A Conversation With Larry Kim of Customers.ai

Larry Kim is a man on a mission. Not content with bootstrapping and ultimately selling Wordstream, he’s swinging for the fences again with Customers.ai. Most of the conventional B2B marketing channels are badly broken, as Larry shares. Cold outbound is dismal; inbound/content is awful; paid ad rates keep climbing. With Customers.ai, Larry’s moon-shot is what he calls “intent-driven outbound marketing.”

What’s that mean? It’s a combination of intent and content; signal-driven intent (like a visitor spending significant time on select pages on a website), and deeply tailored AI-powered content, driven by thousands of personalized data-points, to capitalize on that moment of interest.

Host Matthew Dunn posed the question ‘Why work on this?’ Larry’s response was “This is the most interesting problem in sales and marketing! We want to displace big chunks of the sales and marketing stack through AI.”

When a been-there-done-that entrepreneur like Larry Kim is clearly “all-in” on something, it’s really worth paying attention. It’s difficult to argue with his assessment of the problem — B2B sales and marketing is an exhausting, turgid mess that doesn’t work. Larry’s vision of a replacement is compelling, but also a bit unnerving. Customers.ai is aiming for full personalization — every word and every pixel in the message custom-for-you. As Larry himself says in this conversation, “ It's actually not significantly harder to shoot for something like hard and valuable versus something dumb and easy. So, you might as well go for it.”

Want a glimpse at the future of B2B marketing? Check out Customers.ai and this conversation.

TRANSCRIPT

95 Larry Kim Customers.ai

===

[00:00:00]

Matthew Dunn: Good morning. This is Dr. Matthew Dunn. I am so excited to talk with Larry Kim of Customers. ai. Larry, welcome. I have read and followed you for years.

Larry Kim: Awesome. Thanks for having me, Matt. Great to be here.

Matthew Dunn: So you're one, you're one of the few people like shouldn't need much introduction, but Do you want to do the nickel introduction of Larry

Larry Kim: Kim?

Sure. I'm known for creating a business called WordStream, which was a world's largest pay per click advertising and Facebook advertising software company and agency managing over a billion dollars of ad spend for tens of thousands of companies globally. I started that, you know, in my twenties in my basement as a, as an individual, and then grew that to you know, over 350 employees.

And, and, and it was acquired by USA Today in 2018 for nearly 200, about 200 [00:01:00] million which was incredible because we, I only raised 18 million in, in, in venture capital dollars. So that was a spectacular. experience for my first venture thing.

Matthew Dunn: I would say that's knocking it out of the park, man.

Like really, really cool. And here you are starting all over customers that have former, previous name was a monkey related, right?

Larry Kim: Yes, we started out doing like social media autoresponders and stuff like that. So it was called mobile monkey, but we've really solidified on a much more valuable use case around sales, outreach, automation and data solutions for business to consumer companies over the last two years, I would say, and the company's grown from, you know, zero revenues in that product to, you know, mid single million.

Yeah. Mid single digit millions in, in just two, two and a half years. And

Matthew Dunn: there you go again,

Larry Kim: companies, about 40 people 40, okay. Yeah, [00:02:00] it just feels, it feels great. Like we're like, we've got an incredible product and team and just trying to do it again. That's all.

Matthew Dunn: So let me love you softball.

Customer outbound, right? It's, it's hard.

Larry Kim: Well, look outbounding is, like, you're talking about target, you're talking about outbound, kind of prospecting emails, like it's, it certainly has challenges. On average, you might be lucky to get one or two replies. Per 10, 000 emails sent in, in the meantime, though, what's harder than that is if you're doing this using like MailChimp or HubSpot or Google, Google Suite, like Google Workspace, like those.

Those platforms are dummies. They'll, they'll notice that you're getting this high unsubscribe rate, high complaint rate. And so just scaling that [00:03:00] is very hard because like you probably want more than one lead. You probably want like a hundred, a hundred leads. So that implies, oh boy. So if we're gonna have to send like, you know, hundreds of thousands of emails a month, which is damn near impossible to do because, because you'll just get kicked off of, you know, The traditional inbound platforms like, like HubSpot or, or MailChimp or whatever.

And, and you know, but, but, you know, the alternative is, is also pretty terrible, I think. So like this traditional inbound marketing for emailing, like inbound, you know, traditional newsletter building, like where you compile a list of you know, tens or hundreds of thousands of. People because they signed up for something five years, five years ago, and you just got to keep kind of sending out that newsletter every month.

