Inside Gong: How teams work with design partners, their pod structure, autonomy, trust, and more | Eilon Reshef (co-founder and CPO)
56:42 Share

I want to start with talking about your pod model, which is a really unique way of building and organizing your product teams. That was probably 2016. We're trying to figure out how does the operating model look like for a product and engineering. The first bunch of people we had was essentially a pod. It was one product manager, a user experience designer, back-end engineers, a couple of front-end engineers. But at some stage, we're starting to scale. We were kind of contemplating, do we go like the traditional, the old school of front-end engineers, back-end engineers? And then we said, let's try to replicate what we have.

Let's talk about how these pods work with design partners. From what I've heard, it's very unlike how any other company works. We just took the pod concept to an extreme where every pod is working with sometimes a dozen design partners, sometimes two dozen design partners. This feels like a cheat code of how to build new product lines. What are the percentage of success rate you have with new product? I would say very close to 100% of the features we build end up being used by a significant number of people.

Does it feel crazy for companies not to operate this way? I wouldn't go back. I hate terms such as risk because that's a very ambiguous term, but just the risk of building something you're not going to know if it's going to get used. So when I ask people like Gong what to ask you, the most often term that came up is autonomy and trust. It's a very selfish thing. It's a very personal thing. I just think...

Today, my guest is Elon Reshef. Elon is co-founder and chief product officer at Gong. He was also the longtime chief technology officer at Gong. As I share at the top of our conversation, it feels like basically every company that has a sales team uses Gong. And it's really rare to build a product that is so ubiquitous and so loved across the tech ecosystem.

In our conversation, Elon shares some of the secrets of what makes Gong so consistently successful, including how their product teams work with 6-12 design partners on every new product and feature that they invest in, how he creates a culture of autonomy and trust,

Why and also how he optimizes for making decisions quickly, even large one-way door decisions, what he and his team have learned about building AI-based products since they've been building AI-based products longer than most other companies, and so much more. If you're building a B2B SaaS company or product, you will learn a lot from this conversation.

If you enjoy this podcast, don't forget to subscribe and follow it in your favorite podcasting app or YouTube. It's the best way to avoid missing future episodes, and it helps the podcast tremendously. With that, I bring you Elon Reshef.

This episode is brought to you by WorkOS. If you're building a SaaS app, at some point your customers will start asking for enterprise features like SAML authentication and skim provisioning. That's where WorkOS comes in, making it fast and painless to add enterprise features to your app. Their APIs are easy to understand so that you can ship quickly and get back to building other features. Today,

Hundreds of companies are already powered by WorkOS, including ones you probably know, like Vercel, Webflow, and Loom. WorkOS also recently acquired Warrant, the fine-grain authorization service.

Warrant's product is based on a groundbreaking authorization system called Zanzibar, which was originally designed for Google to power Google Docs and YouTube. This enables fast authorization checks at enormous scale while maintaining a flexible model that can be adapted to even the most complex use cases. If you're currently looking to build role-based access control or other enterprise features like single sign-on, skim, or user management,

you should consider WorkOS. It's a drop-in replacement for Auth0 and supports up to 1 million monthly active users for free. Check it out at workos.com to learn more. That's workos.com.

Hey, it's Lenny. If you want to boost your clarity and confidence, I want to recommend a podcast called Think Fast, Talk Smart. One of the most essential ingredients to success in business and in life is effective communication. Every Tuesday, host and Stanford Graduate School of Business lecturer Matt Abrahams sits down with experts to discuss the best tips and techniques that enhance your professional development. Hone your small talk, influence, presentation skills, and so much more on Think Fast, Talk Smart.

So what are you waiting for? Listen every Tuesday, wherever you get your podcasts and find additional content to level up your communication at FasterSmarter.io. Alon, thank you so much for being here. Welcome to the podcast. Thank you for having me. So it feels like on this podcast, every guest that I've had mentions Gong as a product they use. It feels like it's just like in the ether of tech companies these days.

Also, I've heard so many times how unique it is that you all operate, how you all build your product teams and operate, which is I especially love hearing on this podcast. They're just different ways of operating. So I'm really excited to dig into this, to hear about the journey and things you've learned along the way of building Gong and essentially building something that is so ubiquitous and so loved, which is very rare.

