Speaker 1For those of us who've been around for a while, you can see that this is potentially a house of cards that is going to fall, or at least that's how it feels to me.
Speaker 2That's Amy Conary, the person who literally invented the term SaaS. That term you use every day came from her head, and that's why Amy is such a fascinating guest. She was there in the earliest days of SaaS and helped shape it from nothing into what it is today, and that means she has valuable perspective on the agentic error as it starts down the same road, as well as some of the serious challenges ahead.
Speaker 1I spent some time a couple weeks ago at JP Morgan headquarters and spoke with the head of AI investment there. So I said, how much do you understand or are you looking at the commercial models that are developing right now? And he's like.
Speaker 2We also debate whether circular financing amongst AI companies is a genuine problem. We address why proofs of concepts are playing a very different role in sales cycles today. And Amy shares a new measure to better forecast which of your customers' revenue is truly sticky. What's really important is you can have two customers with the exact same usage, but very different leverage. If you're a revenue leader navigating the AI era, this episode is for you. Welcome to Topline.
Speaker 3Hey, everybody. It's Sam Jacobs. Welcome to the first ever in-person Topline recording that we've ever done. I'm joined by my hosts, AJ. I also just got here, so we're working out the bugs. And we've got our guest for the show, Amy Connery. SVP of something. Yes. At Zora. We're figuring this out today, actually. We're figuring this out in real time.
Speaker 1Anyone at home has some good suggestions. No, my email signature says founder and chair of the Subscribed Institute. But my workday title is SVP of Customer Business Innovation.
Speaker 3At Zora. At Zora. Give us a little bit of your background because my fun fact about Amy Connery is that you invented the term SaaS. True or false?
Speaker 1It is true. Tell us about that. It's not in Wikipedia, though, but I'll explain.
Speaker 2Are you working on that?
Speaker 1I do. Yeah. So I'll get to that. So started my career in AS400 ERP software, client server, perpetual license, very early days, booth babe situation where I would go to the trade shows. I was the booth babe. First job in software, I was fairly good at talking with people and having conversations. And I found our sales model to be incredibly extractive. And it wasn't just us. It was the industry. But back in those days, you'd sell a license, you'd get the revenue up front, you'd ship the customer the software. They spend millions of dollars and lots of time to try to figure out if they could use the software. You got your money. And then you also got your annual maintenance, which was required.
Speaker 3I'm just realizing this. Did you get to book the entirety of the revenue all at once? Wow.
Speaker 1Yeah. So you send an email and they got the email and that was the revenue recognition event. Oh, my God.
Speaker 3They must have resisted SAS for that reason to a certain extent.
Speaker 1Oh, yeah. And me meeting all these customers and thinking, oh, we're going to help all these incredible companies. And the reality is we had no idea whether we were helping them either because we were mailing physical CDs. And I remember one of the companies in the space, it wasn't our company, but another one tongue in cheek used to call the box that they would ship the coffin because it was that big. It was a huge box full of CDs. They'd ship this thing. Of course, it's also for because companies were going out of business and going bankrupt trying to deploy software. So I love the idea of working in the software industry, but thought this model is super broken. How do I like what can I do to fix this? It was really important for me. Sort of a common thread is to figure out how does this get better for customers? And so I decided to become an industry analyst and I joined IDC in a program called ASPs. So my boss at the time had coined that term, which gave me kind of motivation.
Speaker 2So ASPs? Was average selling price at that time?
Speaker 1Application service providers. No, okay. They're totally awful. Yeah, ASP is ASP. But there was a market called ASPs, which were these companies that would, you'd buy your software the way you always had, but instead of having to install it in your own premises, these companies would host it for you. So it was right in the beginning of really reliable connectivity and people getting their heads wrapped around, oh, my software doesn't have to be sitting inside my own data center.
Speaker 3In some ways, the beginning of the cloud.
Speaker 1Beginning of the cloud. Exactly it. But the problem with that model was that you still had to license the software the way that you always had. The only thing they were doing for you was hosting it for you. So around that same time, there were these companies that were developing inside strip malls, people's apartments, and I was one of the newest analysts. And they said, well, we're seeing this happen. It looks like software, but we can't count this in the same forecast that we're counting SAP and Oracle because of what I just told you about how revenue is recognized almost entirely up front. These companies are selling software differently. And if we put them on the same chart, it's going to look completely apples to oranges from a revenue perspective because we were counting recognized revenue in the software forecast. So I was charged with going out investigating what these companies were doing, defining it and giving it a name, which to me, it looked like software being delivered as a service. There was a services analyst at the same time. So at IDC, we had software group, services group, hardware group, services group, services analyst also sniffing around, really interested in covering that. And the way that this worked is I had a P&L and if I cover
Speaker 3it, I had to claim it as
Speaker 1software so that I could claim the revenue in my P&L for when I sold my advisory services, right? Because it's like speeches and reports. And like if the other analyst calls it services as software, that's going to be a different P&L covered by a different person. So I had to really, I mean, and it was, I had to make the case. Did you have a backup option?
Speaker 4Were you like, okay, there's three options. This is like, they all begin
Speaker 1with the word software. Yeah, I was all in. Software had to be the first word. And so I said, it's software as a service. And I defined it by a few parameters. One is delivered over the cloud. Two is paid for over time. So I did not define it as it had to be subscription or seats, by the way. The metric was not defined. It was just paid for over time. And revenue is recognized over time, as the services delivered. And those were the distinguishing characteristics. I had to come up with a forecast. And so at the time, nobody would tell me this is literally Mark Benioff's apartment in San Francisco. NetLedger was, well, NetSuite was NetLedger in a strip mall in San Mateo. They weren't making any money, but I guessed worldwide it was $40 million. So that was the very first SaaS forecast.
Speaker 2Okay. So did you, did you say SaaS at the time?
