Why You Should Throw Away Your Marketing Playbook

0:00If I'm looking to bring an epic marketing executive on my team, and he's going to bring this playbook right into our organization. he's got the golden ticket and you said something that was completely different to what I thought you would say and I'd love you to click on that. So the short answer is I never bring a playbook to a new organization. No two organizations are the same, no two situations are the same. And what I found is sometimes you have a product problem that you're trying to solve as a marketer in the way you generate demand and get the most out of the product. And so the playbook for that is different than when you have a market problem where the market is not ready it doesn't quite know what it is that you offer and you have a killer product it's just they just don't get it yet that's a different playbook you know even a data miner today i'm just blown away by the product i think the market is still trying to figure out how ai can be used in clever novel ways We want to invest in marketing.

0:58We want to get this story out about how AI and public data can be used in novel ways to solve something that really impacts the lives of people. And I thought that was brilliant. I was like, that's a marketer's dream. You have to have the discipline to say, I'm going to come in with a fresh set of eyes and I know what's in my toolkit. right there's a difference i have my toolkit but i don't always prescribe which tools i'm going to use until i've come in and understood the market to build a playbook hey everybody welcome back to marketing trends we are absolutely honored to welcome yet again another visionary leader someone who has really redefined what it means to create and scale new categories in the world of B2B SaaS and product-led growth.

1:56With a track record of leading, I saw six major product-led revolutions, our guest today has not only driven go-to-market strategies that have definitely yielded mass market adoption, but has also been the force behind several hundred million dollar revenue engines. Merle Namani, CMO at Dataminer, welcome to Marketing Trends. Jeremy, it's an honor to be here. Thanks so much for your audience to tune in. And I can certainly, you know, platforms like yours are essentially the kind of platforms that I used to use all the time. I still do quite a bit to learn from people who have experienced and done things that, you know, that can help me on my journey.

2:43And I hope our conversations today is one that helps many of our fellow community learn from each other and say, that's a pretty interesting idea, or that's a, you know, particular sort of strategy that I'm going to go pull into my marketing efforts. So thanks for having me. Absolutely. Now, now I know people will inevitably be Googling you and data miner, but I want to just one, take a note to data miner because data miner is certainly making its mark, right, for growth and innovation. I saw a multi-year recognition on Deloitte Tech, Fast 500, Forbes AI 50, Forbes Cloud 100. And also a very interesting distinction is a built in had them as a best places to work on the best places to work list.

3:27So a lot of momentum inside of the organization. You're within your first year there now. You're coming up on your first year. You have a lot of experience. How's it going for you right now in your world right now? What's happening in your world? It's been quite the exciting period. So when you think about companies in different stages of growth, right? There are companies that are at early stages of growth and there's the excitement from trying to incubate and get a certain product market fit established, right? And maybe it's your series A, B, Browns, and maybe even C, and then you go, okay, there's a certain appeal and a market need, and then you're excited about bringing that to the next level.

4:09I think we at DataMiner are beyond that product market fit, but the world, if you will, has really come to us in terms of a need that we are solving, right? So getting the connection between a market and an opportunity that is emerging stronger than ever to a technology and a competency that we have and uniting those two. So my job, and this is a privilege to me, is to lead a marketing team around this effort to scale the company into this. We're currently a pre-IPO company where it's scale a company with the discipline, the rigor, and the execution to becoming a public company model. And when you look at the market as I reflected on it, This market is about where geopolitical uncertainties are escalating, right?

5:04You're seeing a significant amount of tensions happening around the world. And a lot of the way we live, work, play is all determined by how we handle those situations, right? Whether it be on the physical infrastructure, the physical, let's say, governments, organizations, or whether it be on this cyber side, there's a tremendous amount of volatility that's happening around the world. And so a lot of what DataMiner does is essentially apply the idea of AI as a mechanism for scaling the way we see events, threats, and risks emerging around the world with the use of public data. And that's a fascinating area we can dive into.