I want to start with talking about your pod model, which is a really unique way of building and organizing your product teams. And in particular, how you work with design partners. Let's start with the pod model. Can you just talk about what this pod model is and how you organize your product teams?

Sure. And when we started the pod model, that was probably 2016 before it became popular. I think that might have been even before the Marty Kagan set of books. I don't remember exactly. But at some stage, we're starting to scale. I mean, scaling is maybe from 50 people to 60 people or whatever in the whole company or whatever. And we're trying to figure out how does the operating model look like for a product and engineering model.

And the first bunch of people we had was essentially a pod. It was one product manager, yours truly, a user experience designer, a couple of maybe backend engineers, a couple of frontend engineers and whatnot. And we were kind of contemplating, do we go like the traditional, the old school of like, you know, frontend engineers, backend engineers or however it is. And then we said, let's try to replicate what we have.

So what we essentially did is really kind of replicated that. So up until now, we have this pod structure, product manager, UX,

fractional writing, fractional analyst, and then a team leader from an engineering standpoint, five to I'd say seven engineers. They get an agenda to think like we launched a forecast product. That was a pod working on that. And then they get to be pretty autonomous in identifying how to solve the problems and also working with enough customers so I can go to sleep knowing that they're not like hallucinating is the term that everybody uses now in different contexts, but going in a very reasonable direction.

Okay, so I want to talk about the autonomy piece. That's a really important point. But before we get there,

Let's talk about how these pods work with design partners. From what I've heard, it's very unlike how any other company works. And I think it's something a lot of people can learn from. So talk about kind of the scale of how these pods work with design partners. So I think we just took the pod concept to an extreme where every pod is working with sometimes a dozen design partners, sometimes two dozen design partners, maybe sometimes five if it's a very niche or fringe capability. And they work with them hand in hand.

So an interesting story, the same forecast product I just mentioned, the product manager comes to me one day and he's like, hey, I just kind of played with a design partner on the product. And I kind of know what's going on and I know it's not built yet. So I asked the product manager, but we don't have that working. So he said, no, no, we don't. But I showed him the sort of the half-built stuff. I asked him to hit save. He hit save and got an error message.

And I told him, let's meet again in a week and that save button was going to work. So it's very extreme in terms of working hand in hand with a customer. Customers appreciate it. I later got feedback that they appreciate how kind of the thing was going, making progress according to their feedback. Not what they said, but kind of interpreting what they said, digesting it and building something that makes sense. So every bot has this set of design partners.

So these pods essentially across functional product teams, each one has, do you organize it around like an outcome? How do you describe what each pod is responsible for? Is it like move this metric or build this product or something else?

We tend to be less metric-driven maybe than your average, especially B2C, but even like other maybe B2B companies. Usually it's more around some sort of a job to be done. In our case, it could be sales engagement. How do you kind of prospect? Or conversation intelligence. How do you create a summary? How do you make it easier for people to remove drudgery and consume information fast? Once you get this agenda, you kind of pretty much have a lot of kind of...

You mentioned autonomy. A lot of control over how you progress. Ideally, design partners guide you, not me. Awesome. Okay, so it's like, here's the outcome we want this pod to achieve. They have autonomy to work with design partners to design this new product.

So maybe let's stay on this example of this forecasting tool. Can you just briefly describe what this tool, what it gives you, what it does? Yeah, it's a product we launched a couple of years ago, and it helps organizations forecast where they land in terms of the sales organization. So every sales organization, B2B sales organization, has a bottoms-up forecast process. People submit numbers, people kind of override that. AI helps you kind of, in our case, AI helps you predict the right number. Usually there's an analytics component on top of it that kind of helps you

assess it at scale. So that's the product itself. Awesome. So you created this pod here, build this forecast product, make it successful. How do they find these design partners? Is it like reaching the existing customer base and figure out who would be most interested in this?

Usually it's existing customers. Very, very rarely it would be non-existing customers, but customers with some stage expressed interest in this capability. Not to over give a plug to Gong, but of course we can listen, all of our conversations are recorded, so I can always look up our conversation database and which customers express a need for X, Y, and D. You can very easily reach out to them.