Speaker 1Yeah. Yeah. Yeah.
Speaker 2Did you capitalize the two S's in lowercase, the two A's? At the time as well. Okay. Yeah. She's the queen of SaaS. That was it.
Speaker 3That became the thing. How are you not receiving like millions of dollars?
Speaker 1Many people have asked me this and I was not savvy. In many ways, I'm not savvy today as to, but I also coined other things that did not take off, but that was, that was one that did.
Speaker 4When you look back to the fact that it took off, like what, how much distribution did they have? Like what worked? Like how does something like this stick so well? Because you've said like other things haven't stuck and people are always trying to come up with a term, name it, frame it, that type of thing. This stock, like why?
Speaker 1I think that there were enough companies that wanted to do, have a different way of offering the benefits of software to customers and they all wanted to be associated with the same thing. And it just, it was defined loosely enough, but specific enough that it could apply to any market. So I wasn't defining CRM versus SFA versus it. It was a broad, like this is a business model,
Speaker 4basically. How much was it that the companies themselves were excited to have something to frame themselves against versus investors, analysts, investment bankers, like what role did they play in all of this?
Speaker 1Yeah. So it's really interesting. What my boss now, Tien Tzuo, who's the founder of Zora, he was one of the very early employees at Salesforce. And I remember talking to him when he was there and I've asked him recently, because what was happening at the time where these companies were not taken seriously at all by the enterprise software companies. Right. One, their products were fairly new and they were definitely inferior from a feature function perspective. So the, you know, Siebel.
Speaker 3This is classic innovators dilemma stuff.
Speaker 1It's classic innovators dilemma. The big companies were like, there is no way a serious company is ever going to entrust their, you name your business function, put it in the cloud to these companies that, you know, it's laughable that they would ever, you know, choose this approach. So there was definitely a lot of confidence in the traditional model, that this was not going to take off. Now, of course, we know that it did, but it happened over time. And they're also, the financial community was not a lot of confidence in this model either, because like I said, we talked about already the cash cow that was the existing software industry and the idea of shifting that. So I had done some work back in the day and it held true every single type of software, every company. It would take you about three to four years to see the same amount of revenue from the same unit of software, perpetual subscription, so you were pushing off your revenue, basically three to four years in the future. So you had to have cash three to four years in the future in order to make it as a business. And the metrics
Speaker 2like LTV, they weren't, none of that stuff was invented. So this is the point with teens.
Speaker 1So I asked like, how did you convince, and they basically had to come up with those metrics at Salesforce in the early days. And they had to teach the street that this is the way that you look at this type of company. You look at customer lifetime value, you look at annual recurring revenue, you look at like all of those metrics had to be invented and they had to build a campaign around why this is the way it is. is the way that you should be evaluating these
Speaker 3companies. It's always interesting because I always, like David Skock had like a godfather. I think it happened simultaneously just like, you know, the airplane was invented. Well, it was also, I think it was also not
Speaker 2until 2009. It was like they'd already. Skock? Yeah, Skock was, I think a lot. 2009? No, he's much earlier than that. Was he much earlier
Speaker 4than that? That's like 10 years into the thing they're like, we figured it out.
Speaker 1I mean, the timing I'm talking about was like 1999-2000. That's earlier. Right, when I did that
Speaker 3first forecast. So, you've been at Zora for a while and just skipping ahead a little bit because we're still talking about pricing. We're still talking about usage versus seeds but talk a little bit about what you're doing right now because it's very relevant to what's happening with AI and you know, there's just so many different things that go into how you're thinking about building durable companies for the future including the book that you're working on.
Speaker 1So, most of my work since those early days of defining the market is helping companies realize the potential of that approach. And by companies, I mean both the providers as well as customers because I come with a deep belief in technology can save the world. It can also destroy it. Other podcasts probably, but we're going to focus that there's a lot of good that comes out of technology and you need to find an approach that works for customers and it works for providers. You can't have it be extractive. And so, my work has been helping companies figure out how to shift business models. And if you think of, you know, the transactional era was what the old software industry worked. The relational era was everything maybe up into this point. And now we have this agentic era. And there are different requirements on a business in each era. And so, my work at the Subscribed Institute is helping companies adapt their businesses or build adaptable businesses so that they can flex and grow according to how customers experience value.
Speaker 2How do you feel about minus 50% gross margins like Harvey in June? I mean, actually, it's a question, though, on the durable side of it. You have all of these businesses and we've talked about this. Yeah, we talked about it recently. Does gross margin matter? Yeah, does gross margin matter? Does math matter?
Speaker 1See, I, yes. I believe it does. Right now, it only matters if you're a certain size of business at a certain stage with a certain funding methodology associated with it.
Speaker 3It's almost like ugly if you have it too early right now. Well, everything is growth at the early stage, right? Right.
Speaker 1I'm spending a lot of time peeling, under the covers, and looking at some of these companies in part because I'm at a native SaaS company, let's call it, Zora, a company that's built around this relational era.
Speaker 3Do we like the word legacy? We don't like that word.
Speaker 1No, I'm not going to say legacy, but, you know, eyes wide open, right? If you're not thinking about how you need to change and adapt, you're going to be in trouble. So we're looking a lot at what are these AI-native companies doing differently, but the challenge when you're running a business at scale and you're optimizing the bottom line as well as the top line is you can't spend the way that a lot of these companies are spending on things that we've optimized, like eight different customer facing roles, as an example, in one of the companies that I looked at. All good ideas.
Speaker 3Eight different customer facing?
Speaker 1You have a forward-deployed engineer, you have a hackathon team, you have regular CSM, you have technical CSM, you have, I've lost track of how many people that I've got now, but yeah, solutions engineers. I mean, they're all of the above because they have to get customers to value really quickly on things that in tools, in many cases, that customers don't know how to use or how they're going to build things with them or what they do once they build with them, and they're growing, but at what point in time did that model. Well, how do you discern between
Speaker 2what an innovator's dilemma would be, which is like, you're looking at these like, should we do this? Is this what the world, the future's going to look like versus what's going to last and durable, especially at a company like Zora that's been around for 15, 20 years? Is it 20 years old at this point?
