ve got predictive AI, and you’re doing all that decisioning, then great. And if you don’t, fine, you know, there’s not necessarily an order to it, I suppose, is what I’m saying. 

Spencer  27:16

So it’s, it’s less black and white and more like a Kinsey scale, where you know you’re meeting people, where they’re at.

Toby  27:23

Yeah, there’s a seven-dimensional grid of different ways you can, you can take this, no, I think at the end of the day, just making sure that you’re taking the approach of the needs of the business, and not taking it from an I need AI. You know, a lot of these, these business leaders are like, we need an AI strategy. I was like, No, your problems that you’re solving are the same as they’ve always been. It’s customer lifetime value, it’s average order value, it’s it’s repeat purchases, it’s retention. And what are the things, what are the low-hanging fruits that you need to solve for that? And sometimes it is. We need to have some strategy about, you know, we’re giving everyone 20% off. Maybe we should be a bit smarter about this. You know, maybe, maybe it is. We need to do some kind of cent time optimization, because we’re the kind of business where that has a big impact. You know, food delivery is probably one where send-time optimization is massively impactful. I imagine insurance, car insurance, renewals, I don’t think it makes much of a difference if you send it at 9 am or 5 pm, so the different businesses are going to have different needs about the kind of tools that they need. But I would take it from a business need approach and then say, Okay, what? What is the things? What are the things we need to solve for? And a lot of the time it is, we have generic messaging. We can’t we have very poor open rates. We well, we don’t have great brand affinity, and then the answer is, okay, well, how can we leverage AI or how? What are the tools out there that can help us build customer lifetime value, help us build retention, help us get more conversions. And if you’re either already doing that stuff or that stuff isn’t as valuable like I say in certain businesses, it isn’t then, then suddenly Jacquard the answer. But in other situations, it’s not. It’s all about, what are the what are the needs that you have? And I wouldn’t just take it from a let’s just throw AI at all our problems and see if see what sticks. It’s much more about being nuanced about it. 

Steven  29:20

So Toby question from my end, which is, you know, you’re obviously building a product around AI, and you’re building this, this sort of effort to, we’ll say, you know, generate a business result, right? Whether it’s increasing lifetime value or any number of other metrics that a company’s tracking. What does that mean to you? From Jacquard, right? We’re talking about, I’m gonna change the copy like, how, how much should I attribute that as the value versus the five other things that you listed as being the most impactful thing, right? Is it does the how. Me, it does the how equate to 100% of the value, if I was able to get them. To engage with it, or is it like 10% or 15% of the value? If I’ve done the predictive and the recommendations and everything else.

Toby  30:06

There’s not one answer, but the answer, the answer to your question, is that marketers have always been very good at testing, and you know, the way that we the way that we our product works, is, is that, at least for the first initial campaigns, you put in your human control, and you put in what you would have said if you weren’t using Jacquard. And we’ll test that with a small subset of your users, and we’ll see the impact that how has I mean, the way our campaigns work is, if you leave them running along with us, we will always beat human language. You know, we we wait. The way I think about it, is our predictive language, the way our performance prediction works is it weighs. It’s like a weighted dice, right? And it’s not always gonna, you’re not always gonna win, but if you’re a casino and you you you roll the dice enough times, the house always wins, right? And so the way our product works is we know we have a we have a predictive we generate language, and we know what language is more likely to engage, and we just generate a shit ton of that, right? And then if you put it into a campaign and you and you drop poor performing variants, and you introduce better performing, introduce more variants over time, the human cop, the human control, will always fall out at one point. And so the way that we think about it is it’s not whether we can add value, it’s when, it’s literally how quickly, and at the end of the day, and toward, to some extent, how much. And so it’s just about testing. It’s about taking what you would have done. We can even what we do with prospective customers. We go and get the language that they’re sending, and then we go send them what we would have sent instead, we said, oh, you sent this campaign. You said this. We would have said this. And on our performance prediction, said that we would have got 25% more clicks, 20% more clicks,

Steven  31:55

But all that copy is it’s not what you’re saying is. And again, in this sort of testing example, is one version for everybody versus a separate version for every individual user that you have, depending on how much data you have on them.

