Bruce Cleveland, Traction Gap Partners

Clips
More from theCUBE + NYSE Wired: CMO Leaders

Bruce Cleveland

CEO

Traction Gap Partners

play_circle_outline Evolution of systems from record to intelligence

play_circle_outline Rapid innovation in technology compared to past

play_circle_outline Impact of AI on productivity and economic growth

play_circle_outline Shift towards agentic systems and AI

play_circle_outline Use of AI agents in marketing for research and campaign generation

play_circle_outline Market engineering and category design for startup success

play_circle_outline Thought leadership attracts customers and establishes authority

play_circle_outline Importance of creating a new category in the market to stand out

play_circle_outline Demonstrating traction through customer videos and success stories

Info
Transcript

Bruce Cleveland, Traction Gap Partners

Bruce Cleveland

CEO Traction Gap Partners

Bruce Cleveland, CEO of Traction Gap Partners, joins theCUBE in Palo Alto to delve into the transformative changes in technology and their impact on industries. As an experienced Silicon Valley figure and author, Cleveland shares insights on the evolution of systems of record to systems of intelligence. theCUBE analysts John Furrier and Dave Vellante host the discussion, exploring how Artificial Intelligence (AI) is revolutionizing the way businesses operate, affecting roles from Chief Marketing Officers to CEOs.

In this engaging conversation, Clevel... Read more

explore Keep Exploring

What has given rise to systems of intelligence according to Geoffrey Moore? add

What are some ways in which advancements in technology are changing the way systems are being developed today compared to in the past? add

What are some comparisons that can be made between the AI industry and other industries like the computer industry or the PC industry? add

What are agentic systems and how are they expected to transform the business model, especially in terms of software services? add

What are some of the challenges faced by startups and companies in terms of market research and how can AI potentially help in improving the process? add

What is one reason why many teams struggle with market engineering and launching new products successfully? add

What is important in becoming a thought leader and having companies come to you for knowledge? add

What is the strategy for creating a new market category and establishing dominance in that space? add

What should be shown to demonstrate traction in a B2B product, according to the speaker? add

bolt Powered by CUBE AI

Bruce Cleveland, Traction Gap Partners

>> Hello, welcome back to theCUBE here in Palo Alto. I'm John Furrier, host of theCUBE. We are in studio for the CMO leaders part of theCUBE and the new NYSE Wired communities putting together a new kind of trust network. Got a great guest here and said legend in Silicon Valley, Bruce Cleveland, CEO of Traction Gap Partners, author of the book, my favorite book, which I recommend to a lot of startup entrepreneurs in this generation. Really kind of the seeds of greatness on how to start a company are in this book. I recommend it to everyone. Bruce, you're a legend. We were just talking off-camera the early days of Larry Ellison, Tom Siebel, building out those early ventures. You've done that CMO at C3 AI and now investing, driving, racing competitively. Thanks for coming on.

Bruce Cleveland

>> Hey, you bet, glad to be here.

>> So, thanks for coming on. I mean, one of the things that I'm excited about in this era is that it's like the front end of user experience is changing and the back end of systems. It's like a systems revolution again, and AI infrastructure is booming, all coming together at this point in time, and it's changing all industries, and the CMOs are in that area of, wow, we could be more data-driven, but just every CXO title is looking at a business transformation.

Bruce Cleveland

>> Correct.

>> What's your take on this right now? And I want to get into it because I know you've got a lot of opinions on it, but more importantly, is it a special time? How would you describe this time? How would you compare it to other waves that had this disruptive enablement?

Bruce Cleveland

>> Yeah. Well, that's a great question. I think that I was lucky enough to begin my career when we were considering these things called systems of record. They have been typically built on top of relational database technology, which was just emerging as well. So this was the '80s. And then, on top of those databases, we built these functional systems for sales, marketing, service, support that have now populated every function. I think one of the issues that Salesforce said was, "Hey, we're seeing a saturation of the market." I think they missed their number a couple of quarters ago for the first time in many, many years and attributed that miss to saturation in the market. That said, these systems of record have given rise to something new that one of my partners from the past, Geoffrey Moore, calls systems of intelligence, and the systems of intelligence rely upon the knowledge contained, the data contained inside these systems of record. But, the user interfaces where we have these master detailed list difficult to use, enterprise software being replaced entirely. I think the metaphor going forward is going to be the way that we interact as human beings. We ask questions with prompts, we get responses. Those responses can be verbal, they can be through the machine, the mobile, wherever we are. And I think the way that we're developing this new technology is being built a lot different than when we built these systems of record. System of record were very proprietary. There wasn't anything open-sourced about them, but I think through these large language models, small language models, and through companies like Anthropic or OpenAI or now Deep Sync or whatever, what's happening is we're able to build faster and faster because these things are far more open and accessible. And so you're seeing innovation take place, I think, at an order of magnitude more than when we were, I dropped back at the beginning of my career 30, 40 years ago. We're just seeing lightning changes, which is the problem and the opportunity for the chief marketing officer and for CEOs and other startups and large companies.