Speaker 12008, so almost.
Speaker 2Yeah, almost 20 years where there's definitely an innovator's dilemma piece of it, but there's also like, we can't do a lot of these things. How, like, is there a bucket? We're going to, okay, we're going to try this and put this in this bucket. We can't do these set of things. Like, internally, how does that work?
Speaker 1Yeah, so I think there's absolutely some rationalization. You can't do everything, but from a durable perspective, the way that I see it, it's almost back to first principles and organizing first principles around. Every part of your business optimized to help the customer see an outcome. So if you think everything from strategy to culture, monetization, processes, and then the technology is and then you can start to think about, are we organized in a way? You know, we've got org structures that have existed that do exist in most SaaS companies. Are those the right org structures? Do we really need the different roles that we have? Are they because we've had them or because they're actually serving that customer outcome? And so there's one track which is looking at how do you reorganize or organize around customer value? And another is, of course, how do you do that with a sustainable business at the same time? Because you've got to run both revenue and customer value simultaneously, obviously.
Speaker 4Can I go back to Salesforce for a second? Back then, Salesforce played this role, you said, on helping everybody make sense of this new sort of company. Do you see somebody playing that role in the same way they did now? Like, is that an open AI? Is that in place? Like, who is helping the market understand how to look at us and how to evaluate us and what made sense and what doesn't make sense? So I had
Speaker 1this conversation with a team. What a good question.
Speaker 3We're glad we waited for that.
Speaker 1I was like, team, so back in the day, Salesforce, who's doing this now? What do you think the metric is going to be? He's like, no, no, no. Back in the day, we had to convince everyone that this was a viable model. No one has to be convinced of that
Speaker 4today. Because the revenue
Speaker 1is real. The revenue is real, and the amount of, so I spent some time a couple weeks ago at J.P. Morgan headquarters and heard from and spoke with the head of AI investment there. And really all they care about is that in seven to ten years, some of their bets become the next Google or Amazon or, not all of them, but.
Speaker 3This is like their venture arm?
Speaker 1It's the head of AI platform investment. Okay. Or from their perspective,
Speaker 3the investments are, we're making these investments, to drive productivity growth for J.P. Morgan as a global conglomerate, so it's not worth we're not investing in per se, we're not equity shareholders, but we're spending money.
Speaker 1And as long as, in seven to ten years, they're free cash flow positive, they're good. So I said, how much do you understand, or are you looking at the commercial models that are developing right now? And he's like, we're not. We're kind of communicating directly with the founders, and they're telling us what they're expecting to see, and how they think their business is going to grow, and what margins they're going to see, and we're, this is my editorializing, we're taking them at face value, because we're seeing growth.
Speaker 2Because founders are never wrong. Founders are never wrong.
Speaker 1There's some great ideas out there, don't get me wrong.
Speaker 2Yeah, but isn't this part of the problem, though, is like that no one's really looking, I mean, there's this whole, and I still don't even understand it, concept of double-counting revenue, where, you know, Harvey counts this revenue, Anthropic also counts this revenue. I think that is, but that's not. It's unrelated, but it does feel like. No, I'm not saying it's
Speaker 3unrelated, I'm saying that those examples are not double-counting revenue at all, it's just a supply chain. I think whoever's advancing that meme is like, totally, seriously, is it double-counting revenue if somebody buys a Coca-Cola, and Coca-Cola paid the aluminum company a bottle to make? No, it's not double-counting, it's just how you buy. This is the whole
Speaker 2point, though, Mike, the concept, I hear it, and people say it, and I'm like, there's a lot of fancy words that are being used, and I don't understand it. Yeah, I don't understand it.
Speaker 3It's like, it's not circular at all, they buy compute from OpenAI, and they make stuff with it.
Speaker 2But to the point around forecasting and financial modeling, it does feel like, on the VC side, and everyone's is like, we have no idea, we're 150, 200x multiples on AI companies.
Speaker 3I mean, Instinct, I think, is $10 billion and 15 people, and it's, you know, it's like. 100,000 users. I use it. But it's only 100,000 people. I haven't paid anything yet. Thank you for the substance compute. It's at 5,000 to 7,000
Speaker 4per person per year, and compute, basically.
Speaker 3I read 3 to 5, but yeah, something like that.
Speaker 1So for those of us who've been around for a while, you can see that this is potentially a house of cards that is going to fall, or at least that's how it feels to me. And the same day that I was at this J.P. Morgan conference, my cousin was in D.C. to meet with members of Congress. Who's your cousin? He is an author, and he was one of three people that won a copyright lawsuit against. Oh, like OpenAI or the LLMs? Exactly. Oh, that's cool. So he has been doing a lot of speaking, because they ripped off his book and many others to build these models back in the early days, which, can you imagine spending 8 to 10 years of your life's work? I mean, we all remember when Google was buying all these books and scanning them
Speaker 3for exactly the same reason.
Speaker 1That's right. So they had to pay a penalty of about $3,000 per book, per author, is what people got in the end who were participating in the lawsuit.
Speaker 2So it's a class action lawsuit.
Speaker 1It's a class action lawsuit. But my cousin is texting me pictures from this AI is going to destroy a humanity conference while I'm at this J.P. Morgan. Literally, while I'm looking at this slide, seeing the amount of investment that is in this ecosystem right now, and I'm like, this isn't going to stop. Like, there is nothing stopping this. Certainly not the government as it exists today in the United States, because. What do you think?