Toby  32:09

We have multiple products. So we have something called Audience optimization, which, which gets you the best-performing language for that campaign, assuming you basically have no data. You know, we can come up with better copy. We can take a billboard and say, this is language you should use, right with no personalization, and because we know over 60 billion data points over the last 10 years that this language, on average, engages with a greater number of people than that then, and so this is what you should use. And then we have our personalization, product, contextual relevance, whatever we want to call it, where it takes in the user data and then finds, instead of finding the bell curve of what is the largest cohort of users, we can take that up. We can do the whole thing. And then we can actually say, for this group or this person, you should be urgent and exciting or casual or friendly, and it should be a long message, or it should be a short message. But this group, it should be something completely different, and we found that. We found it with different brands ironically, maybe not ironically. Maybe this is obvious. The older cohorts, people who are 60 to 70, prefer completely opposite sentiments than those who are 20 to 30. All these brands spend millions of dollars trying to speak to Gen Z right, trying to work at, how can we be, how can we remain relevant? And actually, whenever you change your tone of voice, and you try to appeal to a certain group, you’re going to not appeal to another. You know, you can, you can be an old man brand, or you can be a Gen Z brand, very few of both. But actually, if you are able to change your tone of voice, to change the way you say something on a personalized basis. You can be something to everyone, to some extent. And there’s the final thing I say this. There’s a very philosophical question that comes out of this, which is, what does it mean to be a single brand and to have a single brand identity? If you can be 1000 different brands to 1000 different people, is there a point in a brand? Is there a point in Adidas or whoever it might be, Nike, to try and have articulate themselves in a certain way, if they can do it so many different ways to so many different people, and I think different brands are going to have a have to have to reconcile with a decision on actually, we do want to have the single identity, and we don’t want to be too personalized, or they’re going to go, no, we want to maximize, maximize engagement, and so we want to be different things for everybody.

Spencer  34:29

One of the fun things about all the of your different product lines, and also marketing in general, is that the first thing you describe, the first product, is sociology, where you’re putting information out to everyone, and then making judgments based on large swaths of the population and how they respond. Then as you get more data over time, it becomes anthropology. You are dealing with smaller groups, and you get to know these segments, these groups, and you you, you cater to them. But then finally. Tell you what, what you guys can do. It comes down to psychology, which is like one-to-one messaging, where you’ve worked your way down to the point where you can now personalize it, literally for that person.

Toby  35:11

It’s interesting. It’s, you know, I’m in Williamsburg, Brooklyn, and if I, if I personalization is not a new thing. If I look at the ads, the billboards around here, they are to someone like me, right? And if you go to Deep Queens, the ads look very different, right? If you go to the Upper West Side, the ads are very different. And that is taking to location. It is a zip code, is a fascinating data point, because it takes into account, typically, income level. It takes into account age. It takes typically, takes into account the number of kids you have, right because everybody in the Upper West Side are a certain age, has a certain number of kids, and they earn a certain amount of money. And so you can really that is a level of personalization to your point, even at an anthropological level. So it’s nothing new. We’re just able to do it in a digital interface, knowing those data points, and I’d love to be able to get zip code for what we do, because that that tell that tells you a lot about a person. 

Spencer  36:08

Steven weren’t we talking about something? There are some pitfalls about, like, we were playing the podcast with Laura. There’s some pitfalls about, like, zip code based marketing?

Steven  36:20

It’s the geolocation stuff where, like, if the, if the, if the geolocation point isn’t accurate, you’re making a lot of assumptions about where somebody is, that might be completely wrong. 

Toby  36:33

Yeah, geolocation and zip code is very different. I’m still the same person on a business trip to Indianapolis, right? My zip code is here, right? And if you’re telling me about some, you know, Country Club in Indianapolis, it’s not relevant to me, because I don’t live there, right? Yeah, but if you’re doing, if you’re telling me about a shop down the road in Williamsburg, then it is, it is relevant, even if I’m in Indianapolis because I’m coming back. So you zip code is geolocation. Interesting. Geolocation, I think, is very interesting for transactional and I land at the airport and Uber sends me a push notification. Geolocation is great for that, but, you know, I think in e-commerce, retail, financial services, all that kind of stuff, health care, zip code is far more valuable to me as a data point. 