>> You've seen many waves. I remember it's funny they call it the computer industry. Back then, it was the computer industry. I was pre-punch card generation. So when I was in college getting my CS degree, I said, "I never did a punch card." Because you started getting into structured languages, assembler and systems was around then, but they were databases. You made a query, you got some data back. They were great systems of record. You mentioned interfaces. Now it's like going to school, I get trained, I graduate now I'm inferring, I'm reinforcing my learning. So, this AI it seems to be another kind of big movement around similar impact as the computer industry. So, systems of intelligence feeds into the data model. So if you see the way... First of all, do you see this inflection point similar to the computer industry, early days, or open systems with the OSI model, or do you see it more of a whole nother thing? Because everything seems to change: the data layer's changing, the AI infrastructure's changing, pricing's changing. How would you compare the AI industry, whatever we want to call it? Do you see it similar to the computer industry, more like maybe the PC? I mean, what would you compare this scope of magnitude change?

Bruce Cleveland

>> Well, I think it's something that we have never seen before. Just take a look at two years ago, AI was sort of a term. I mean, we were exploring it, but now we're using it. There's hundreds of apps and tools that are emerging in all different functions. The difference, I think, today is that AI can actually generate things on our behalf. Whereas before, we were the ones that, we, the humans, had to generate it. So what's happening is that we're able to delegate authority to something to generate, create something on our behalf. That could be code, it could be a marketing campaign, sales campaign. It could be feedback on business plan, feedback on your business model. So what's happening is we are, I think, going to become a generation of far more productivity. With respect to companies, we can't cost control ourselves to success. As a country, we can't... Let's just take a look at the debt that we have. We take a look at the biggest elements, the bites at the apple of our economy. They're things like Social Security; they're things like Medicaid, Medicare. We can't really touch those programs without affecting people really negatively. So those are really off-limits, but by not touching those things, we can eat around the edges, but we're not going to make dramatic changes in our deficit. The only way we're going to be able to do this is to dramatically improve the output the GDP of our country. Well, I think we're going to be able to do that with these massive productivity enhancers through the form of these agentic systems and AI.

>> That's a great point. It's funny you mentioned that. I was talking to a friend about how impact the word impact is a term people use impact investing, market impact. People want to go to market, they have market engineering market traction, as you point out in your Traction Gap book. But what's interesting is that speed and agility become critical to make this work. And we were talking before we came on camera around some of the structural changes involved in compute pricing models. And it seems to be changing. Can you take us through your thoughts on this because you were talking about how standalone package-based software moved to the cloud, and now we have different model with more data-driven, which is usage. Take us through that and then what that does to the compute and the architecture that powers it. Because it seems like the two levers that are pulling right now is infrastructure advances, compute and hardware, and software and algorithm. You got Deep Sync, innovating; it's kind of clever, but it's not really new chips. Talk about the changes structurally in pricing of products and then the underlying infrastructure.

Bruce Cleveland

>> Well, there are definitely new chips. Just like we saw advances in chip architecture. We went from sys-based processing to risk-based processing. That was a major leap. We had to change the byte order and a number of different things, but we saw how Apple was built on top of Motorola CISC, and then it moved to PowerPC RISC, and then the entire OS got thrown. So those were really complicated problems, but we were seeing the Moore's law always in place at chips as we got faster and faster, but those were to address a certain type of compute. And today, with these new systems coming online, we need different types of processing. And we've seen what's happened with NVIDIA with their GPU. One could argue that a new chip architecture is needed, and there's a company called Groq, G-R-O-Q, that looked at that. The guy that created TensorFlow or one of the creators of TensorFlow is the-

>> Jonathan.

Bruce Cleveland

>> Jonathan is the founder of Groq. And so he would argue say, "Well, we need a different compute architecture to be able to handle this. So we're not paying the kind of prices that we're seeing under GPUs." And I think there's going to be a war down there at the infrastructure because the type of compute that we need is a different type of compute than we had. And so, the type of compute, we need real-time computational capability. Whereas before, we would have systems that would run these application software, yeah, that would be real-time transactional, but then we'd run business intelligence. We'd run this retrospective technology that would go through and process online analytical processing versus online-

>> Spit out an answer in a dashboard.

Bruce Cleveland

>> Yeah. It didn't matter that it took a couple of days to get those answers and give us insights. Today, we need instantaneous insights. We have milliseconds to respond to an offer, a prompt, whatever it is. And so we need different type of compute, and we can't just take the old technology and apply it to the new and expect it to be able that we can afford to do this. Things like tokens, we're have to start paying for these things. They may be sort of free right now, free from the standpoint to suck you like in the heroin dealer.