Speaker 3Because to the point of durability, there's a spectrum, right? On one side is pure AI optimism. It's going to be great. Don't worry about it. We've got to fund the On the other side is pure skepticism. of this is a house of cards, it's all going to collapse. The middle ground, I guess, is it might collapse. Let's be responsible. What's your perspective on how we should think about?
Speaker 1So I will explain it the way I explained it to my mom, who was also in the middle of these text chains because this is my cousin. They're like coming to her. She's like, what do you think, Amy? I'm like, okay. There used to be people that were elevator operators and there used to be people that would change the traffic lights.
Speaker 4This is a great family chat, by the way. If you saw mine, is this a voiceover? Mine is crazy.
Speaker 1This is amazing. And I was like, yeah, yeah, yeah. I said, so those jobs don't exist anymore. No, you're right. And also there are guardrails in place to make sure that elevators don't plummet to the ground and that traffic lights don't change, you know, all turn green or all turn red.
Speaker 4You know something on that? When I moved here from Pakistan in 2007, we still have those people that press the elevator button. So it was very new for me. I was like, where's the person who's going to press the button? Especially if you're a germophobia. I'm not touching those buttons. I do everything myself.
Speaker 1I know. And I have a memory of being in an elevator with an elevator operator. And I have to say, I much prefer pushing the button myself rather than making small talk with someone in an elevator. But the point is that there was an innovation. We didn't hold back technology just because it was going to displace people in jobs that, presumably, didn't need people to be doing them. And there are guardrails in place to make them safe. So that's what I think will happen again.
Speaker 3As it relates to durability, are you worried?
Speaker 4Maybe durability doesn't matter.
Speaker 3Well, she's writing a book called Durable.
Speaker 1Oh, that's right. You missed that part.
Speaker 4That's the question. Durability means everything. So let me ask the question then. So I think the argument people would make right now is that this is not the moment in time to worry about durability. This is not the moment in time to worry about the equivalent of CAC versus LTV. This is a moment in time to just figure out how, like this thing is materializing, mutating. We need to make sense of it. Some things matter, some don't. So maybe the thing that matters is figuring out is tokens the right thing to measure? But the equivalent of CAC versus LTV, this is not the moment to worry about. And we're building these businesses. If we try to build them with an over-obsession with durability, with efficiency, with profitability, with things, crazy things like gross margins, that we're going to- Crazy. Crazy. It'll be too slow. And that we'll get overtaken. And we actually won't figure out what these businesses need to look like. And it's all of us who are from the previous generation that saw what played out in SaaS. We're looking at this and saying, you should do it the way we think you should do it, responsibly. But the core difference is that that movement didn't have the revenue acceleration this movement had. So while if we did what we're doing now, then it would be a house of cards that falls. We're doing it now with such revenue acceleration. It gives us the boundary walls to be a little bit wild and crazy.
Speaker 1So I think there's a difference between durability and rigidity. And what you're describing, what I heard in your question was experimentation. Isn't there room for experimentation and trying this and trying that? Absolutely. So my metaphor for durability is bamboo. And if you think of bamboo, it's one of the strongest construction materials that exists. I heard it's going to be on the cover of your book. It's been used for generations. Yes, if I have my way. There's a whole chapter on it.
Speaker 2But it's also incredibly flexible,
Speaker 1like if you've ever, and it's very easy to control, yet it also has a really expansive root system. And so there's something about building a business that has the flexibility to adapt, but also has a North Star that has a principle around, and in my particular case, or what I believe is that having that principle and understanding who your main character is, your customer and your customer, is more important than ever before, because now we have technology to accelerate whatever it is you're focused on, whatever it is your culture rewards, whatever it is your processes and your organization, your operating model is designed to produce, that result will be accelerated and amplified because of everything we have available to us.
Speaker 3This episode of Top Line is brought to you by Compound Growth Marketing. The way your customer's bias shifted before they ever reach a rep, they're asking an LLM who to trust, who to shortlist, and why. Decision-making is impacted by investment at the top of the funnel, and most B2B marketing teams underinvest because it's difficult to measure. That's the gap, and it's expensive. Compound Growth Marketing bridges this gap for many of the fastest-growing companies in B2B SaaS and cybersecurity. Their PDS, or Predictable Demand System, unites paid media, answer engine optimization, and GTM engineering into one discipline engineered to produce pipeline you can actually forecast, backed by over 100 million in media managed, no vanity metrics, no guesswork, if you'd like a free audit on funnel media, including AEO, reach out at compoundgrowthmarketing.com and mention Pavilion. AJ, have you reached out, and did you mention Pavilion? Did you say anything? I did. Last time
Speaker 2we had an ad read, I actually did reach out, and I have a call scheduled next week.
Speaker 3Oh, that's great.
Speaker 2So I truly do. And I mentioned Pavilion. I put the name in as well. Thank you very much. We will chat back in next week. Can you do that, Asit?
Speaker 3Reach out to them, mention Pavilion. I don't spend anything on marketing.
Speaker 4Okay, fair enough.