5:49But as we connect those two worlds where AI has the scale and the breadth to be able to do things that humans cannot possibly digest in around the threat landscape and what's happening, and then to be able to do it with information that is freely available in the public. And often at times, Jeremy, it's the public data that is seeing these threats faster than any private intelligence organization or corporate organization. And that's where the power of what Data Miner is doing is coming to help protect, keep people out of harm's way, keep organizations, resources, infrastructure safe, and subsequently impact the way we live, work, play, educate, govern.

6:28And we're seeing that as the mandate for the marketing organization is to get that word out of how we can be a solution to finding resolutions and comp in a world that is becoming increasingly volatile. So is it accurate to say that data miner, because I'm not an expert in the world that you're in, but I just want to understand the offering. Essentially, it solves this need for kind of real-time, mission-critical, actionable information by detecting kind of early alerts, emerging events, things like that. Is it only in those sectors, security and crisis management, those kind of things? Or is this protecting assets and protecting making decisions across a whole bunch of categories?

7:16You're 100 % right. Think of it as an early warning, early detection system for threats, events, and risks that are emerging for the things that you care about, right? And that could be infrastructure. It could be people. It could be resources that matter to you. And I'll give you a couple of examples. But we think about it as informed, like one is we are the organization that informs the news agencies. So about 95 % of the news organizations, whether it's the BBC, the CNN, the Al Jazeera, NBC, they're all customers of DataMiner. And the reason is we tend to provide these early warning alerts to these news organizations, at which point reporters know what stories to go and go tackle.

8:03So when you see an alert coming out that it's a breaking news coming out on BBC or on CNN, well, it's really, these are data minor alerts that are then acted upon by the reporters and the news organizations that go dispatch to go to these hot sites where conflict or some sort of incident is happening. So that's one area where it's happening. The second is around public sector. Public sector in terms of the Department of Defense and the federal agencies, the State Department and all these, these guys are all customers of data miner. And if you think about NATO, NATO is a big customer of data miner.

8:43The mission is force protection. The idea of protecting our troops, keeping our troops out of harm's way. And so we can essentially use public data about emerging threats that are happening. So for example, when the Israel-Hamas conflict started, we were 27 minutes ahead of the Israeli government in terms of being able to notify them that an incursion was underway. And we actually sent them the video of the paragliders with AK-47s coming into the kibbutz. And we sent them that and we triangulated that with audio squawk box of 911 emergency calls with telegram posts, social media posts, that an incursion was happening.

9:27And we were able to package that all together with audio, video, text, and images to be able to come together as an alert that says, with sort of a summary that says, there's an incursion underway. And that kind of early detection warning saves people's lives and gets people motivated in responses and so forth. So that's on the public sector. And we do the same thing for corporates, especially when you think about what's happening in the Red Sea, right? You had the Houthi rebels that are active in the Red Sea. 15 % of global trade goes through the Red Sea. Now, that impacts, if you're a Tesla or BMW, you have these factories that are in Europe, and those factories are just in time, right?

10:14And those just in time factories depend on smooth flow of free trade and without any disturbances. So our customers in Europe and EMEA are very much tuning into the data miner signals to be able to know how to adjust the manufacturing, the logistics, the supply chain, based on what's happening in the outside world. I'm interested in kind of the balance of, because I see the value in speed, real-time information delivery, and also it's got to be accurate, right? And especially in the context of these high stakes environments, but how does data miner balance that, you know, speed, accuracy and getting it right.

10:53Cause you got to get it right. Especially in this scenario as you're talking about. Yeah. A hundred percent. Right. So, so it starts actually, you know, this idea. So remember like there is this notion of, do you have enough signals coming? That's the first part. And so we have over 1 million data sources. We're feeding over 8 billion data units per day. We're generating about 200 ,000 alerts per day. And so the volume of the landscape we cover, for example, Twitter or X is one data source out of the million, right? And so we cover deep and dark web. We cover everything from network sensors to flight aware for all air traffic.