Like, I think that's equally futile because they're no longer in market for whatever it is you were selling and. You know, that results in lower engagement, which, you know, the [00:04:00] Google, Google is not stupid. They'll just dump all that stuff into promotions tab. And so that's why you end up with like 5 to 10, 5 to 10% open rates.

So,

Matthew Dunn: So how are you guys threading the needle of this? Those two big challenges.

Larry Kim: Sure. Well, the thing that I'm trying to pioneer or evangelize is, is it's kind of a combination of, of, of both of those. It, it, it's I'm calling it intent driven outbound marketing. It takes certain elements from inbound marketing, like intent signals, but combines them with you know, outbound emailing.

So I mean, right. The, the, the, the strength of inbound is. Yeah, there's, there's intent and and it's helpful. It's, it's delightful. It's like, it's supposed to be something that you're interested in, in the first place. Yeah. But the downside is just like waiting, like, Oh God, I put my, you know, website SEO page and now I have to wait.

Like, you know, it would be neat if you could just, you know, more quickly connect that content with the types of [00:05:00] people who are interested in it. And so that's kind of the intent driven outbound marketing combo, which is then we can, you know, combine the intent signals with you know, the, the audience, you know, that that's interested in this stuff at the moment.

And. Fire off automated sales, outreach email cadences to, to connect with them. And so then you kind of have your cake and eat it. Do you remember the weakness of the strength of outbounding was that it was like really cheap and you could just. Get a big list and you don't have to do a ton of work.

The downside was, it was just not very scalable because it caught or, or you know, it just low, low low, low engagement. So if, if you combine the two ideas, you actually have something pretty interesting. So what, what, what could be an intense signal? So [00:06:00] it, it. Varies depending on your industry, but if you think about your target market and what are the things that your ideal customer persona does prior or in the period before buying something from your business.

It could be that they start installing software on the website, you know, so there's ways to monitor, you know, the kind of installation or uninstallation of software on websites as an example, or perhaps, perhaps if you're like, you know, like a realtor agent selling office furniture or sorry, office leases or, or, or an office furniture company, maybe you're looking for people who are in the, in the market of moving from one office to another.

So there's, there's a certain scent that it's given off by that. Like you can subscribe to data feeds of, you know, offices that are moving or offices that have raised a ton of money or offices that are doing a ton of hiring. Like there's different signals that you can harness, [00:07:00] which you can then, you know, use to trigger these outbound, generative AI.

Because now you don't have just a standard pitch, but a pitch that really sounds very helpful and thoughtful, and that can be like radically customized on a one to one basis you know, depending on what signals was kind of intercepted and also customized based on. The identity of the individual who is receiving the, the content based on everything we know about that individual, like their, their social profile, like their whole identity fingerprints, like we can, you can work that into the copy.

And so when you do this at scale, and maybe I could give you some more examples in just a minute, but just the general idea here, it's, it's. You know, this is the [00:08:00] most exciting thing in marketing, in my opinion right now, right? Like if you're trying to break into the inbound marketing scene in 2023, good, good luck.

Like

Matthew Dunn: the noisy out there,

Larry Kim: you're not the first person to come with a blog and content and this kind of thing. And, and if you're Trying to do like in my background is an ads like Facebook or Google ads, like good luck at just 10 million competing advertisers. Those, those costs are, are going to be extremely high.

And so like this is a really killer go to market motion for, for companies like right now, I think. What do you think?

Matthew Dunn: That it, it's exciting. It's also. A daunting lift. I mean, you're, you're describing sort of a sitting right at the hourglass, right? Between what are people after and when are they ready to, to hear from you and trying to do a, a rifle shot job of [00:09:00] saying, Oh, this guy's probably looking for my widgets right now.

Let me reach out. Let me be as personal and targeted and relevant and a lot of the stuff as possible instead of bugging him when he's not interested. But how have you started to get your arms around all of the signals and sources and triggers? Sure.

Larry Kim: So you know, that's one of the things my company is actively engaged in, in developing, you know, we're not the only companies that provide data and Emailing solutions like there, you could, you could build this all yourself if you wanted to like there's tons of signals out there and you could use Zapier to, to send signal from, from a to B to, to trigger it, to trigger an email, but, but like you know, we do have a.

My company customers on AI and you know, which I think is onto something really interesting. We, we do provide a lot of, you know, signals, particularly interesting for for businesses like like the website identification capabilities. So. This is like [00:10:00] someone like you probably have a website, you'll probably have a few thousand visitors a month coming in to research stuff.