Speaker 3Well, anyway, folks. Please hire compoundgrowthmarketing.com so we don't become unhoused. And please mention Pavilion, but only one L. This episode of Top Line is brought to you by DevRev. Guys, tell me something. Tell me something. I'm excited about. I'm really, really jazzed. Jumping out of my seat. I am energized by computer. It's not the computer that you're thinking, but it's called computer by DevRev. If you work in GTM or sales, you know what it's like. You're running a dozen deals at once. It's on you to remember every detail. Who said what? What you promised? What needs to happen next? Things slip. A follow-up gets missed. A deal goes quiet. Computer by DevRev is an AI teammate that fixes that. Computer uses something called shared memory to hold the full context of every deal across every communication channel. Granola, Salesforce, the whole bit, even across the rest of your team. So nothing falls through the cracks and you always know where you stand. Your deals keep moving and you stay one step ahead. Take a look. Get computer.io forward slash top line. Acid. Yeah. React. I think computer is a classical thing. You need it. These people are cool. You know what I want to meet? The founders of DevRev. Because they're interesting. They're here. I hope so. I want to go demo a computer. Yeah. And also understand shared memory. Well, we've been talking about it for the last couple of months. Yeah. I'm definitely going to see it. I think the three of us need to go up there. We're going to go right now. This episode of Topline. is brought to you by Pavilion. We're coming off the best Pavilion GTM we've ever done. Ten years after Pavilion started as a few operators getting together for dinner, we now have a community of over 10,000 across 50 chapters and 15 countries. And at GTM 2026, we laid out what we think the next decade looks like. AI is making capability abundant. But judgment, relationships, and access are becoming more valuable, not less. The GTM 2026 is a community of over 50,000 people and over 100,000 members of a community of over 50,000 people. The GTM leaders who win from here will be AI-fluent, P&L-fluent, cross-functional, performance-oriented, and deeply connected. So we're building Pavilion around exactly that. We launched GTMSkills.com along with new programs like Leading with AI and GTM Engineering School. We introduced DreamFuel, bringing elite performance coaching to operators. And we're expanding access through Pavilion Scale, Pavilion Gold, and Pavilion Ventures, connecting members not just to ideas, but to the people, opportunities, capital, and ownership that can change your career. The next decade of GTM is going to look very different from the last one. If you're an operator and you want to be ready for it, come join us at Pavilion. Go to joinpavilion.com to learn more.
Speaker 4Is this a lesson for everyone other than OpenAI and Anthropic? Because, like, if you think about all of us, it feels very right to think like this. You know, I run a business. I would think closely to what you were saying. If you look at the application layer, in AI, you can see the sort of focus there as well. Like, Harvey has ICP. They want to serve that ICP. They want to do a certain amount of things. But then when you look at the companies that are driving most of this change, they're kind of the opposite of that. They're doing many things for many people. They're trying a bunch of wild. They're wild companies, right? Like, they're figuring out a new genome, and they've solved a math problem, and they're like, let's try this. It's a little bit wild and crazy, so it doesn't apply. Like, there's a bit of a.
Speaker 1Yeah, I don't know if those companies are going to be looking to read a book because they're probably not feeling the pain that everyone else that I'm likely talking to is. Which is, how do I distinguish between experimentation and leverage as well? Because if I've got something new, and there's something I'll be talking about at this event later on this week, or three different assumptions that you should rethink for your 2027 plan.
Speaker 3Will you tell us what the three different assumptions are?
Speaker 1Yeah, so one has to do with POCs, the amount of POCs. That are out there in the world. Everyone is doing a POC. And it used to be that hands-on product was a really good signal from a pipeline perspective, but everyone's doing POCs just to kick the tires, but actually has no mandate to buy. So that's one thing.
Speaker 2So be careful about your POCs. Be careful about your POCs. And paid POCs, obviously, you're talking about. Like, they're actually putting money in skin in the game.
Speaker 1No, there are a lot of unpaid POCs. Oh, really? The majority of POCs that I'm.
Speaker 2Do those companies count those as revenue in some way?
Speaker 1Nope, that's just a cost of sales. Yeah, they're not. It's like, can we build something that's productive? So that when the customer says go, we flip the switch. But I'm talking to a lot of companies that very small percentages actually pan out. So they're doing a lot of free production ready builds. And the customer says, oh, I actually have another one with another vendor, or I've just decided to build it myself or not.
Speaker 2Because there is like the pre-seed seed world right now. And those BOCs that existing, I would assume that if I'm a founder of that company, and I need to get into the investment committee of some big fund, Andreessen or Sequoia, and I'm putting that logo up. I mean, I don't care if it's paid or not, I'm going to be putting it up there. So I just wonder, I know we've gotten in trouble like 2021 and 22, when everyone was raising in the zero interest era, like traction sort of mattered, but the idea mattered. Today, it feels. It matters. Idea, obviously, in the AI world, but traction, especially if you're doing something that's like a verticalized slice of it, definitely still matters as well. So those companies are probably-
Speaker 3It matters a lot, which is why I'm sure it's being faked. I know. It's a fake it before you make it. Yeah, then I think it's not, you know, it might not be ARR, it might not be contract.
Speaker 2I haven't heard experimental revenue in a while, thankfully, but that was a thing like two years ago.
Speaker 1A couple of weeks ago. Brought up, so.
Speaker 4Great. I have a question for you. I think it's still a thing. Would you rather 200 million in experimental revenue versus like 30 million of whatever other-
Speaker 3But will this define experimental revenue? Does it- You define it. What do you think?
Speaker 4You're the one asking the question. What do you think experimental revenue is? People have been subjective about it. What I would think
Speaker 3is the following, to the point of POCs. I would think if 300, let's say $100,000 for a three-month POC, and you call that- You call that $100,000, you call that $400,000 ARR-E.
Speaker 4That should be illegal. But I think the other version of experimental revenue is that your customer is experimenting. JP Morgan has, you know, $100 million budget to buy AI, and they've told a bunch of department heads, go spend this money, figure it out. We don't know what the use case is, and so they don't know what they're buying, but they're ready to spend to try to wrestle with this. With this, scratch at it, and maybe some use case pops out. Every company that they're going to give the money to is going to be like, I have revenue. They're going to take it. But this is downstream of a massive experiment, and the customer is not buying this with the sort of mindset when they buy a CRM.
Speaker 3This is why I find SaaS pricing useful, because I don't- Listen, I teach that the customer decides if it's a subscription, not the contract. But nevertheless, if you get a 12-month contract in place, at least you know there's some commitment. I would call that ARR. I don't think that-
Speaker 4But even if they sign, in this case, a 12-month commit, right? The mindset of the customer is, I think, the biggest factor here, where the customer's like, I'll sign a one-year contract with you. Let's run a bunch of experiments, see what your building made sense of.