One of the maybe unique things we've done at some stage just became, I think we have like 25 pods right now, maybe 30, it depends what you call it, pod. But it's a lot of effort, like this whole management. At some stage, I borrowed an idea that comes from talent acquisition. And in talent acquisition recruiting, there is a person called a recruiting coordinator.

If you do that scale, which we had done over the years, and that person, all they do is just set up the meetings for the recruiter, who then sets up the meeting for the hiring manager. So in our product team, there's one person who's basically a research coordinator, and she's responsible for reaching out. She basically talks with the PMs like, what's your target market, ICP, whatever you want to call that? Give me some, what do you want to learn from them? She reaches us. We have a micro CRM for that.

and then sets up the meeting. And the PM comes in, they have already a meeting in their calendar. And these are the people at these companies that are going to be their design partners.

Exactly. So I might say, hey, what I want to speak is with a head of RevOps at, I don't know, mid-sized companies or, I don't know, IC seller at an enterprise company. And then she can kind of, of course, sift through our customer base, slice and dice it, run a micro email campaign and get those going to come in. I love that detail because, as you said, coordinating 12 companies and people at these companies and timing is really stressful and complicated and it's like...

the PM's life up. So that's really helpful. It's interesting. There are some companies where product teams aren't even allowed to talk to customers. Salespeople are like, no, don't mess with these people. Customer success, like, no, we got this. You're like the complete opposite. Each pod is working directly with, say, a dozen customers helping build a new product for them. Exactly. And in the early days, we didn't even like...

tightly coordinated with customer success. Nowadays, we do it much better because there's always going to be these customers like frustrated about something or in a negotiation about something that's probably not the right time to ask them about to be a design partner. So we kind of double check, but it's not like a process where we have to get signed off by three customer success managers to talk with a customer. That makes sense. Obviously, yeah, you don't want to surprise people and mess up relationships. So is there any structure to the design partner process or is it just teams have these people at

available and they talk with them when they want? Or is there more structure to how to effectively build a product with design partners? It depends on the context. So when you build a new product, there's a more linear path, right? Ideally, you want it to be launched on time. Ideally, you want to measure some progress. So usually what we've done in this case is some sort of a weekly meeting where we kind of show them progress and sometimes biweekly, depending on our own cadence.

And some other capabilities that we built might be more, I would say, freeform. Maybe it's an enhancement, maybe it's a tweak, maybe it's an extension of the... Let me give you an example, right? One of the things we do right now is we let customers ask a question about an account. So coming to Gong, like, what's new with Cisco if Cisco is a customer? And

At some stage, you do it, we're doing it in all languages, right? So you want to recruit a specific design partner, a set of design partners who are non-English speakers. So it doesn't have like a strict timeline. You want to get enough of them so you see the thing works, gets your reasonable quality results. You're not going to get to all languages on the one hand, but at the same time, you don't have like just Spanish. So that might be less structures. Maybe it's a couple of meetings with each one. And then when you move on, you launch the feature and you move to the next kind of quote unquote project or feature.

So one thing that people might be thinking as they hear this is how do you, all these customers are telling you, here's what I need to be to use this forecasting tool, for example. And as a PM, it's always this balance of doing what customers ask you to do versus like, oh, we have this vision and here's how we keep it simple. What guidance do you give your teams for what to do with this feedback essentially?

Yeah, I think this is kind of core skill that I expect PMs to have around this. This is exactly your job, right? Try to figure out what request is like must have versus not must have. We typically, they ask the customer, you know, what do you have right now? How happy are you between zero and 10 or whatever? So if you're at the six, we want to get you to an eight or a nine. So that's maybe a high level principle. But at the same time, I expect them to say, hey, this is a unique, I didn't hear it from anybody else. Maybe I'm going to proactively reach out to more customers, but that might be a one customer thing.

And we still do one customer things in a different context, right? So if we have, I don't know, a seven or eight figure deal that they have like one customization that they really, really need and we know that they can't work without it. I mean, like every enterprise facing company, we're going to do that. But from a design partner perspective, it's the opposite. It's more like let's try to build something that works across our customer base versus for a specific customer. Which is why a dozen is probably smarter than one or two or three.