And you know, the holy grail would be like, if we could only figure out who the heck these people were, right? Well, if you install my pixel, we can ID, you know, 15, 20, 25% of those, those people. Okay. Like this is Matthew Dunn or this is Larry Kim. And it can then be kind of systematically, you know, dumped into an email, an AI powered email followup cadence.

Okay. So this, the ROI on this thing is, is, is remarkable. You know, compared to traditional email capture, you might be lucky if you capture one or 2% of the people who visit your website. So we're talking like, you know, 10 times more. Email capture and you can configure the AI to kind of reflect [00:11:00] back the user's website journey.

Like in the outreach, it's not just, thanks for stopping by. It's thanks for looking at the rentals. Like, let me help you find a rental in Somerville. Like it can be very, very specific. Like even if they have a thousand offerings, you see what I'm saying? And then The other thing that we, we, we have is we, we have a lot of information about the identity of the individual.

So it's like, this, this is Larry. He's got two boys. He's married. He lives in Cambridge. Now he leans this way politically. He's got these interests. Like, so we not only can figure out the identity of people who visit the site, but also. We have a crap ton of data around this individual's interests, his behaviors, his past purchases, his, his or her you know, their you know, persona, like their identity.

And, and that was always hard from a marketer perspective to, to [00:12:00] to incorporate into actionable. Campaigns because it would be like, Oh, that's really interesting to know. Like a lot of our customers like reality shows or a lot of our customers, like, you know, like it was, it was like these one off insights that you would glean.

But here, you know, using the power of generative AI, you can subtly make remarkable, like infusions of personalization. Into sort of a template, a templatized kind of email that makes the difference between like a, you know, what is it? I think I was quoting one in 10, 000 replies for traditional cold, cold email.

Like these, these typically get 60, 70, 80% open rates. Yeah. So this is well, I think this is the future of email marketing. And the crazy thing about this, Matthew, is that if done correctly, It'll actually rejuvenate your crappy inbound [00:13:00] marketing efforts. Right. So it makes sense. Yeah. You, you probably have some crappy, not you, like people probably have some crappy email lists that they've been, you know, like I call it the, the, the pile of emails.

Okay. Like they've been hoarding for 10 years. It's like a house list. Okay. Yeah. That, that, that thing, it's like. It's full of all dead emails, basically people who like, there's, there's like five, 10% open rates. Like, so, so you have a few like that are interested still, but you know, most have moved on the reason why those emails get characterized into promotions folder is because there's not enough, like engagement.

To like the engagement score that Google uses for the promotions to have algorithm. There's not enough juice to get it out of promotions jail. And so if you invigorate this with like a few thousand using the same, using your domain score and all this [00:14:00] stuff, like if you start doing inbounding, I'm sorry, targeted outbounding.

Yes. With, you know, instead of 5% open rates, but rather 60, 70, 80, then Google starts noticing, Oh, emails from, from this domain. They're very interesting people that care about them. I should probably not be dumping them all in the promotions tab. So, so there's kind of a, I call it the virtuous cycle. It's, it's it's, it's, it's super super exciting.

Matthew Dunn: What debates have you had with yourself, with your team, with colleagues, with whoever. About the checks and balances around around privacy, volition control in doing something like this.

Larry Kim: Look I don't want to do anything like bad. Like I, I care very much about privacy. And you know, I, I don't want to do a business that does anything illegal.

So one thing that we, we, we don't do [00:15:00] is we don't do Europe. So like, there's just a lot of like regulations around GDPR and stuff like this, and there's, there's just no way to thread that needle at all. So we just, if we see something that looks European, it gets a self destructs. You know, USA, it's a little different.

We have good old First Amendment, Congress shall pass no law that in any way impedes the right of free speech. So, you know, they, we do have some regulation around can spam and stuff like this, but that is largely, if you look at the, the summary, it's more about opting out. So giving people the right to like.

You know, swiftly and easily identify who is sending these emails and to be able to up themselves out of them without much effort and in a quick and expeditious manner. And frankly, we have to do that because if you don't do that, [00:16:00] you'll get more complaints and the more complaints you get, your reputation will drop and it'll just screw up your entire delivery.

So, so. So we are a hundred percent cancer compliant and and you know, the way that we're like, ultimately the customer is the arbiter of whether or not something is interesting or not. It's not me. Like I might be like really excited about these emails, but like, if, I mean, you're, you know, it's whether or not they're engaging with it, are they clicking opening forwarding replying to them?