>> First hit's free.

Bruce Cleveland

>> Yeah, exactly. But-

>> Can we say that? I mean, we're digital we can say whatever we want.

Bruce Cleveland

>> But to your point, so that's one thing that's at the lower level of the architecture, but there's also business model. So, I'll just take what happened to us at Siebel Systems. We were the creators. Tom Siebel was the creator of the CRM entire category, and it was built on top of, it was a system of record, and it had all of the elements of what you would expect IT companies or IT departments inside large companies. They'd implement us on their servers, and so-

>> It was a gravy train for many, many decades. That was a great innovation.

Bruce Cleveland

>> Yeah. It was a pretty good model. It was a pretty good model. Tom did pretty well. And so the problem with that was, of course, then came along a different thing. And so where Marc Benioff didn't create the CRM category, what he did was he thought about it and thought different what Steve Jobs would've said, "He taught us to think differently." So the point here is, at first, people didn't really buy into this cloud computing thing, and I don't think that's what he called it initially, internet-based computing or whatever. But eventually, the term cloud computing caught on. I actually remember when Larry Ellison made a joke of it. What is this cloud computing thing? Is it going to rain or something like-

>> That still serves like the famous Churchill Club rant with what's his name from the Sun at the time?

Bruce Cleveland

>> Scott McNealy.

>> Scott? No, it was the other guy. Okay. Dave would know that. I forget. It was the classic Churchill when was like, "Cloud computing, it's just servers somewhere."

Bruce Cleveland

>> Well, names matter. So but, eventually, we all caught on that we could have this new infrastructure. So we went from single-tenant models to multi-tenancy models that caused a lot of companies a lot of concern. In fact, a lot of companies that started with single tenancy still haven't really been able to convert their technology to a multi-tenant cloud-based native solution. So just when we're getting ready and through that giant Gordian knot where we're arriving is we have these new systems called agentic systems, which are these AI-enabled systems operating autonomously in the background under a certain set of rules and guidelines that humans have given it to do processing for us. Workstream processing, a lot of different processing. Well, that is going to, I think, also transform the business model. So we don't see perpetual licensing in place really any longer. We see SaaS-based software as a service. Well, there's a new sheriff in town, a new type of SaaS called services as software, and this will be usage-based. And so companies that are going to be building these agentic systems, of which Salesforce will be one of many, are going to have to compete with a potentially highly disruptive model. Wall Street's going to have to figure out, in fact, this is a Back to Future thing. If it's usage-based, it may be more perpetual license-like. The good news here is I think it's an opportunity for SaaS companies. The opportunity, I believe, will be get in front of it and figure out maybe you can actually influence a quarter. Because remember today, if we don't have a customer on board in month, like on day one of a quarter, the revenue as we move through the quarter, by the time we're done, you could close a great deal. That's a couple of million-dollar deal. That's worth zero for the quarter, and then it's only amortized over the length of the relationship. So we're going to have to learn a new different sales motion for these things. It may end up being that we can end up saving a quarter with usage-based pricing, where we couldn't do that with SaaS.

>> It's a new kind of recurring revenue. Well, just to lock this in versus great insight. You're saying single-tenant architecture, which is how people run their software. I'm an enterprise... I buy, I'm a single tenant. I get the CRM, I get the system. Multi-tenant is like Salesforce. Hey, multiple people can use the service charge by the perpetual-

Bruce Cleveland

>> Sit in the same apartment building. Yeah.

>> He said it was software as a service, the same building. So, multi-tenant yields his way to SaaS. SaaS is subscription-based, and the next evolution is service as software, meaning you're paying by the usage. We're seeing evidence now the cloud meters, a startup in San Francisco, we covered them. So that's going to change the go-to-market and reporting. So valuations, you kind of hinted at The Wall Street angle where it's like, okay, I close a big deal for the quarter, the revenue recognition, okay, they're not using it today. Or maybe they will, or if they like the electricity model, they can use it forever perpetual. So what's the impacting from your standpoint as someone who's seen how investments get harvested, how people build teams, valuation? Is there a traction gap there? Do you market engineer the change? How would you advise people? Because I think people are generally agreeing with you that, okay, I can charge by usage, but what's the playbook? How would you attack that problem?