Yeah, I think at some stage, I guess you get like seven or eight or nine at some stage. Based on my experience, the request starts to converge. There might be one outlier, but generally you're going to hear the same things. I imagine this approach is rooted in how you all started. We worked on a post back in the day on how you all got your first 10 customers. And I remember the story was you got, I think, 12 design partners when you first designed Gong. And then you told them, we're going to start charging now. And 11 out of 12 were like, we will buy this now. Please charge us. And we love it.

Exactly, exactly. So in a way, it's replicating this, but it was successful at the time. It was like maybe 80% intentional at the time. And at some point, you take the stuff that works and you make it 100% intentional. What are the percentage of success rate you have with new products? Because you would think this approach is the best way to consistently build products people will end up buying and using. Is it like 100% of the time you end up building things that people will buy and use? Is it something below that? What do you find?

I think it does increase significantly the utility of the products. I would say very close to 100% of the features we build end up being used by a significant number of people. We don't charge for all of them. For most of them, we don't charge, which doesn't mean other things couldn't happen. Maybe people use it, but the value is not huge. So it's like, yeah, design partner likes it, but it ends up being applicable to a smaller fragment or segment of our customer base than what we had hoped.

It may be they're not as willing to pay for it, although that's a little bit of a different process, like real product launch. It could be the quality is not good enough because we're kind of focusing on, you know, is it providing value? Is it understandable versus like, you know, did you find a bug when you used, I don't know, Safari or this kind of computer? Because we're not building the design partner program to kind of solve for this. Maybe we should, but we are not doing it right now. But generally speaking, I would say yes.

But more than 95% of the capabilities we build are being used in a very significant way. Which I think is probably higher than most companies. And this feels like a cheat code of how to build new product lines, expand product expansion, TAM expansion, ways to add new ways to charge your existing customers. It feels like a cheat code, basically. Just tell us what you need, we'll work with you and build it, and it'll sell to you. It'll be great.

Yeah, respectfully, you know, yours truly, I do believe customers know much better than what they need. And then myself or my colleagues in the executive team, whoever else it's,

You talk to a customer and they kind of describe the pain. They might not know how to build it or what's the right way to implement it, but the pain should be there. Coming back to something else you mentioned, this word autonomy. So when I ask people at Gong what to ask you and what stands out about you to them as a product leader, the most often term that came up is autonomy and trust. How much autonomy you give teams, how much you trust teams to do the right thing.

Can you talk about that way of working, where that came from, and why you think that is the way to operate? It's a very selfish thing. It's a very personal thing. So I think even beyond trust, it's just, for me, it's selfish. I'll tell you why. I just think you get more from everybody if you kind of let them be themselves and do things in the way that they believe is the right way, of course, within limits, right? They're not going to develop, I don't know what it is, software in a different business. But the

The story I always like to tell is I was when my son was in primary school, which was a while back. One of the parents told me and we had this picnic where all the parents and the kids were going to meet. And usually there's like a list of ingredients that people need to bring in, you know, to bring on a battle of water, whatever the thing is.

And what's usually happened, there's even like people are joking about it, is people run to the list because it's usually like a physical list and then, or make night, probably now it's already in Google Sheets always. And people run to just like mark the item that is as easy to get as possible, like kind of battle of order and then I'm done. And then you always get like the lowest common denominator because everybody brings the sort of, I don't want to say cheapest, like the easiest thing that you can bring to such a picnic. And this lady told me, here's a different method. Just tell everybody, bring your own thing.

I'm like, are you crazy? You know, people are just going to not bring anything or I mean, whatever you want. Right. Or people are going to read the same thing. Like multiple people are going to, I don't know, bake some pie or do something. Right. And she's like, no, that's not going to happen. So I trusted her. That's maybe a trust word, but we tried it out. And what happened was really kind of fascinating. People were going out to the specialty stores and bringing like specialties, like

I'm based in Israel. So they were going to this hummus place, which is like an Israeli thing. And she's like, driving 30 miles to your favorite thing. People were baking and making stuff. So we had literally a feast. And the funny thing, two things happened. Everybody was much, much happier, right? They were happier because, of course, they got better food.

and then you're like and also most people kind of just their personality they brought it to the tables like i really like hummus i don't like the whatever the other thing i would have to bring and we did it every year afterwards because we did this thing at least annually and it worked every single time so if you take it the software that you can't tell everybody you know just develop your own thing but if you can guide them towards hey do the thing that you know give you more autonomy essentially bring yourself to the game be yourself don't try to sort of

put yourself in a box. I truly believe you're going to get much better results, short and even more importantly, long-term, because it keeps people thinking, it keeps them being motivated, and they're like, how do I kind of contribute in the way I think is the right way?