If yes, then that truly is, is interesting. And if they're, if it's like a five, 10, 15%. Engagement rate. Well, then who's doing this? Who's doing the spamming? Like, I think, I think these old newsletters are just a bunch of spam. Like if you have like a, you know, 15% open rate or 20% open rate, that's ridiculous.

Like there's no, if I send you an email, Matthew, there's no reason you wouldn't open it. Right. Like you would, you would open it. It should be a hundred percent like, or it should [00:17:00] be very close to a hundred percent. And you know, if, if Companies, you know, have this kind of false indignation, like, Oh, like this is bad, blah, blah, blah, like this, that's just dog dogma.

That's just like, I mean, how can they be pointing the finger somewhere else when they're sitting on these 10%, you know, engagement rates of total, you know, garbage, you know? So,

Matthew Dunn: well, what part in part, what I hear you describing for, for what you're, what you're doing is tightening the relevance cycle up considerably.

Right? Six months ago, I may not have been at all interested in on our fill the blank website on golf clubs. And if now I've decided I am and I start shopping and you've got a customer who's got essentially got that instrumentation and I'm using a consumer example, but leave that aside them reaching back out to me without making me jump through fill in the form hoops.

I might go, oh yeah, I really was browsing [00:18:00] those guys and I really did look at this, this, this, and this, so I gave them a big old signal, yes, I'm engaged. So them continuing the conversation seems a lot more acceptable than getting a cold email about something I'm not interested in.

Larry Kim: Right. Yeah. Intent driven.

Outbound marketing. There is an intense signal here. They have heard, they've heard of you in this case and they've actually demonstrated like interest in buying your project, you know, 80% of people who are on your website will be your. Future customers, like, you know, within whatever, whatever your, your sales cycle is.

I can describe the marketer's dilemma here, Matthew. And that is that marketer it's, it's very hard to capture leads right now. Like if you're doing ads, you're paying 50 a click and, you know, maybe you have a 2% conversion rate. So like you're paying a thousand dollars for these stupid email opt ins.

Yeah. And so like from the marketer's perspective, like they have worked so hard and spent so much money that they [00:19:00] will never give away that lead ever, even five, 10 years from now, like if they haven't unsubscribed, so I can, I can, so there's like a. Almost a scarcity mentality at play here. And so what we're saying is like, maybe email marketers wouldn't feel that way if leads were abundant.

Like if I could truly give you like, I mean, say, say you have 20, 000 visitors, your website, I have no idea, but just say the 20, 000, like if I could give you 4, 000 emails of visitors per month. That's probably like, you know, 20 times more than you're capturing using your pop ups and stuff like this. Then, then would you be like moved from a mentality of scarcity to one of abundance, where you're saying like, holy crap, I have so many emails coming in.

And what if the price of this was like, you know, like less than the cost of a postage stamp for these leads, not like a thousand dollars from Google. Okay. Then what you would be, you would change your mindset to being [00:20:00] more like. Now that I have so many leads, I need to maximize for deliverability rather than, than size of, of, of, of list.

So it's, it's, you know, in target outbounding, less is more actually. So like one thing that we aggressively do is we, we just ruthlessly, like the platform ruthlessly throws out emails that aren't engaging. Like automatically. So like, oh, smart. Yeah. So like MailChimp, there's a stupid feature of MailChimp that says like, you know, resend this email to non openers, non, non engagers.

Yeah. That's the stupidest thing you could do in your life. It's just, it's just, it sends all this signal to Google, to the people, like to the promotions filter. These are the people who weren't interested in the first place. And now you're going to send another a hundred thousand emails to these people so that, so that they're going to get even more signal saying that nobody opens this stuff.

That's the death [00:21:00] spiral. You see what I'm saying? So, so, so like, our thing is like, if, if the, if the engagement. Level like, you know, falls below us or you know, a high bar of engagement, we just throw them up because, you know, it's not, it's never been the case that if someone hasn't engaged with you, it's true that you do have to touch a contact a few times, but like, it's usually not the case where you know, after 12 emails of, of, of ignoring it and going straight to trash, they start becoming super valuable engaged on the 13th.

It's usually what happens is you show some kind of interest on the first email, but not enough to fully, you know, get over the hump, if you will. And and so that's where we would be okay with that follow up because, you know, there's, you're not beating a dead horse. There's still signs of life. You see what I'm saying?

Matthew Dunn: I, I get it. I got a message email this morning. From a sales guy at a [00:22:00] CRM company shall remain nameless. And he'd sent me a couple of follow ups, you know, Hey, we talked months ago and you said you were doing this. How's that working out? And I read it and closed it because they really didn't, there was nothing I was going to respond to him about.