Spencer  37:14

Yeah, being a person who is always thinking about making sure that what we do is profitable for ourselves and for our clients, for our final point, you know something that comes up whenever we introduce a platform partner, whenever we talk about platforms, if you’re talking to a marketer, they might have to dig in a little bit, but they’ll understand it pretty quickly, like what the what the value is for them, and they may even value what understand what the value is for The business they usually do, probably, but when you go to a CFO or a CEO and you have to talk about, okay, so it costs x, the ROI or RO AI, in this case, always comes up. So when you have those conversations like, How do you tackle those conversations? Like, how quickly incrementality things like that. 

Toby  38:03

There are hard ROI-type stuff and there’s soft ROI-type stuff. We think about it in two ways. There’s obviously performance. There’s you get more open to more clicks if you use Jacquard. And what is the value of a click on open to you? Well, our average increases by this much, and so our average ROI and performance is this. It’s a very simple, you know, we are quite close to the money in that regard. We can attribute, you know, if the human copy generated this, and we generate, we create, we generate that in terms of clicks and opens and conversions. It’s quite a straight line to show ROI in that case. And then there’s productivity and efficiency. Those are little bit more nebulous, then it’s okay. Well, your creative team spends this, this amount of time coming up with this content. When you come up with scalable solutions like Jacquard, when we can generate 10,000 subject lines, you know, that’s a hard, harder ROI, because you weren’t doing that anyway. And would you be able to do that with a with a team of copywriters, you know, 1000 monkeys with 1000 typewriters, like at the end of the day, the ROI in that case, is a bit more nebulous, and it’s a bit more tough, but, but what’s nice is that we can demonstrate both, you know, there’s quite an obvious productivity and efficiency gain by having a platform generate lots of content, and then as There is an immediate performance gain of using better content in place of what you’re doing today. So that’s how that’s how we think about it. At least, there’s also a risk element that I think is kind of understated, where there is risk in using AI when hallucinations come into play. But at the end of the day, there is humans in the loop. There’s approval processes, and it’s a bit like self-driving cars. To some extent, self-driving cars don’t need to be perfect, they just need to be better than human. Better than human drivers, right? And in our case, we can come up with brand compliant. You know, you’ve received emails that’s the say first name in it, rather than your actual first name, and they didn’t do the curly bracket for the F name. Yeah. Yeah. It’s not like humans. Error doesn’t exist, and we can ensure that your content is calibrated on brand compliant, performant in the right place to the right person. And there’s ROI in that that’s also hard to measure. But what’s nice is we have the fallback is your content will just be more performant using us, and everything else is gravy. 

Spencer  40:20

So for our copywriter friends out there is Jacquard, friend or foe?

Toby  40:25

I would answer that question as the invention of the paintbrush didn’t prevent artists, right? It’s another tool, and it enables you to be more creative. Jacquard enables to create copyrights, to be more creative and do let it’s always been about productivity, and productivity isn’t about replacing jobs. It’s about doing more with what you’ve got. And copywriters still use Jacquard, and to be honest, they actually us going into into a business and saying, content creative is important, and it makes a big difference. Copywriters are like, I’ve been trying to beat this drum for years, right? And everyone’s been talking about which offer and product to do and the time and channel, and actually, it’s the content that makes a big difference. And so we haven’t seen at least anybody shying away from or fearing what we’ve got. If anything, it is a tool to be used. And in the same way that Photoshop is a great tool for artists that allows productivity, it doesn’t mean that that role goes away. It means that it’s just another way to express that creativity. 

Spencer  41:26

And we’ve seen our creative team really embrace tools like Jacquard and Movable Ink, DaVinci platform, and, you know, other similar things for exactly what you said. But I know that there is a real fear out there of AI replacing jobs, so I think having that as a little vignette is really helpful, because we do agree with that.