Speaker 3That I would not consider ARR. I would consider that true ARR. I can get into the mindset of the customer.
Speaker 4But remember, like, in this, you've got usage base. So the customer can sign a one-year contract, and how much they end up spending with you is a big, big difference.
Speaker 1It's the delta between the commit and the burn. Yeah. Is what you're getting. Yeah, yeah. It could be huge.
Speaker 2Amy's point, though, earlier, which is, like, it used to be hands-on product was a great sign of adoption and buying signal. I think in your question, though, I'd probably take the $30 million in durable revenue.
Speaker 4That's crazy. If this is the form of experimental revenue that I just mentioned.
Speaker 3Yeah. Well, so
Speaker 4he said $200 million in revenue
Speaker 3versus $30 million in revenue. $200 million in experimental. But we haven't defined- But his definition of experimental is, they don't really mean it. So
Speaker 1this is the key. I'm not sure. This is the key in the talk, too. Is how do you define experimental from leverage? So activity versus leverage, where they're actually using it. Is there a proxy? Because we, one of the great advantages of these systems is you can actually see what customers are doing. So if you can do that, is there a way that you as a business can both predict-
Speaker 3Right. So they're using Zora sometimes to buy this stuff, right? Or yes? Yeah.
Speaker 1So what Zora would allow you to do, like a prepaid drawdown model, as an example. And so let's say you have a prepaid drawdown model. And you're observing customer behavior. How do you differentiate between experimentation and actual leverage? And I'm seeing activity and leverage as the two terms that are being used in the industry today. And so what are the proxies that you can determine in system that will help you understand what your next play should be? Very interesting. Because it changes your behavior for how you actually help a customer and what different resources you fire off. It changes potentially your sales compensation model. It changes what you're motivating sales to do. It changes a whole host of things.
Speaker 2So what are the signals in the prepaid drawdown that are giving you that idea that you have leverage?
Speaker 1Yeah. So I've been looking at our own, so Zora's own software. And I've developed three different categories that I'm fooling around with filters right now. It's experimental. But one of them is scope and workflows. So how many APIs are they connecting to? How many workflows are they running at any given time? The other has to do with- Mm-hmm. Yeah. Yeah. Yeah. Yeah. Yeah. Basically, how much are they trusting the system? Are they actually taking outputs and changing them? Are they taking them and actually continuing on with their work? And there are a few other measures there. And then the third is context. How much of their own data are they uploading and putting into the system? And so if you look at those three categories, and there's some subcategories there, I've developed a leverage index and a score. And what's really important is you can have two customers with the exact same usage but very different leverage. And so what kind of workshops you'd offer them or what training? And what kind of training you would send to them are very, very different. To the point, because if
Speaker 3they're low on the leverage index, then that revenue is less durable.
Speaker 1That revenue is less durable. That looks a lot more like lots of experimentation. The other thing with usage is that over time, you would expect it with some of these tools to decrease as people optimize. So when those experiments end, I mean, think about our own use of LLMs and how smart we've gotten about which model we need to use when. You would expect similar things to start to happen with some of these tools as people-
Speaker 3I never- I always defaulted to the latest model until recently. And now I'm like, when I'm using ChatGPT, I'm like, Sol's fine. I don't need to use Astra.
Speaker 1Right. Yeah, for certain- I mean, I'm going to write an email. I absolutely do not need to use Astra.
Speaker 4I need that PhD to give me a coffee. Okay, fine. Yeah, exactly. I think there's this story of a company called Rogo. I think it's from New York. They are an AI financial services agente platform. And Sequoia is one of the investors. And they present- They presented this plan to the board of how much they were going to grow and how many salespeople they needed to hire. And Pat Grady from Sequoia says, listen, this is an aggressive plan. But if you believe that in the next 24 months, every financial institution is going to make an AI purchase of some sort, then they're going to buy something. It's imperative that you are in each and every one of those doors. So make this more aggressive. Hire more people. And I think that's a little bit of what the investor mindset is. That's everyone's going to make a purchase. Everyone's going to have a vendor of some sort. What you do for this company today is going to be very different than what you're going to do two years, three years down the line as the technology gets better and this world materializes a bit more. But you need to get in. You need to be the one that gets in now and then holds on to this customer and navigates with them. And I think that's a lot of the game that's been played. So they don't care about the efficiency. They don't care about the durability. They don't care about the things that we traditionally care about. It's this game that they're playing.
Speaker 2What's interesting. Though, is that you hear those stories. I rarely ever hear the like after. Like, what did that actually happen? Did they hit the plan? Did they overhire? Or I mean, definitely.
Speaker 3LaGuardia and Harvey are hitting the plan.
Speaker 4Well, they're flying.
Speaker 1Yeah. That's the challenge of writing this book now is by the time I hit data, we won't know. Right. These will not have run their course. But I'm I have a hypothesis just based on what I've seen in history is that the behavior and what we're seeing now will end at some point in time. And how we don't know. And so the question is, what remains? And if you're trying to build a business that you would like to outlast that, what are the way what the principles do you think people are trying to outlast it?
Speaker 4I think about this a lot. I'm like, what's what are they thinking of? And I think a lot of people at the early days are like, let's give it a shot.
Speaker 3Secure the bag.
Speaker 4This is a casino. Let's give it a shot. Let's try to build something great. We'll make money along the way. There's secondaries, etc. So it's not like I'm playing this. Binary founder journey of like, I have to build something durable. And that's the only way or something that survives a lasting company. That's not the only way for me to win. There's a world in which I try something has a nice start, hits a wall and then just goes to hell. But I made some money along the way and it was a fun adventure. And I was in the field. If that's your objective, then that's a good strategy
Speaker 2unless you get
Speaker 4divorced.