Bruce Cleveland

>> Yeah. Well, I mean, I've learned, I've taken over two businesses that sat within a larger business that used a different business model. And in one case, it was, I'll just reference Siebel. I took over the on-demand business. That was their software as a service piece. So you would go to talk to a rep about here's the offering, and they would take a look at their quota and what their commission plan looked like. And they'd say, "Well Bruce, I'm incented to close basically $5 million of business a year. Your stuff, the average cost, average contract value, half a million, million bucks. Your ACV, maybe 60K." He said, "Why am I going to sell your stuff and not this stuff?" And so, how does that manifest in the company? How it manifests is we get called in at the last minute. Salesforce, we find out, is in the middle of a company. We've been trying to sell our perpetual license product to them. And even though the Salesforce product isn't all there with everything that a company needs, they have an enterprise roadmap that expresses they will have it, and boom, we lose the deal.

>> Yeah, because it's 60K versus the big number.

Bruce Cleveland

>> Or worse, they bring me in two weeks... This is like I fly planes, too. And so it's like having your instructor put the plane into a stalled spin and then telling you to come up under the hood and look out, and there's only a hundred feet above the ground. It's your airplane, fix it. And that's kind of what happens. I felt like, oh my God, I'm being brought into a deal at the last minute because the account team knows it's going away. This is a Hail Mary. Let's see if we can right the airplane up-

>> -

Bruce Cleveland

>> I actually want to push down get off power, but to your point, you're giving these last-minute things. So, I encourage teams to get ahead of this. If you're an existing mature company, I think you really need to be doing the modeling. You need to be doing the financial modeling of what this would look like, and maybe what it looks like is a hybrid model. Maybe what we have is both subscription plus usage, and so there becomes... Maybe it's not as disruptive as from perpetual to subscription, which there really is no hybrid there. I don't know.

>> So, the innovation dilemma is basically compensation of deal. The old expression goes, your team's brains are connected to their wallet, and they'll be motivated, but whoa, whoa, whoa. What do you want? You want to do this? It could be great for the company. I don't get it. My brain's looking at the bigger number. I get comp; I can see the bonus. It's like, hey-

Bruce Cleveland

>> Makes a lot of sense....

>> sell me on this.

Bruce Cleveland

>> That's why they're called compensation plans. They're incentive plans. They incent us to do the thing that we want the company to do. The problem is that you can have the CEO say, "This is strategic to us." And the sales rep says, "Yeah, but I won't be around to see the strategy unfold because you're going to fire me because I didn't make my number." So I think we have to resolve those issues, and it sounds relatively simple, and you would think that we would've learned by now, but I believe that there's a bunch of Harvard Business case studies, thousands of them that point to the fact that business models are probably the most disruptive thing possible. It's not really technology. Technology can-

>> So you got to kind of burn the boat... So use that expression. But if you're a CMO or a leader, you have to affect that change and pull people through the change. So you either have to fire the sales force to bring in a new... What would you do if you were to affect the change? If you were a leader out there at a company now and you advise companies... So do the homework, hire from scratch, do a little skunkwork, spend a month, what's the operational risk management or accelerated plan to...

Bruce Cleveland

>> I'm going to point again to my great partner, Geoffrey Moore. I mean, he's phenomenal. And he wrote a book called Zone to Win. And that book was a manifestation of working with Satya Nadella at Microsoft and Marc Benioff at Salesforce, where this is probably a very familiar concept. He calls them zones. We have the performance zone of the company, which is the one driving the revenue. We have the productivity zone of the company, that's the one that supports the performance zone. And then we have these skunkworks group, and Geoff called that an incubation zone where we have different things working. What usually happens is a startup comes into the market and starts to eat away at the company. Sales rep says, "We're losing deals." They're kind of your sensor, if you will, to the nervous system of the company. And they go, "Okay, what's going on? What do we got going on? And skunkworks." "Oh well, we've got this great stuff." What's the knee-jerk reaction by companies? They take that skunkworks, and they throw it into the performance and productivity zone, and then boom, it fails. It fails. In fact, the data shows it fails roughly about 60% of the time. So why is that? Well, because comp plans, because the contract to give you an idea. When I took over the on-demand division, I was told I need to use a contract that we use for our big ticket enterprise deals that thick. And I said, this is a true story, "We have a deal going with Alaska trailer sales. The guy, he's not going to look. He doesn't know what residual clauses are, whatever. Can I just have one little tiny contract?" "No, that's the contract we have." This happens over and over again. And then-

>> Friction.

Bruce Cleveland

>> Right. Friction. So Geoff's idea, which I thought was brilliant, is why don't we create a transformation zone? We take somebody from sales; this is not a permanent group; it's a virtual group that comes together finance, sales, service, marketing, support, all the different groups, and they serve as a virtual team to take a look at the new product, figure out how do we have to package it, how do we have to price it, how do we have to promote it? What are the comp plans going to be? Revenue recognition: how is all that going to work before we simply smash it into the existing group? So, I think this is a great playbook for companies. I would highly advise people to read this because Geoff was not just pressuring about it, but I mean, he physically went in and worked with some very brilliant people to create this concept, and if we could use that, I think it'll pay off.