Toby  41:45

And I don’t want to be I just want to add to that last thing, like, I would say, in the world of these Movable Ink’s Da Vinci’s and these offer fits, there is certainly those are tools which are, in my eyes, an arrow in the quiver of a marketer. We, I think we operate in a very different space in terms like, I say the how and the creativity of the language, rather than the selection. So I don’t want someone to think, well, I’ll either need one or the other. It’s very much like all of these tools represent ways to improve value in what you’re doing. So I think it’s important to say that we just take it from that approach of the quality and the performance of the language, rather than selecting between assets.

Spencer  42:23

Yeah, no. I mean, in a perfect world like our clients use Jacquard and Movable Ink and Hightouch AI decisioning for different reasons, but obviously.

Toby  42:34

And that’s how we’re seeing that we’re seeing that a lot, it just comes down to we are. Where do you put those dollars? 

Steven  42:39

Toby. I mean, I think, I think this is great in everything that you’re describing, like, you know, I can see the pile on value, so to speak, of having that, you know, the point at which we can get somebody to engage with the message. I think the only thing that I hear from CFOs and from CMOS around anything involving AI content is, you know, what is actually incremental, versus what is just changing the timing of somebody who is going to do something anyway, and channels like email and the certain ones that you affect, you know, if you will, if we look at these trends over a larger period of time, you Know, before and after looking you know, what was the performance of this audience in the control and in the treatment group beforehand? What is their performance afterwards? Like, if that’s not meaningfully changing over time, then regardless of how much-increased engagement we see in clicks and opens, right like, is this solving that, or is there something else that we should be doing to better leverage the tool? And I think whether that’s a, you know, do I need to look back at what am I feeding the platform to make it effective? Like, do I have enough offers? Do I have enough good enough things that people want to buy? So where do you kind of draw the line between, you know, this is where we’re helpful. And like, everything here is, like, you got to focus on, you know, your product, the your audience. Like, like, how do you kind of see that where, like, maybe you’re not solving, you’re solving, sort of a lot of things working. Well, yeah, where does, where do you solve it where, you know, it’s less within the message itself.

Toby  44:24

Yeah, I mean, it’s a good question. It’s going to be different for different brands, I think at the end of the day, though, I go that example of that brand that had all these different propensity models and offers and products, and then they send a very generic message, maybe to an interesting group of people, obviously having a biased opinion on this. But I think the how is almost the first thing, how you say something should be the first thing you think about, because there’s no point in spending a million bucks on all these different platforms if the last thing is generic. And we could, like I say, we can take a generic billboard and make it more valuable, or. Can take the product that’s recommended, the offer that you want to provide, the weather and the location of the user and their location themselves, and come up with a contextual message. And so there’s value to be had at whatever stage you’re in. And what you don’t want to do is, whilst these platforms are valuable in deciding what product and offer you can get and which hero image should go in there, and which footer goes where, if no one’s opening that email, right? If no one opens the email with the product recommendation, if no one clicks on the push notification to get to the in-app that’s doing it, what’s the, what’s the point? I think the how, at least, can be valuable no matter where you’re at but you don’t want to waste that it that that can be valuable, but, but the value compounds, right? If you do if you do it all. So at least that’s my that’s my view,

Spencer  45:43

Gentlemen, we are at time. I want to thank you so much Toby. You know, I got really excited for this conversation. We did the prep call, and it turned out exactly like we wanted. So thanks so much for being here. We are obviously big Jacquard fans and big Toby fans, fans, but yeah,

Toby  46:01

Thank you for having Thank you having me on. I’m going to do a shameless plug that I host my own podcast called The Array for any of your listeners. And I’ll do, I’ll, I’ll shamelessly plug your podcast on, on mine as well. Yeah, hopefully, hopefully your listeners got understood a little bit more about your cars. 

Spencer  46:18

Yeah, and share the link. We’ll put it in our, you know, in our LinkedIn posts and stuff, and, you know, crank those views.

Steven  46:26

Yeah. You know, I’m a big follower of The Array of Objects, so I’m really excited to listen to The Array now.

Toby  46:31

Yeah, yeah. So you’re, you’re exactly the kind of person who

Spencer  46:37

Nerds. Well, not kidding. I was gonna say Just kidding, but I’m actually not kidding. But I’m actually not kidding.

Spencer  46:43

All right, guys, thank you guys. 

Toby

Go birds, 

Steven

Bye.