Speaker 2But Asad has said something to me. I'm switching gears maybe slightly because a lot of SaaS era companies are thinking more about services. It just kind of makes sense. AI tech enabled services, AI native services, however you want to define it. Finally, the services in SaaS like me. Well, services is actually a thing now. And so Asad always says this at the beginning of the year as a true service company that his revenue is zero. That would freak me out. So like I'm a founder CEO and now I have to start to think differently about my own company. And just how the metrics are changing and transforming for 2027. And I am like very, very heavily focused on finding the right ICP, right product market fit for our services and what this looks like. But how do you see and actually curious of Zorro? is thinking about this in this way as well. Do you see a lot of SaaS-era companies that are like, okay, we have to find it. It seems like the puck's moving towards more of a services, they just outcomes. They just want everything taken off their plate. Like for a sales comp, easy. We'll just do it for you. Zora, we take care of everything. You don't have to think about any of the payments. Is there something like that that you're seeing for other SaaS companies that we're moving towards a services era?
Speaker 1Yeah, I mean, absolutely. If you have a really keen understanding of your ICP and what their ultimate objective is, it's probably not owned software. It's probably not subscribed to software. It's probably get some sort of a business done, process done more efficiently, effectively, more compliant, whatever your game might be. So Zora in particular is selling to finance organizations. And, you know, most every function is going to, in the future, have some sort of an AI teammate that's part of their organization. And so if you're a SaaS company, you might start to think of, am I in the business? Of software or I'm in the, am I in the business of AI teammates in, in whatever function it is that I serve. So interesting
Speaker 2because then I saw like rippling and I think workday was headed towards this where like they're going to make agents like FTEs and that feels like a weird pricing switch on their, their side as well. But, um, does Zora have AI teammates today? Like, do you, do you, does Zora think about this in that way?
Speaker 1We would think about it in this way, but do we practically have that today? I don't know. Not yet, I guess is the answer to it. And I think that a lot of, you know, who you serve and their progression along the maturity curve, but I think McKinsey has the ask, act, orchestrate sort of the, the, the progression. And I think you're in the finance role and like you're signing your name on things that if they're wrong, you could go to jail. You might not be ready to hand over an end to end process. Yet, but in the future, why, why not?
Speaker 4What are your thoughts on, we, we, we have this conversation around tokens happening in the industry right now, the cost of tokens. And so you got the frontier intelligence and you got companies like open AI that have a range of models and they want the most intelligent model in each category at the best price possible. And there's a lot of pricing pressure that's coming from open source, which is making it cheaper for some companies to theoretically be able to use this. AI themselves. But then you think of like just the idea of a token. Is that the right way to measure it? Because one model might use a million tokens to do one task with your 90% accuracy. And another model might use a thousand tokens to give you the same outcome with the same accuracy. And they'll have a different cost. Maybe the thousand token model is more expensive. So it's more expensive, but gets you there faster, right? And the cheaper model takes more time, takes, so there's this thing happening of like, how do we actually measure intelligence right now? Do you think this is the first building block that we have to master before we get to everything else? Like, we really need to figure this out. And do you think this is the right measure?
Speaker 1Gosh, that's my, that was a very scientific question.
Speaker 3My brain that
Speaker 1got up really early this morning. What do you think about tokens? So, I mean, tokens as a layer of abstraction is a really good one and it's works really well to get you closer. You are to infrastructure. When you're talking about applications in a lot of what we talk about in the SAS world, credit models, which take those tokens and level them up and make them a little bit closer to what a business user, like the task that they're trying to get done, makes, makes a little bit more sense. I think what I'm seeing right now in the conversations that I am having with companies who are trying to figure out tokens versus credits, it's like, figure out first of all, if you need those. And if you're really close to infrastructure, you're if you're selling to a business user, some sort of functionality, getting involved in credits and tokens might be too confusing for your buyer. They're like, I don't, I cannot grok what I'm buying, how many I need. I can't predict it. It's confusing, et cetera, et cetera. So not everyone's going to need a token or even a credit model. I think having something that's as simple as possible initially, as opposed to trying to find the thing that is like laser, the best model and a spreadsheet to align price or value. That first one is probably better because at the end of the day, you're selling likely to a human. And if you're selling to a business person, that human's got to explain it to somebody else. So it's like a game of telephone. So having, I call it narrative coherence in your model is incredibly important. If you're selling to a sophisticated CIO, what you're describing, like that's their jam.
Speaker 4What about like outcomes? Like how far along to providing real outcomes and pricing against those do you think this market is going to get to? Because like initially when all this took off, we're like, now tech is going to charge for the outcome. We'll charge for closing a sale for you or helping you actually- Making the hire or whatever. Making the hire, et cetera. But you see right now, the form of outcome-based pricing we have is very far from the actual outcome. You know, this is not really that. Like how close to it do you think we get?
Speaker 1It's really hard to have a very, very close to outcomes pricing. And by the way, outcomes pricing has been around for a very long time. When I first got into pricing very long time ago, there were companies that did it. These were very highly negotiated. You had to have a ton of trust between the customer and the provider because let's say you're pricing based on revenue. So many variables that contribute to it. It's not always easy to track the best outcome metric in the system. And back in those days, you really couldn't. So you'd have to like sit down and they'd show you their revenue. Like we'd figure out what you would pay. Now, obviously now we can- Orchestrate and put these in the system more. But it, so for example, one area where you see more outcome pricing is in customer support and support resolutions. And so it's pretty well accepted that an outcome is ticket is resolved. But the way that they determine that is if the interaction stops and like two days go on, they consider that closed. But that could be because the customer gave up.
Speaker 4It could be for a number of things. But they have to, yeah,
Speaker 3it's closed. Life is too short. Someone complains resolved. I see it's resolved.