Reminds me, I'm looking for daycares for our son. He's like 17 months, almost 17 months now. And there's this Montessori approach to teaching kids. And it's a very similar approach, which is just let them, if they're ever busy with anything, don't even make eye contact. Don't interrupt them. Let them keep doing the thing and let them choose what they want to work on.

Yeah, there's many, many of these education systems or principles that are along those lines. I mean, the person who told me that, I don't assume she's invented it, but we all err on the side of wanting more control. But I do the same thing with my kids. So I never installed any piece of software on my kids' devices. So not like firewall protection, antivirus, air attack, nothing. Because I'm like, this is your problem.

And, you know, if you want to protect yourself, it's your responsibility. So this is autonomy. And there was one time where I had negotiated with my daughter. She was like, I told her, I think you're kind of using your computer too much. We negotiated. She said, maybe an hour is enough. I told her maybe more. I think we agreed on a two-hour thing. And then she came to me three days in a row. Could you please install the software on my machine so I can help me, like, control my limits?

And I love it when it's the other way around because now she's responsible. I'm helping her versus the other way around. So absolutely, I take it to my personal life as well. So how does this look day to day at Gong on the product team? Like when someone hears, oh, you give them a lot of autonomy, what does that actually look like? Help people understand what that actually means.

It means that if you're working with design partners and you get like an idea from the customer, it's your responsibility to decide, are you going to do it or are you going to talk to your manager or to me? You know, now I have four script managers, but it's your responsibility. So we're not going to quote unquote punish you if you decided that, you know, you kind of took it an opinion from a customer and went ahead and did it.

It's your responsibility to decide, do I know enough? Do I need more input? How up do I go? So that requires them to think, how confident am I in my decision?

So is the culture basically you give them feedback and advice and the teams can operate the way they want, they can build the features they think are important, work with design partners that they think are important? Yes. And of course, you are expected to solicit feedback, right? If you're going to build your own thing for six months and it's going to be, well, we're going to review it along the way, of course, but we expect you to initiate the review. We have a weekly session where you can bring up your reviews.

But it's not us forcing you to do it. You have to bring it, you have to solicit it, and you have to sort of drive the process.

This episode is brought to you by Vanta. When it comes to ensuring your company has top-notch security practices, things get complicated fast. Now you can assess risk, secure the trust of your customers, and automate compliance for SOC 2, ISO 27001, HIPAA, and more with a single platform, Vanta. Vanta's market-leading trust management platform helps you continuously monitor compliance alongside reporting and tracking risks.

Plus, you can save hours by completing security questionnaires with Vanta AI. Join thousands of global companies that use Vanta to automate evidence collection, unify risk management, and streamline security reviews. Get $1,000 off Vanta when you go to vanta.com slash lenny. That's V-A-N-T-A dot com slash lenny.

Is there an example of a product or a really important feature that came out of this way of working where a team just like, "You don't think this is a good idea. I'm just going to do it anyway." I don't think it goes up to a whole product. It's kind of very hard to... Because you got to have resources to build a whole product.

But I do think there are substantial features that came out of it. Even the AI fine tuning example I gave you before, it's like something came up in a hackathon and people were like, "Let's start to build it. Let's incubate it." And then they come and move forward. Of course, we had to give them resources at some stage, but it wasn't like a top-down, "Let's do this." It was more like, "Hey, we're trying it out. Hey, we need a couple of more resources. We're trying out more." And at some stage we realized it's super important. And then we kind of quote unquote funded it completely.