11:34That's where the transponders. So if the flight has changed directions unexpectedly. It's going at an altitude that is lower than what it's supposed to be. Whether it goes dark and stops beaconing, these are all signals that we get that something is awry. It doesn't mean it's an alert, but it's one signal that something is wrong, right? Or something is interesting that is happening there. And so we get these million data sources that are feeding us and our algorithms ultimately then have to use something called predictive AI to be able to then stitch together what we would see as a potential event or risk or threat.

12:13And the way we do that is that we have 12 years of archived data that we have built on what is an event and what isn't an event. What is an urgent event? What is a minor? Now you can grade it to say, this is a priority escalation. This is an information you should be aware of, and this is something that you should act on right now, right? And that, so we have this 12 years of data archive. And how that's important is these LLMs have to be trained, right? And Jeremy, you know that from models, these models, we are not using public models. We're using our own foundation models. We built over 50 unique LLMs that are purpose-built for this application.

12:56And the reason why that's important is because the closer, the more proprietary the data set, the more tuned that data set is to a particular problem or an application, the lower the hallucinations, the higher the accuracy, and the lower the cost to operating. Because you don't need massive models. You need smaller models. You need those models to be highly tuned and highly accurate to the fit and purpose of what you're bringing. I was going to ask about hallucination with custom LLMs. And so how do you navigate that? You said it's more of a custom LLM. And then are you removing hallucination altogether?

13:34I mean, how accurate are we getting there? So essentially, our models are, we've now achieved a tremendously accurate way of operating because of that 12 years of data. But we are never going to be 100 % accurate, right? And that's just anybody who tells you that your models are 100 % accurate is just not living in. It's truly disingenuous to say something like that. But we have something called human in the loop. And the human in the loop, basically, the models are out of the box. It's proprietary. I won't give you the exact number, but it's a very, very, very high number out of the box. We are in terms of accuracy and confidence.

14:16So it basically ships out of the box for a very high percentage. And then the sections that we feel we need, the model will be flagged and says, we should have a human review this before it goes out. because there's something odd or something that requires a further review. And at that point, we then bring in what we call domain experts. And these domain experts are industry specialists, right? And they will then take a look and look at it. And we are able to then publish that as something that is now 100 % confidence, or that very close to honor. Because now we're essentially saying, we've had a quick review.

14:56we've made whatever correction that is needed with a human or maybe the human says, no, this is actually right. We'll just publish it because it makes sense to me. Because sometimes it's the context of what it is you have to get right. And that is where the machine can use a human in the loop. So we try to use those DEs and those domain experts for that purpose. But to do this at scale with a million data sources, with 8 billion data units hitting our platform every day, machines have to do, I mean, almost, you know, the large, large volume of this on its own and get it right. And that's the genius.

15:30And that's what this company has done really well. Wow. Incredible. Do you ever compete on any of the benchmarks, like the AI benchmarks? I know you can't because you're not competing necessarily with ChatGPT or Claude or Llama, but do you ever compete to see accuracy? Hey, are we beating them? Are we close to them? I've been looking at the MMLU Pro a lot lately and seeing kind of the results of that. But do you even look at that at all and see how they're doing in terms of accuracy and how you're doing? Because we build our own foundation models, right? At a foundation model level, we look for accuracy benchmarks, et cetera, to help us assess our performance.

16:13So with the foundation models, you can do that well. And because then you could say these are, you know, general purpose kind of foundation models that we can do multiple things off of. But then when you get very specific and these 50 LLMs, you know, we have an LLM for translation. Like we do 150 languages that we translate. You know, we have an LLM for computer vision. We have an LLM for audio. We have an LLM for, you know, like we have all these LLMs that are purpose built for all of this, you know, because our computer vision has to take a look at a picture and be able to decipher what do I see in this picture?

16:49Oh, I see a fire truck in front of a McDonald's. Okay, so how do I know it's a McDonald's because I see the golden arches, right? Okay, then I see a fire truck. Okay, what do I know about this fire truck? Oh, there's a little identifier at the bottom of the fire truck that is a unique identifier. And if I look at that, I know what that county, that fire truck is from and what fire station it's from. Right now I know it. And then I look at another, you know, so now it's piecing together the location of where this fire is happening at a McDonald's facility. That's all computer vision. That's an LLM, right?