So the follow up this morning was, Hey, I've been getting these notifications that my, that you're opening my messages. Is there something you need from me? Like I responded, I'm like me opening a message. And not saying anything back is pretty much the signal you should be acting on. Like, if I want to, if I want to mail you, I'll mail you.

Get off my back, right? And that's the experience with Outbound. It's like, God, the flood.

Larry Kim: So yeah, that's a frustrating experience. So I would argue that the guy is doing it right. Okay. So Well, from a deliverability standpoint it's

Matthew Dunn: not the same thing as a customer standpoint.

Larry Kim: That's you know, you, you would want to send more emails to the engagers, the people who are opening.[00:23:00]

Yeah. Yeah. Because then Google will see like, Oh wow, there's everyone who ever gets one of these emails from, you know, customers on AI seems to be opening it and that'll send the signal to the, the, you know, the process have to not dump them into the dumping ground. You know, it's, it's a little annoying.

And so, you know, if, if there was a connect, like, you know, stop sending me emails and. What, what I would suggest is that the company, you know, rapidly remove you from their list because you're at the biggest risk for complaining and it's the complaints that kill domains. Like if you're, if you're going to burn

Matthew Dunn: it.

I'm not making a point very well. I agree with what you said about the deliver, the value of the deliverability signal in that case. And yeah, the guy's doing it right. It's like, Oh, this guy's opened my last three messages this week. There's some interest there. Thank agreed, but if I wanted to reply to him, I had to reply to him, shut up and leave me alone is [00:24:00] what I don't agree with.

Larry Kim: Sure, sure. There's a pecking order of engagement signals. You know, replies are the best, forwards are the best. Link clicks. Click clicks are, you know, second best, and then the worst is just an open signal. Yeah, yeah. Like that's, you know, Especially

Matthew Dunn: in an MPP

Larry Kim: world. So, it may be that company should increase the kind of minimum bar for engagement to be like a click, more like, like a a reply or a

Matthew Dunn: forward, or I wouldn't have, I would be, I'd be okay with, and my guess is pretty sophisticated companies.

So my guess is this is what they do. If, you know, if I open three, four times. Nothing happens after that. There's a point where they should go. Just

Larry Kim: because I opening Oh yeah. Mul multiple opens. So you could like, and, and you know, our software does this. It's like, you could say like, I'm looking for three opens, or Yeah, five opens.

Or, or, or, but you know, as soon as they reply, send me a tech, like a Slack notification or something like this. Yeah, yeah. [00:25:00] Like there's a, you know, you can come up with a pecking order of, of What you consider to be a, like a fish on the line versus just a nipple and you know, you want to be efficient and you shouldn't set that bar too low because you'll just be annoying people who aren't interested.

Matthew Dunn: And, and they're the ones who, as you said, you really got to be careful of. Cause if I, you know, if I dumped them in a spam folder. That, that, that actually starts to create a quite a big signal that he doesn't want.

Larry Kim: Right. It's a, it's kind of a death spiral. Yeah.

Matthew Dunn: So yeah, yeah, yeah, yeah. It's interesting how big a role the inbox guys, particularly Google play here.

I mean, in, in, in, in a really simplistic sense, customer customers. ai is an AI grappling with Google's AI. Around domain reputation and promotions, tab placement and stuff like that. It's like you got, you got two Silicon entities doing a massive game of chess with each other.

Larry Kim: There's nothing like when I use [00:26:00] AI to send emails, there's nothing to finger.

There's nothing to fingerprint. Like if every email is different and customized, like not just in a kind of a template manner, but really different. It, it, like we can talk about different things and different features, different benefits and, and, and, and, you know, different products then you know, like Google can recognize a newsletter from a mile away.

Like, it's, it's the scent of a newsletter is very obvious to, to, to an, to an AI. It's the same email, like to everyone, except for like the name and a couple, a couple of placeholders

Matthew Dunn: change. Yeah. Yeah. And you change a much more, a much bigger. You've changed a much bigger map of the content than that, it sounds like.

Absolutely. Yeah. And how recent is, I mean, generative AI sort of blown up in public consciousness in the last six months, but just from what you're saying, I'm guessing you've been working at this for a good bit longer than that.

Larry Kim: Yeah, we've been working on the generative AI [00:27:00] for About a, just about a year or so, you know, I wouldn't say that we were ahead of the market by a super long, long time, but we just felt that we had such an incredible use case for AI so that we had.