>> And his work has been phenomenal over the years, Crossing the Chasm, Inside the Tornado. I mean, that work is legendary. So the new zone thing is really guardrails to bring in innovation while maintaining the posture of the performance of the revenue and slowly bring it in, phase it in-

Bruce Cleveland

>> Phase it.

>> Like phasing in a new offense-

Bruce Cleveland

>> That's right....

>> in sports, kind of bring it in. All right, so I got to ask you, since we're the CMO leaders, you've been a CMO, but you also run businesses, you'd invest. If you're a CMO today and you're advising CMOs, you can go on one or two directions, become obsolete, or make it a superpower for the organization. How would you see the superpower side where CMOs become much stronger leaders like a CISO or a CIO or CXO where they can use that because they're marketing to the customer? So, at the end of the day, that's changing too. So, the old MarTech stack CRM systems can be abstracted away potentially with AI and agents. What would you do if you were a CMO or advising the CMOs to think big enterprise CMO? Look at: you have a superpower, you get the data. What do you do? What would be your advice?

Bruce Cleveland

>> Yeah, well, I mean, obviously, the classic answer, it depends, right? So, are you a startup? Are you a large existing company and anywhere in between-

>> Say B2B classic kind of mid to high-end enterprise; I got institutional legacy stuff.

Bruce Cleveland

>> Well, problem number one is that if you're a CMO, you didn't use any of these new technologies personally. So I mean, that's the problem, and you're running something. So, your personal knowledge of how to actually use this technology is pretty limited. And your team may not be that experienced with it either, certainly at the higher levels. Now, maybe new people coming in, maybe they've had some exposure to it. There's a lot of great courses to take, but that's not really your job as a CMO. Your job is not to run these systems, even systems of record. It's not really your job to run them. However, I do think it's your job to become aware of what's the possibility, what can these things do? And then, I would put together groups where I would have them go investigate certain technologies certain capabilities that I think might be relevant to how I want to go to market. The great part about it is we can actually make these technologies become members of our team that can somewhat operate themselves. And I'll talk about something shortly that I'm doing that I think is pretty fascinating with these agentic systems. But the truth is we'll be able to take these new agents, and they can actually plug into systems of record and other agentic systems and run them on your behalf. So you'll be able to instruct them. And I'm not talking about this isn't fantasy, this isn't like many years, this is now people are already building these. You'll be able to say, "Look, we need to accomplish... We want a campaign that does X, Y, and Z. We need these assets. We need whatever it's going to be. Come back and tell us what should be the buyer's journey on this. What would you recommend as the stages that we need to execute in our marketing automation campaign to drive X amount of growth?" And so these new agents will be and currently are, they may not be in commercial right now, but they are out there being developed. Will be able to take those ideas and generate campaigns on it, right? They become working members of our team. Now, some people find this frightening because now we're not... Oh my God, I'm going to be out of a job. Well, you might be out of the job that you currently have, but there's going to be new jobs. And so these are going to be jobs where we are creative, where we take our skills. A good example of this is large language models have broad understanding of things. They have book knowledge. So, I go to become a lawyer. I learn book knowledge. I go through all the classes. I study constitutional. All right? I gain knowledge. That doesn't mean I'm a fantastic lawyer yet. Now, I have to go through and I have to train as a lawyer. Same thing like a surgeon. I go and I go to medical school, then I have to go train to be a surgeon. And then, even beyond that, I have to actually do thousands of operations. So skills. So there's going to be book knowledge, and then I call it street knowledge. So we're going to have LLMs and all these different AI models that are going to have great amount of regular knowledge, but then we need to take human beings and take their skills and convert that into expert knowledge. And so the two combined, I think that's where the puck is going to go for us humans, is to become experts skilled in some areas, and then we can train our own agent to operate on our behalf.

>> I mean, you'd be winning, you'd be doing more campaigns, doing more things. It's like, I know you're a big driver. I now know you fly planes. To be good, you need to get miles under your... Air miles, and then driving is a skill. I mean, you know-

Bruce Cleveland

>> Very much so. Very much so....

>> first hand that driving the AI really is going to be something where the human is in the loop. They are in the driver's seat, so to speak.

Bruce Cleveland

>> Yeah. And I don't think you can just delegate this to an AI. We've seen the that gets created out of these, write a paper for me, and it comes back, and I went, okay, I'll never get those minutes back that I just spent reading that garbage. What I want to do is I think it's great to have the generalized knowledge put into it. So for research, because that's a laborious thing that's needed, but is of little value. All it does is authenticate and provide you with credibility that what you're saying has some tangible relationship to reality. Okay, great. When I was going to Johns Hopkins, it's a research university, and so you're taught how to write a research paper. Well, you have to cite everything. You have to be able to... And it needs to be peer-reviewed. It can't just be some random act of writing that you're citing. So I think-

>> Or written at two in the morning, and it's hallucinating.