17:27But it does that using multiple LLMs to piece it together. So we don't do benchmarking on that because it's very proprietary and unique to us. But it's the efficacy of how well that works through the accuracy, the other measures that we use to look at that. Amazing. So we talked a little bit about in the prep, and this was surprising to hear you say this. I definitely want to click into this because I loved your answer here. Yeah. You said something that was completely different to what I thought you would say. And I'd love you to click on that because your language was beautiful around that. Do you bring a playbook into these organizations?

18:02What's your perspective there? And it's interesting. So the short answer is I never bring a playbook to a new organization. And it's really a bit jarring because I've had boards and CEOs ask me, what's your playbook? And I say, I have none. I don't use, I say, no two organizations should ever, that I've been to, should use the same playbook. And then it's like, well, tell me why and what do you mean? And to me, and this is Jeremy, no two organizations are the same, no two situations are the same. And what I found is sometimes you have a product problem that you're trying to solve as a marketer in the way you generate demand and get the most out of the product in what you're working with.

18:50And so the playbook for that is different than when you have a market problem where the market is not ready, It doesn't quite know what it is that you offer. And you have a killer product. It's just they just don't get it yet. That's a different playbook. The other is sometimes you have a really strong sales organization and you have to just feed that organization and it's starving and it needs it. That's different than an organization that has a weak sales organization, but marketing is particularly, let's say, strong. And so I think the personalities, the culture, the needs, the product, the market, the All these are factors that come into where you look at an industry and you say, what's the right channel?

19:33Here, the channel that really works may be physical events and physical high touch, when you're trying to sell at a sales-led motion kind of a thing to executives who are the buyers and the decision makers versus when you're really going at a PLG motion where it's much more bottoms up in the sense of, let's say, a developer-led. where developers are your buyers, and then you have to go up market. So every market, every work is very different. So you have to come in and have the discipline and the fortitude to not want to repeat what worked in the last organization in this organization. And there are elements you're going to carry over for sure.

20:17But you should not come with a preconceived notion of what the playbook is going to be because it worked in your prior organization or maybe it worked in your prior and you're the one before. You have to have the discipline to say, I'm going to come in with a fresh set of eyes and I know what's in my toolkit. There's a difference. I have my toolkit, but I don't always prescribe which tools I'm going to use until I've come in and understood the market, understood the business problem, understood the operating model, understood the strengths of where we have the greatest opportunities. And I think even at DataMiner today, I'm just blown away by the product.

21:01I think the market is still trying to figure out how AI can be used in clever, novel ways. And so much investment has been put into building these foundation models, building the silicon, building the infrastructure. but what are the killer applications that only AI can do? And as you see the world sort of evolving and you see cybersecurity threats escalating outside of your four walls and things are happening in the deep dark web that can be early warning, early detection, how do you shift left? How do you get to that point where you are acting before it becomes a ransomware attack, before it becomes a credentials leak that undermines your brand reputation and your brand trust.

21:44And those are novel, those are things that customers are struggling with, but they haven't applied said technologies in a very effective way to solve that problem. And that's where a unique perspective, a unique approach, a new idea, a new way of doing things, using public data, using AI becomes really powerful, right? So that to me is why I don't come with a playbook. I don't try to, I discipline myself to try to say, I'm going to look at a situation as it is, and then use the tools that I have in my toolkit to build a playbook. Could you talk about the state, like the, maybe the stage and phase data miner was in, right?

22:20As you're kind of outside looking in, you're coming in, kind of talk us through what you noticed, what you saw needed to shift there. Now you're coming up on your first year. What are some of the things that have happened? So when I came, you know, data miner has been tirelessly building for the last 10 plus years. It's interesting because somebody said this to me, they said, we're an AI company getting into cybersecurity. We're not a cybersecurity company coming into AI. Our core competency is AI. So then it's like, okay, because we had a thesis of the world and we wanted to solve that, that we wanted to keep the world safer.