All this data around users, website sessions, and all this data around people's identities. And so we felt that this wasn't just another stupid, like AI startup where you add a chat, GBT integration, and suddenly you're like an AI company. Like we, we do use our own LLM. We ingest the content of your website so that we can.

Speak your language. So we, we do develop our own LLMs for the, for the customers. And then we we, we customize that using our unique data that we have about customers [00:28:00] in the United States of America, 250 million consumers. So like it, it's I think this is an idea. I think this is very disruptive to the way that email marketing is done, which I think is really stupid.

Like the current way of like, well, let's make a big pile of emails and you know, a million of them, half a million, whatever, and run like weekly newsletters. I think that's crazy.

Matthew Dunn: Well, my thought bubble as you were describing the generative AI work there was like, this is not your grandmother's dear first name mail merge, right?

You're a long ways past that. And it hasn't, it hasn't worked. By the numbers, it hasn't worked in forever. Right. If I get a dear Matthew email, I'm like, Oh, for God's sake. Right.

Larry Kim: MailChimp even publishes like benchmarks for, for you know, open rates for different industries. It's like, you know, mid, mid to high single digits, which means, you know, that means like 90, 95% of people [00:29:00] aren't that interested in this.

Matthew Dunn: Yeah. Yeah. It's from a human communications perspective, it's a 95% fail, not a 5% success.

Larry Kim: It should, it should be the opposite. You should, you should be shooting for 90%. Yeah,

Matthew Dunn: no, absolutely agreed. But you can't get there. I think this is what you're saying. And I certainly agree with it. You, you can't get there with a conventional sort of, let's just, let's nibble at a couple of little things and hit send to a million, right?

It's like, it's the same content. Why would anybody give a crap? We can all smell it a mile away.

Larry Kim: Yeah, that's, it's, it's, it's. It works both sides of the equation. So the, the algorithm can't really fingerprint it as easily. But also like you're generating like these really delightful subject lines that really resonate with people.

And so it generates the engagement that is required to have the escape velocity from promotion to [00:30:00] have purgatory.

Matthew Dunn: Yeah, well, and, and to get the open, I mean, you mentioned subject lines, let's go there, right? It's like, I hate to say it, but I scanned down my inbox on who and what did they say before I ever opened the message?

And I don't think I'm the only person that

Larry Kim: does that. That's right. That's right. So this is it's just not really fair to like a you know, a A product like MailChimp, you know, we, we, we, we, we use that for, you know, our small, our modest newsletter and the, don't worry, we don't keep people on there for 10 years, but, but like, we, we, we do have a MailChimp, we do have a MailChimp subscription and it does have an AB testing, you know, subject line thing, but you know, we, we, we put two, okay, so it, it, it tries to choose between one or the other, like, This other one, like what we do for our prospecting, it's you know, customized at scale.

So it's, it's a one on one everyone gets a different

Matthew Dunn: pitch. So you guys, you guys are already at one of the future points [00:31:00] of email, which is honest, honestly, really real personalization of the message.

Larry Kim: Based on, based on the users signals based on their, their, their browsing thing based on their identity.

And then we're, we're expanding the universe of different things to key off of. I listed a couple ideas, like maybe it could be like a website signal or maybe it could be like a. Kind of like a search signals, like there's ways to figure out what people are searching for you know and like, like I'm looking for, you know, a lawsuit or something like, you know, either there's ways to, to you know, generalize this to, you know, a ton of other signals.

Matthew Dunn: Texting is. Part of your world as well. Correct. It

Larry Kim: is. It's becoming extremely regulated. I wouldn't recommend using something like postscript or something to do like [00:32:00] mass texting and like treating it like a distribution list, like an email distribution list. And I'm going to blast a hundred thousand emails.

Like you'll get hit with so many fines and et cetera, et cetera. Yeah. You know what I do. I think is acceptable, like nobody is, nobody is like reading your texts. Okay. Like what they're looking for is the complaint rate. Okay. And then if you get a high enough complaint rate, then, you know, Twilio or whatever your provider will show you off.

Okay. So you know, there are, you know, interesting ways to combine what we're talking about here with intent driven outbound emailing with texting in a way that doesn't generate a ton of complaints. And that is to, in your follow up sequence. Not just sending up follow up emails, but also, you know, Matthew, but with sending you a text message, like that's again, [00:33:00] personalizes scale.