Bruce Cleveland

>> Right. But finding doing the research work it's necessary but boring and not really very valuable. What we need to be able to do, though, is take that research, and then we need to convert it into something that's new and different and creative, and I think humans still have the upper hand in this.

>> So, the CMO has great opportunities. Let's get back to some of these things that fast. I like that the job that they think they have today won't be their job tomorrow. If you look at the B2B environment, there's always opportunities in these to do breakthroughs. So one of the things that Dave Vellante and I always talk about on theCUBE is stovepipes and silos, the old IT environment. You mentioned some of this new agent stuff. Some of the big successes today on the entry-level now is RAG-based systems retrieval augmentation generation, which companies like Glean and others it's enterprise search. But that was a hard nut to crack 10 years ago because of all the different schemas and different databases, you'd have to figure it all out, and it was just a lot of slogging. And then you're like, "Okay, I'm done." Yeah, you probably could punch it through for victory at the end, but it's like, what did I actually get? Right? Okay, now it's changed again, new data. So you see that really working the retrieval, okay, just abstract away. That's a breakthrough in my mind. That's something that was hard to do. Old architecture easier and achievable with AI. In B2B marketing, what would you say are some breakthroughs that would be in that similar vein where if you throw a lot of dough at it, a lot of bodies, we might be able to get through the other side, but would it be worth it when you get there that with AI is doable? Because what you were saying about agents, that sounds very provocative in the sense of if I can understand the metadata of all the different databases that are all unstructured and structured and different things, that would have been hard to write software for to know about it.

Bruce Cleveland

>> Well, marketing's a very broad spectrum of things. I mean, we could be talking about corporate marketing and events, PR, and we could talk about product marketing-

>> Both....

Bruce Cleveland

>> which could be market research. So the one thing that tends, and I write about this in the book, the thing that kneecaps a lot of startups and I... So here's true data I did for the book B2B: roughly 80, 85% of all startups fail. B2C, 96% of all startups fail, and then new products launched from existing companies. This is not in the book, is around somewhere between 50 and 60%-

>> Of new product launches?

Bruce Cleveland

>> Launches, right.

>> Net new products or line extensions or-

Bruce Cleveland

>> Net new products, new Coke versus Coke, that kind of thing.

>> That's still a high number.

Bruce Cleveland

>> It's a huge number. And so one of the things I posit I need to do... I'm going to try for the second edition of the book to do this, as market research is expensive. It takes time, and it takes a lot of capital, and so these are two things of which is in short supply for certainly startups, even medium-sized companies. Large companies, well, if you're in the consumer packaged goods industry like Procter and Gamble, fine. I mean, you've got PhDs and market research. Every time I go do a presentation to, I don't know, a conference, I just did one. I asked, "How many of you have a degree in product management or product marketing?" There's no hands that go up. Okay. So we look at, I said, "Well, how many of you have a PhD and understand how to do market research, stats and coming back and doing a proper survey and doing a null hypothesis and using things like one-way ANOVAs and know chi-square nows?" Nobody does that. I mean, well, not nobody but few do this,

>> Very few.

Bruce Cleveland

>> And most venture capitalists look at what the hell are you doing? Where are the MQLs? Where are the SALs? Right? Well, I think this is one of the reasons we're the victims of our own insanity, which is we need to do the real market research, but how can we do it faster better? We can have AI do this for us because that's what it's really, really good at. It can find all of the data that we need, and then now there's new companies emerging that will do the market... Will build the panels, they'll build the B2B panel or the B2C panel through APIs so we could connect the AI with the API. I'm throwing a lot of acronyms around here.

>> Yeah, I know it's all good.

Bruce Cleveland

>> But I think we're going to do that, and maybe we can change that horrible, dismal performance of startups into something maybe we just don't do it. We don't do the startup.

>> We get lost in our own legacy process, and I get handicapped when they keep the eye on the prize, which is the end-user or the customer. We were riffing earlier in the first segment today; I think I said the term low latency marketing. I mean, whatever you call it, being fast around researching what's happening so you can move on it, throw some new content, dial up a campaign in the same day things are happening, whether a DeepSeek came out that was happening. We got the weekend; texts were blowing up, we saw it a month earlier with their first paper and some early indicators. So a smart marketer would've seen maybe the DeepSeek stuff coming and then lasered it in on it. That would be a low latency execution, not, "Hey, we should do something with DeepSeek. It's all over the news." Not three days, only three days later. So, this brings up this concept of what's going on on the tactical front lines. How do you see that evolving? Because that's a super exciting area. Because what you're saying is agents will be assisting the human driving who wants to one inform, delight, get customers, or renew customers, or figure things out. How do you view that? Because I think that's an exciting area we see bubbling up once you get this new tooling or weaponry or intelligence.