23:01We wanted to use real-time information in a way that can arm people so that they can make better, faster decisions that ultimately save people's lives. And that was the thesis of the founder. And that was, you know, and a great visionary in Ted Bailey. And a lot of that work was where, but it was a very underrepresented story in the sense that the company was under the radar and purposely by design, it was under the radar because it really wanted, it was focusing on a lot of the customers and so forth that it didn't really need that mindshare visibility. But now as a pre-IPO company with a global market need and appeal for what we believe is a real sort of problem set that we think we can help with, then the idea is, okay, how do we elevate our point of view around the world around how we want to see technology be applied to solving real world problems.

24:01And then how do we get mindshare established around that? So that was the mandate. And I appreciated the board's sort of willingness to say, we want to invest in marketing. We want to get this story out about how AI and public data can be used in novel ways to solve something that is really impacts the lives of people. And I thought that was brilliant. I was like, that's a marketer's dream. You have a phenomenal product, a phenomenal capability, but the market doesn't know it. It doesn't understand it yet. I want to sign up for that problem any day because then I can do my best work and we can sort of create that market together.

24:41So in this kind of PLG model, right, how do you balance the need for user acquisition, of course, with this importance of nurturing existing customers, right, to become advocates and drive organic growth? How do you balance that? Yeah, I think one is there's two parts. There is the, what you have to do to establish a very strong point of view around the world, around the way that you see the world and the role of technology to be to solve that. First, that's the category creation kind of efforts that we talked about, which is very much, do I have a compelling view around, you know, how public data and AI can be reutilized?

25:28utilize. It's not. People today are like, whoa, public data. I'm not even sure. Isn't that just social media? Isn't that accessible to everyone? What's so unique? What's so compelling? Why is that any different than the intelligence I already have? Because I've got all these other tools and I am the Department of Defense or I am the big governments. I've got all this. And so I think there's a need to create a narrative that one, public data generated by humans and peoples and machines are really valuable. And when applied at scale with AI can really change the way that you protect and serve your customers, right?

26:08And that is that point of view. Then, so that sets you apart in the way that people think of who you are and what problem you are solving, right? Right. Then it's the you have to establish the motions, the go to market motions to capitalize off of that. Right. And the go to market motions could be, OK, I need to create brand awareness, an umbrella air cover. Right. That includes thought leadership. That includes, you know, the ability to identify and have a have a sort of a brand identity, a brand sort of connection to that particular point of view. and there's an investment there. Then you have to have particular channels in place of how you go drive that.

26:50Then these are the demand gen programs that you would be leading, the campaigns that you're building to be able to go create demand off of that perspective. And that could be very specific to cyber, corporate security could be specific to news organizations and you could essentially go to market that way. So I think this structure of, I have a strong compelling point of view, My brand sort of echoes that point of view, and I establish sort of an understanding of that. And then I have these demand gen programs that then drive from there the actual campaigns that actually create pipeline. And so when you connect that, you have a good go-to-market motion that aligns to the vision, to the awareness, to the programs that drive the pipeline and drive the conversions that you need to support the growth.

27:42Wow. So are you finding there to be an increase in demand, a need to educate the prospect, educate the customers now? Do you feel like that's been a real focus on the content side of things specifically when it comes to product-led growth, where people, like you said, do they really understand the capability? We have to go out and teach them and show them what's really possible, or is it a different angle in terms of the content? No, I think the first most important one is really just educating and informing. Think about it. If you're a security professional, right? These people are, they do God's work.

28:22They protect organizations. They keep us safe. They keep us going home to our families. They make sure our bank accounts are what they should be. You know, these are, you know, they do God's work, right? And these people are overwhelmed in the way that they are tackling issues and threats coming at them. And I think what we have to do is figure out, understand their pain point enough to be able to then say, I get you. I understand your problems. I understand what you're trying to do, how you're trying to just stay afloat. And I think what we have tried to do is create mindshare around, don't be afraid of AI and don't be afraid of technology and new things.