So when people listen to this, some product leaders might be thinking, oh, I want to work this way. I want to give my teams more freedom, more trust. What needs to be true in your org for this to work well versus it become chaos? Firstly, you as a leader need to sign up to let go a little bit, not control everything, willing to make some more mistakes than maybe you'd make otherwise. So that's the one thing.

I think the harder thing is sort of, at least for me, is you also have to sort of get your peers in the same, on the same boat, right? Because, you know, the head of sales is going to ask you, hey, what's happening? How do I know what's happening if you don't have a control over every feature? And the CFO is going to ask you, hey, what's the, I don't know, what is the ROI of this or how do you justify those type of decisions? So there has to be some fundamental trust within, you know, you and your team, you and your colleagues, right?

to at least experiment that way. And of course, if you do it on an ongoing basis, you lose some visibility. And I think that's maybe one thing you got to acknowledge, right? Because if you give people more control, by definition, you're going to have less visibility in what you're doing. So give up a little bit of visibility, hopefully get the benefit of higher velocity and higher quote-unquote morale or sort of like engagement from people. And that should result in better products as well. I love that. That's a really good example.

With the sales example, which is great, do you encourage the sales folks to talk directly to the pod to ask about these sorts of things? Or do you discourage that sort of communication? Oh, yeah. I think sales, from my perspective, are part of the virtual pod, right? So the core pod is, as I mentioned, product engineering. But part of the virtual pod is there's product marketing, customer success, sales.

In all fairness, salespeople are usually busy doing their work versus actually sitting with us and helping us kind of, "Is this going to sell? Did you hear it from customers?" The type of questions that product managers usually want to get. But if they're happy to spend time, they will be very involved. Coming back to the design partner way of working, does it feel crazy for companies not to operate this way, to not work this closely with design partners on new products and features they're building?

I wouldn't go back, right? So I think it's just like,

Even like, I hate terms such as risk because that's a very, I don't know, ambiguous term, but just the risk of building something you're not going to know if it's going to get used. And I was asked by sort of a very senior product manager of a very successful big SaaS company. It's like, why do you even do this? I'm like, what do you mean? It's like, hey, we launch products and then we see if people like them. It's like, well, I don't think that's a great idea because that company is successful, bigger than gone, but at the same time,

I just think it's leaving too much in the hands of, I would even call it luck, right? Because how do you know? Yeah, like I was thinking, it just feels like a cheat code and just feels like something a lot of companies can learn from how you all operate there. Something that has come up a bunch so far in our conversation is your focus on speed and optimizing for velocity. Something that I've heard about you is that you're really big on just making really quick decisions.

Even like one-way door decisions that are really big, your philosophy is just make it quickly before you have all the information necessarily.

Talk about that approach. That's maybe a little bit of a personal thing, but I would encourage people to look up in Google. Maybe I'll do a spoiler for Isaac Asimov. He's a science fiction writer, I think, beginning of last century. And he has this short story, The Machine That Won the War. So you can look it up. It's a fun story, pretty short. But it's basically the computer, the big computer at the time, supposedly won the war. But the only reason it won the war is because they wanted the people to trust this superhuman machine.

but what they realized the machine was just giving crap, just like our LLM hallucinations. So the head person, president, whatever it is, basically ended up saying, "I end up tossing a coin." But people wanted us to really believe that this is a smart machine thing that's going to help us win the war. So

It's kind of obviously a funny kind of story, but I think there is truth to it. So many, many decisions when it's not a close call, it's like, should Gong open an office in China right now? Well, probably not. There's so many reasons why not. We don't really debate it.

But should we develop feature A and feature B and you look at them, they're kind of the same, I don't know, call it same value or same cost or whatever. However, whatever kind of mental framework you have for deciding, you end up being like 51, 40, 49. No decision is going to be super wrong.

So, yes, you can try to bring in more data and you can try to sort of like bring more people. But like both decisions are OK. So just go ahead with one. Hopefully it's not like a huge, huge mistake. I'll tell you, I had this discussion with my co-founder, Amit, who is the CEO. And a few years ago, we were considering buying a company. And, you know, it's like pretty strategic decisions, right?

And we're like, oh, we don't know. There's pros and cons. We're like on the fence there. And we end up not buying the company. We kind of look up Gong and we haven't bought like big companies. And I asked him, maybe it was a couple of months ago. It's like, what if we had bought this company? Do you think we would have been in a radically different position? He was like, no, no.