Bruce Cleveland

>> Yeah. But I think all these things are tactical. I'm going to go back to what I wrote about in the book. I'm going to make a major point of this, which is one of the biggest problems is that we, and the reason I use... So, I'm going to explain market engineering is a term, I actually didn't invent it, but I did co-opt it. It meant something else for the stock market. But for market engineering, as I describe it and use it, is category design, thought leadership, storytelling and messaging, and positioning. Those are the five tenants that I describe in the book. The reason that I bring this out is because most teams are terrible at market engineering. They don't do a very good job of it, and the fact is if you try to play in somebody else's category, if you just launch a new product or a startup into a category that's already owned by an existing company, the likelihood of you succeeding is between slim and none. The data shows it.

>> Why are people failing so miserably in your mind? Not prepared?

Bruce Cleveland

>> One, because I don't think that they give category design its due. They think, oh, that takes a long time. I've read all about it. No, we need to start generating leads now. The problem is, and you'll see this, my best clients are the ones who've gone through this, where-

>> Scar tissue....

Bruce Cleveland

>> as the new investor, they go, "Okay, well, where are the leads? Where's this coming from? How are we going to make our revenue number?" We're short on MQL. Sales goes, "Those are terrible leads."

>> leads are crap.

Bruce Cleveland

>> Right. The fact is that no one's going to listen to... You can send out all the email you want. You can send out all of the LinkedIn ads that you want, but if nobody recognizes the problem that you're talking about and they don't have a name for the problem, the name for a problem is a category, then they're not going to be looking for it. And so, we first need to engineer the market engineer our category without a market. This is why I always say product market fit I don't believe in it. I believe in market product fit. Without a market, there's no need for your product. So we need to engineer the market. Engineer, oh, there's something that will solve this problem that I have. Oh, what's the name of that? Oh, it's a category name. So we go looking for it. This is where most teams fail. They don't do this work, and so they come back, and they say, "Hey, we're not getting great conversion rates. Our sales guys are calling out." All this stuff, and you come back, and you look, "Ah, it's terrible. The lead counts are horrible."

>> And the lag on that indicator is what? Months after? You're already kind of dead on the vine at that point, right?

Bruce Cleveland

>> I mean, it's kind of like the stegosaurus that gets hit, its tail cut off, but it takes a long time for that to travel to the head. So I think the problem that we have-

>> And so the time to react too is also-

Bruce Cleveland

>> Is also now probably-...

>> catastrophic....

Bruce Cleveland

>> different. So what we need to do, and this doesn't change with AI? If you say intelligent things, if you become a thought leader around a problem and you show other companies, if you can tell that story in an educational way, not in a sales way, then people will begin to listen to you, and they will begin to come to you to find out more about what you might know. You might happen to have a product or a service to solve that problem for them, maybe or maybe not. But they'll come to you, and I saw this in space. Tom said, when I came aboard, he's a master of this stuff. He said he goes, "Bruce, we need to create a new category for C3 AI. We're going to call it enterprise AI." Great. The great thing about a great CEO I don't have to... I'm not training him. He already knows this. So he goes, "Let's build this." Well, at the time, nobody was using -

>> No one was talking about that term.

Bruce Cleveland

>> Nobody. I mean, there was a digital RAG, a digital newsletters, not retrieval law manager, but a digital newsletter.

>> I know what you meant.

Bruce Cleveland

>> Yeah. Called enterprise AI, but nobody was using the term. And I go, great, we have a baseline. We're going to measure that. We're going to measure Tom in the news, Tom, and now I did have an iconic CEO, so it was a little easier, but still, you can manufacture this. So here's the thing. We engineered the market for enterprise AI. Then, we installed what are the attributes of that category. What do you have to do to be an enterprise AI player? And we threw under the bus all these point solutions and point things out there to say, "No, if you want to build enterprise AI applications, here's all the stuff. We built a billion dollars worth of stuff over the 10 years." If you don't have all that stuff, you can't be in the enterprise AI category that's what we declared. So my point is for these startups is, you need to create a category and maybe another analogy if you bear with me.

>> Absolutely. Yeah.