Speaker 1Even in a case where it's accepted, it's imperfect, right? And there also are things other than the tool that contribute to those tickets closing that you might, well, our product is just more reliable or et cetera, et cetera. And so they're open to debate. I would expect to see more of it though. Thinking of the conversation that we had a few clicks ago on activity versus leverage, if you're just tracking usage and, you're in the same vulnerable position you were in with seats. That's how you were doing it before. So there is got to be some percentage associated with outcomes. It probably won't be 80%. Maybe it's a hybrid model and your outcome piece is much smaller percent, but you have to figure out what that is and has to be trackable and transparent.
Speaker 4At some point, these buyers are going to move away from like, right now, I think everybody buys with a level of excitement and fear and then they're going to buy with a level of excitement and fear. And then they're going to buy with a level of excitement and fear. And they're willing to be a little bit loose with their money. The economy is good. Everybody's making money. Money is flowing. There's new intelligence. Let's try to be on the, be aggressive here.
Speaker 3At some point, people are going to rationalize. And your point is, I'm telling you what the
Speaker 4other argument is.
Speaker 3You're like, everybody's making money.
Speaker 1So the argument is get as much
Speaker 3market share as you can so that as things consolidate, you can be there making your case for why the way you did it was the right way.
Speaker 2Well, and Austin and I have differing a point of view on this where he says everybody's making money. I, I don't, I think the majority of us do not feel like that that's the case. We're actually all looking at our bin where our businesses and be like, and our health benefits are going up 15 to 30% this year. This year is not better than 23 and 24. Well, not yeah. Better than 23 for sure. I don't recruiting businesses are doing great. Recruiters. I don't, it's, it's really hard to say how we get through the end of the year. I feel okay. There's some things that are working really, really well. And some things that feel fundamentally broken in the world.
Speaker 3Amy, we're almost at the end of our time together. I want to get the last. So one assumption is POCs. Be careful about POCs for your talk. What are the other, you said three things for 2027 planning.
Speaker 4We'd walk through one.
Speaker 1The second one is activity versus leverage.
Speaker 3Okay. We covered that. What's the third one?
Speaker 1You have to come to my talk to find out. I'll be there. I'm holding you back because.
Speaker 3Before we go to the point of your family group chat, my friends always get mad at me when I take it to a different, weirder place, but what's your PD about the birth rate?
Speaker 1What's my, that's a big thing for me.
Speaker 3What's your P doom, what's your percentage that AI is going to kill us all? Like, are you, are you a tech optimist or tech pessimist? You know, is this a good thing or a bad thing? Obviously there's qualifications, but generally speaking, you know, how do you feel about the next 10 years?
Speaker 1I'm a tech optimist. Today was the pavilion. I'm a tech optimist. Today was the Pavilion Women's Conference. And every single woman that got on stage mentioned her family. And so I will say I'm a mother of three children. Two of them are in college. I've met one of them at the Pavilion Gold Retreat. Yep. And I am, I've got them all over AI because I think that they absolutely have to have knowledge of how to use these tools for their advantage. And they also have to have grit and adaptability because who knows what they'll be doing and how they're going to be doing it. but I'm also teaching them to look people in the eyes and be able to carry a conversation because I feel like that is going to become more and more important. And it's always been important.
Speaker 2No, 100%. 100%. I do wonder on that last point, I have three daughters, 14 and twins that are 11. How young would you start in the AI? Because it feels like there's just a weird, the anxious generation and phones and social.
Speaker 1Yeah, I'm taking the cues from my son's school, which I've chosen well. It has a long history of educating leaders and they do not have the kids using the tools all the way through high school because they're teaching them to learn. And so it wasn't until now, I mean, there's all the way through high school
Speaker 4and didn't want to write his college essays with AI. But like for you being so progressive with technology, doesn't that sound like, I remember when I was in high school, they didn't let us use calculators. And I used to find that really frustrating because I was like, but in my entire life, I will always use the calculator. It's a good point. And spell check.
Speaker 1And like, there's something about learning how to think and building the confidence that something happens with this tool. I know the process behind it. I know what good, I know how to diagram sentences because my eighth grade teacher was the worst at the time. But my eighth grade teacher, English teacher, made us diagram sentences. And so I, AI slop drives me incredibly now because I know how to structure a sentence. I know what good writing looks like. I know what-
Speaker 3Those used to be stylistic choices, the way AI writes. And now it's, and now nobody knows that it's a stylistic choice and that, you know, you need a subject and an object and a noun and a verb.
Speaker 1Right, so you develop your own filter. So, you know, there were all these schools, the alpha schools and there are articles about them. And maybe, maybe those, maybe I'll, you know, in the future find out I did the wrong thing. Maybe both ends of the spectrum work really well, right?
Speaker 4Like the alpha school on one side, which is like the very progressive implementation of AI and like education. And then like the complete opposite end also sounds really cool. Where like you actually have none of that and you live life a little bit more.
Speaker 3On a farm, growing your own vegetables. Amy, tell the audience where they can buy your book. You know, let's sell some stuff for you. That'd be
Speaker 1great.
Speaker 3Before we go.
Speaker 1Amy Connery, look me up. The book is not printed yet.
Speaker 3But it's called Durable.
Speaker 1It's called Durable. We're working on the subtitle. Yeah, we're working on the subtitle. I've got to set up some sort of pre-purchase. Bamboo will be in there. There'll be some bamboo imagery.
Speaker 3Currently SVP of Undisclosed at Zora.
Speaker 1That's right. Working on that.
Speaker 3Awesome. Thanks for being our guest on Topline. It was great. And thanks for being a loyal member of Pavilion Gold.
Speaker 5Of course. If you are looking for more Topline, check out the Topline newsletter at toplinemedia.substack.com. That's toplinemedia.substack.com. Or if you're a video person, check us out on YouTube. Topline-media is what you want there. Have a delightful day, everybody. Topline-media.substack.com.
PodcastTranscript AI