So like has, has something to do with, with your identity and your intent. And that also references that this is a follow up. Yeah. Okay. So, so Meaning, don't take a list of 100, 000 things, phone numbers and blast them all your offer, you know, do an email offer to 100, 000 people, see who's interested, see who's clicking and replying, and then as one of the follow up options, use generative AI.

To generate the text message and then fire that off as a follow up to this engaged audience. And then like that, that won't generate the, the complaint levels that result in like you getting shut down or anything. In fact, it's just a follow up. And in fact, you can there's other forms of follow up.

Like I can use generative AI to then pump that into text to speech, like, like have a. It's, it's, it's, it's remarkable. [00:34:00] You can't, you can't tell that it's, it's, it's not a person. And you could then, you know, take the text of speech, take the speech file and fire that off as a voicemail.

Matthew Dunn: Oh, absolutely. Yeah.

Yeah. There's a camera, the name of the AI, but a relatively well known actor trained on his voice. It's like you can't tell the difference and I suspect that's going to become commonplace all too fast. You know, what I like a synthetic, a silicon Matthew, like Hey, take care of recording this message.

I'm busy. Sure.

Larry Kim: Oh, we, we, we, we can do it with your own voice. Like, yeah, that's what I mean. Yeah. Yeah. That's, that's, I mean, there are, there are ways to do that. Yes. Yeah. Yeah. So, so, so like, I think what I'm trying to paint here is like you know, a emerging channel that is not, not quite traditional inbounding [00:35:00] which.

Sucks for all sorts of reasons, competitiveness, you know, slowness, like amount of work to, yeah, to build, to build up that content empire. And that, but not quite pure outbound now either, which is hard to succeed with because of you know, low, like low interest and high complaint rates marrying the two using signals and.

AI outreach as well as multi channel follow up. So like all this text to speech, you know, text to video follow ups and, and that being kind of a significantly more interesting way to go about doing email marketing in the 2023 post, you know, AI world.

Matthew Dunn: Making it a much more intelligent and informed.

Channel, intelligent, informed marketing channel.

Larry Kim: People, companies have these stupid newsletters and [00:36:00] like, they have no idea how much damage they're doing to their email reputation and how little value that's creating. And you know, you know why they're doing it. It's just because that's the way that they've been doing it before.

And, and It's basically a strategy that was designed prior to the advent of the Google Promotions tab, like 15 years ago, like that's how pathetic it is.

Matthew Dunn: Yeah, yeah. And, and. And design in an era of very different volume.

Larry Kim: Yeah, yeah, where the signal to noise was like, you know, it used to be that everyone, you'd read every single, you know, email that came in.

It was a novelty.

Matthew Dunn: That was a long time ago. Right? You've got mail. No, no, no, no, no, no, no. I don't want that anymore. Wow. How, how, how fascinating. What a great what a great vision as well. What, like, what leads you to be pushing on this particular problem? Like you, you, you could probably do anything. Why this one?[00:37:00]

Larry Kim: Oh, this is the most interesting problem in sales and marketing. Like, like this right now, the customer acquisition is so broken after you're like, like those are your, your options. It's to compete in an arena of content marketers. You know, which requires, you know, not everyone can do that. No, no. Or to compete in a arena of, you know, overpriced kind of paid channels.

Yeah. You know, $50 a click kind of thing. A thousand dollars CPMs. Yeah. Yeah. So, you know, it's, nobody is making any money. It's, they're doing it at a loss. FOMO because because they're worried, you know, the boss will be on their Facebook and not seeing their ads. Yeah. You see what I'm saying?

Yeah. Yeah. It's, it's not this is such a, and, and this is such a strategic company. Like we can provide the, the AI. [00:38:00] The data and the tools, you see what I'm saying? Like this is a killer killer combination.

Matthew Dunn: Wow. Quick side question. Cause I want to, I want to keep track of your time. You've got a company to run.

You guys announced an integration with Clavio recently, correct?

Larry Kim: Yes. Funny story. Andrew Beilecky, the CEO who I teach a MIT business you know, MBA class with, he actually interviewed at my last company WordStream in 2010 for a VP of product. Yeah. My partner didn't like him. Can you imagine? Wow.

Like this is a 12 billion empire. Yeah. But but yes like those customers are great because they're typically selling like these, you know. 50, 100 e commerce products, like a blender or something, or, or rims for a, for a car talk for tires. So, you know, this is like, you can get leads for less than the cost of a postage stamp you [00:39:00] know, that have expressed intent you know, it's like extending the notion of a a cart abandonment sequence.