Bruce Cleveland

>> So, a category... I think this is great. We played the Super Bowl yesterday. Some people are probably happy; some people are not as happy. But we played the Super Bowl. Well, we played it in a place, it was called the Superdome. Well, a category is a stadium. It's where we play the game. Now, we can play different games in that particular stadium, but in this case, we played football. And football has rules. So, rules are the attributes of the category. So if we're going to play a game, which is we're going to be in the market, then we need to create a category for ourselves, a place where we play. We can't play in somebody's home field all the time. We have to create a different one because the home team tends to win. So we want to create a new category, something that the home team or that somebody else doesn't already own. So once we have that, the attributes, then we install ourselves as the thought leader of that, as the number one team because we have better players, etc. Players are features in a product. Teams that have the players, they're the brands. So we need, and the ironic part is once we build the category, we actually need competitors. So we have to invite other people into our category. This is what I spend time on with my teams to explain this metaphor. And so, with C3 AI, we didn't have any place to play, and so we created the category to play the game. And now, if you go look at enterprise AI, you'll see a lot of people are playing that particular game.

>> Yeah. And by the way, he was on theCUBE. I remember that eight years ago, seven years ago, he was on theCUBE at re:Invent. And when you guys were first on that, it's funny you mentioned the arena in sports. It's why New England always won it in the winter because they had home-field advantage. They had a great running game and good defense, and they beat the crap out of the Colts, who was a dome team. They were not used to the cold. So their attributes were designed to-

Bruce Cleveland

>> A different game....

>> different. They were a home game.

Bruce Cleveland

>> That's right.

>> They played the home game. So, the winning strategy is make the winning the incumbent play road games.

Bruce Cleveland

>> That's right. So you wanted to affect us at C3 AI was we had no SCRs, we had no BDRs. Sales development reps, business development reps, great model for subscription. We were a subscription company. We had none, but I had 3,500 inbounds a month as a result of doing this. The hardest part was vetting all of them to find out who's a student, who a consultant, right? Who, whatever. But I didn't have to spend money on outbound. Now we did it some outbound, we did some Adword stuff or whatever, but I didn't have anybody doing cold calling per se. We are getting people to call us, and that really is your objective. And you can measure that. You can measure it by how many people are using your new category name. How effective are you by going out to conferences and speaking the book that you write, the definitive guide to whatever, your technical white papers, white papers, you can do all this. It's a digital marketing game.

>> Bruce, how does a startup who might know that they have a superpower coming, that it's a category creator, a net new category, knowing that they don't yet have the resources to defend? Do they stay stealth? They go incognito mode in another category? How do you look at the resource-based competitive strategy for a startup that may want to be nimble and either silent running the field submarine analogy? Because if you expose, because too many people get, hey, I want to tell the whole world all my secrets because you got thought leadership. So thought leadership is a double-edged sword because if you're out there talking too much about what you're doing, some would say... Or someone would say, "Tell everyone your idea." Is there a playbook to say, certainly low profile, before you open a category up if you don't have the resource to defend it or if you're too early, how does a company deal with that? They have traction, but they want to go a little further before they unveil their thing. How would you look at that, or is that kind of off? Not relevant?

Bruce Cleveland

>> No, I think it is. I think people are too worried about it, though. I think what you first need to do is to have people... You need to get out there and have people listen to what you have to say and then come to you. So how do you do that? Well, you do it in formats like this. Don't create your own podcast. Go to somebody else's podcast. It already has a whole bunch of people who listen to them. Don't write your own newsletter. Go to something else. So don't have your own blog, go to... Don't start with yourself. Remember, the category is a problem. You don't have to say exactly how you're solving the problem, but you can talk about the problem. You can market the problem. If you market the problem, people will believe you're trying to solve the problem, not necessarily sell your product.

>> And might share information about their needs.

Bruce Cleveland

>> Their needs as well, which will give you more insights into what you need to build. One of the things I counsel people is that don't take your sales deck and use it at a conference like TED or whatever. People aren't there to be sold like that. I mean, maybe if they go to your booth, they want to see a product demo, but if you give a keynote speech, it better be something interesting. They're there to learn, and so an investor deck is really a combination of a sales deck and an educational deck. So make sure that you've got the appropriate deck for the appropriate audience. Talk to them. And also, if you're a startup, remember you're selling a financial product. I know you love your baby, that's great. Awesome. But the person you're talking to, especially if it's a B2B product, will never have even held a job that did that job. You'd be better off bringing videos of five different companies you videotape while you gave a demo to them and have a VC watch that demo. The demo okay, I get it but don't-

>> Show the traction.

Bruce Cleveland

>> Yeah, show the traction.

>> Well, Bruce, great to have you on. Again, going a little bit over because it's great content. The Traction Gap is a great book. And we heard here he's got some new additions coming, a lot of new stats. Again, really an excellent book for anyone either starting a company or trying to do product-led growth inside a company, as AI will give more entrepreneurial thinking and process inside the companies. Bruce, great pleasure to have you on theCUBE and part of the NYSC Wire CMO leaders great to have you on. If you're willing to drive the bus and drive the AI car, race car, you got to get the data. You got to know the process. Of course, you got to check into the CUBE here and get all the data from us. I'm John Furrier, your host. Thanks for watching.