So it's like, I mean, it could have been better. It could have been worse, but it would not have like made a huge difference. And the reason is it was a 51 49 decision. It wasn't a 70 30 decision. So it's hard for humans to make decisions. You probably know there's like, it's like, there's like experiments that show it's, it's almost like running or jogging or doing something that is a physical requires your physical capacity to make decisions. So just make it, I know it's hard and you want to postpone it, but just, just, just do it.

It's such a freeing way of thinking about it. It's interesting because there's been recent conversations on this podcast where Spotify has this kind of value they call talk is cheap and it's meant to be a virtue. Talk is cheap, so let's just talk a lot before we make a decision. But it's specific to Spotify because there's like a lot of regulatory challenges. And if they make a big decision, it's a long term. It's like they put a lot of effort into it.

So it's interesting, there's such different ways of operating. There's like, let's just talk for months and make a decision versus we're just going to make it and then it'll be fine. Yeah, of course, like big, big one door. Yeah, I mean, this is, of course, you're going to spend more time on, but people tend to overthink, I think, decisions. I also found out personally, the quality of my decisions, if you sort of kind of wake me up in the middle of the night and ask me, you know, what do you think about X? And I'm going to be like, I have no idea, I'm slipping.

And then you were like, you know, you got to force me into decision. I'm going to make a decision. Now you're going to give me two weeks to ponder over it. I don't think the quality of the decision is going to be much, much higher, which is maybe personal. It could be personal, but that's at least what I found out over my too many years of existence. Yeah.

I think something that's probably necessary for that to work out well is having a deep experience in that space. Like you've been at this for a long time. So I imagine your instinct often is trained based on your past experience of the market and customers. You feel like that's a necessary component of

trusting your gut and instinct on these sorts of decisions. Yeah, of course. You got to at some stage know what you're doing. Yeah, if I were now to sort of make a decision around, I don't know, entering a different space, there's no way I would be like, yeah, let's flip a coin like the Asimov story and go for it. I go to a conference, learn it, whatever the thing is, and then make that decision. But most of the decisions all of us make on a day-to-day basis is our domain of expertise versus like totally new things. Awesome. Okay, let's talk about AI for a bit.

You guys were very early on AI. You were working on, basically, it was like your product was built on machine learning back then is what it was called. Before it was cool and everyone probably thought it was like a waste of time and like, no, it's never going to work. Now everyone's building AI, building AI into their product. What have you learned about working with AI over the years that you think people maybe are not yet aware of or that will likely cause them pain that you can help solve and avoid for them?

Yeah, funny. You know, when we launched Gong, we didn't use the term AI because people thought it was a bad thing. It makes wrong decisions or they just thought it's an action item, an acronym. When we founded Gong, I was in sabbatical and actually went to this deep learning course because I was bored, you know, in all fairness. And

And after that course, I ended up buying Nvidia stock, which I wish I'd kept up until now. But I did send an email saying, hey, this is the next thing. So we understood it's the next thing. Of course, we didn't know it's going to be LLM and GPT and other acronyms that evolved over the years.

And probably now that we're talking kind of end of 2024-ish, I think people should not go from one extreme, which is, hey, we need a bunch of data scientists for every small project, which was the case five years ago, three years ago.

to the other extreme, which is, hey, LLM is going to solve everything. Because LLMs don't solve everything. They have huge utility. We use LLMs all over the place. Most companies that develop AI kind of stuff use LLMs. It's a great thing. But at the same time, don't assume it does everything. You still need some of the core competencies of AI. So you do want to have expertise, people who actually know what they're doing.

and help guide us VMs around, you know, is this something that can be built or no? Because if you're going to spend many, many hours on asking an NLM to do, I don't know what, like...

In the case of GON, for example, tell me what a good sales cycle looks like. LLMs don't do that. It's just like maybe something else does, but we have a deal prediction model. LLMs cannot predict deals because it's very, very specialized. So I think you still need to have expertise. You still want to have some measurements.