To, you know, a wider universe of people who haven't created an account on your Shopify store. Gotcha. And and haven't, or have an account with your Shopify, but haven't logged into it.

Matthew Dunn: Right, right. It's.

Larry Kim: Which is like, you know, yeah, 90% percent.

Matthew Dunn: I am struck and I've brought this up in a few other interviews with people.

I see some really good ballgames in the Shopify Klaviyo ecosystem. Some of the best executed. Email marketing, I've seen, at least personally, is smaller stores, not big

Larry Kim: brands. It's the Olympics of, of of email marketing. Like when I say email is busted and everything, like, you know, kudos to the e commerce people who are still like managing, using their wits [00:40:00] and their design.

Execution's still, you know, getting a decent return off of that. But you know, for the, for the rest of you

Matthew Dunn: Yeah, just, just kidding. Well, see, the nice thing, usually I wrap up, I wrap up these conversations and I say, hey, if someone wants to get in touch, where do they go? But in the case of customers.

ai 25% of people who are really curious to go visit your site are going to get a smart AI driven outbound. Message. As soon as they're done looking around. It's

Larry Kim: like, don't, don't call me. I'll call you. Yeah,

Matthew Dunn: yeah, yeah, yeah. Knock on the door, we'll take care of it from there.

Larry Kim: But we'll apply some filters and like, kind of size up your identity and your...

Try to determine whether or not you're ICP or not. If so, we'll follow up. If not, you know, then we will spare your inbox. How do you

Matthew Dunn: keep adjusting the machine, right? Because that's it. That's a complex balancing act and nothing's ever right. Perfect. We're in sort of [00:41:00] a probabilistic best thing you can do.

So there must be a, there must be an incredible amount of you know, relearn, retune on a more or less minute by minute basis.

Larry Kim: You're right. It's not minute by minute, but you know, algorithms can change, you can burn, you can burn an IP, you can burn an email you know, someone can make a bad complaint or something like things you know, pop up all the time.

But I think that's okay. That's, that's, that's fine. Like, like it's still on balance. Like, I think about like content marketing, like you still have all this work of like updating old content with, with new, with new videos and like, it's, it's, it's not like, it's not like any. Channel hasn't required excellence in execution before.

All right, I'm just saying that this is a lower bar to get started. The cost structure is significantly cheaper. The engagement structure [00:42:00] is significantly higher. You know, you, your marketing is, it's, it's your, your value prop. You're, you're you know, your, I see your unique, unique selling proposition.

And how you tell that story. I'm merely suggesting that, you know, this is a very strong, fertile channel to amplify that story. You know, at the moment. And for the, for the foreseeable feature for like, I, I like, I like, I don't see content marketing, sort of, you know, getting easier in 2024 and easier

Matthew Dunn: in 24.

Oh, it's gonna be a mess. Are you kidding me? Everyone's going to be dumping stuff in ChatGPT and hitting publish. Bah! It's going to be awful. Yeah,

Larry Kim: yeah. And I don't see the ad prices getting cut in half either, so.

Matthew Dunn: No, no, no. Yeah, yeah. It's it's going to be Mr. Toad's wild ride. It is kind of exciting, right, to see the, you know, the AI Sputnik moment sort of juice everybody.

And I said to a colleague the other day, [00:43:00] I haven't seen this kind of energy. Like since early dot com where everyone's starting to think about what's possible and speculate wildly and it's fun to see that come back, but it's going to upset a lot of structures that were, you know, working in an aging out and, and I think it's going to make some of the things that were already broken, obviously broken.

Now, content marketing is one of them.

Larry Kim: You're just seeing the tip of the iceberg here. It's just like, think about, our vision isn't just to have the best email campaigns and lead generation strategy. We want to displace the entire, you know, like big chunks of the Sales and marketing stack, like you know, through AI, you know, we're starting here, of course.

But there, there's more ambitious plans. You know, I

Matthew Dunn: love it. I love it. He's swinging for the fences again, man. That is really cool.

Larry Kim: Well, you know, it. [00:44:00] It's actually not significantly harder to raise a target than like to shoot for something like hard and valuable versus something dumb and easy. So, you know, might as well go for

Matthew Dunn: it.

Got to do the work either way. Well, that that that's the closing statement right there that what you just said. Lead line when we publish this episode. So, wow, we ran over time. My apologies. I knew it would be a fun conversation though. I'm really glad to have the chance to talk with you.

Larry Kim: Awesome. Have a great day.

Thank you. You as

Matthew Dunn: well. All right. My guest has been Larry Kim at customers. ai.