So yes, version one, you can just go to an NLM and say, create something, I don't know, whatever. But if you don't have measurements, like in the old machine learning, whatever metrics you use, you're not going to advance. You're going to have V1 and then you're going to have V2 and you have no way to know if you've made a progress. So we kind of pay a lot of attention to... We have people who are going to specialize in how you measure. We use Elo system, which is kind of the one used in chess as well. And we do have experts who kind of help us make the right decisions.

You can make very good progress without these, but I think there's a glass ceiling if you don't figure out how to create more operational rigor around this whole AI thing. So what I'm hearing is don't assume you can just outsource all your AI magic model building to the foundational model companies. You need to have your own AI expertise, ML expertise.

Yeah, or even if you end up outsourcing the core work, at least you have to have the expertise to understand what is doable, what is not doable, what's the right way to approach it, what's the input you give to the LLM, how is this going to be good quality or bad quality. There's even like, if you just take the product management aspect, if the LLM gives you something that is 90% accurate, or I don't know, people are going to think is good.

the product's a little different than if it's 50% good. So just the way you even think about it, the way you... I think Figma calls their AI feature like first draft.

Which is a term I like because they kind of realize it's not best, it's not great, but it's a good first draft. So if you know what it is, it's easier not just to name, but how to conceptualize, how to build a workflow around it, and what to train users to assume for it. And I think there is an expertise there that comes on top of LLMs, even if you just use LLMs and you can't afford or you don't want to go deeper.

For folks that want to do this at their company, what are the functions that you have that help you do this slash skills of people you hire that you think are important?

So I think you still have to have this kind of quote unquote data scientist role. And data scientists could be in the company, could be advisors as well, right? Not everything has to be a full-time in the company. And the role of a data scientist is help guide the company, right? Deal prediction model. Is this an LM thing? Do you need to build a model? If you need, what input do you need? How long it's going to take? Right?

Also, in our world, at least, data scientists are the people who know how to measure these things. Is this model better? Is this model better? Is this prompt better? Is this prompt not better? And they're not going to do the judgment, right? So when Gong creates an account brief, the data scientist is not going to know if that brief or this brief is the right one, but they can kind of guide us through what's the right tool set you need to sort of put it in front of customers and how do you measure this and whatnot.

And then I think at the end of the day, you also need this myth, now it's becoming common. The prompt engineer, the person who's actually working with the NLMs and guiding them, that is a bit of a technical skill, but you got to have it. It doesn't have to be a full-time person, but there needs to be that expertise of somebody who's actually optimizing things. Many, many customers tell us that Gong AI is more accurate than others.

Yes, there's some combination of models we built from scratch, fine-tuned, because we have AI expertise. But some of it is also how we kind of... The prompts we give to the LLMs, how much rigor we put into optimizing them and kind of finding the edge cases and ranking them and improving them over time. If you want to get really good AI, you have to invest in it. As you're talking, I'm thinking about how

Your pod model matches really well with this world of things moving so quickly, AI changing constantly. Just giving teams autonomy feels like a huge advantage in this world where things are just changing weekly.

Yeah, so we have a couple of, maybe now it's three different pods. We have like an embedded AI specialist team, either a specialist or a team, or I don't know, a couple of people. And then they can iterate very, very quickly on using LLMs or using non-LLMs, you know, SLMs, people now say small language models, but whatever the thing is, they can iterate very, very quickly. Awesome. Okay. A couple more things I want to touch on. One is the spiral model.

So you mentioned that you just went to learn deep learning on your own. You like went off to the side. I'm going to understand this new thing that everyone's talking about deep learning. And you got really smart and machine learning basically really quickly. And you have this thing you call the spiral model or the spiral method for how to learn something complex quickly. You wrote a medium post about this or blog posts. What is the spiral method? How does it work? How do people learn things really quickly that are really complicated?

Yeah, I think it's even beyond just the speed, but also like, how do you even know that you learned that you actually learned it? So it's kind of there is a mathematical kind of or physical concept called annealing, which is how certain kind of material kind of becomes the way it is. And it's sort of the temperature goes slightly down and eventually become a crystal or whatnot. There is an element to this, I think, in learning as well, which is.

You want to know what deep learning is, like you know nothing. You go find the person next to you and you're like, what is deep learning? They tell you something. Of course, you don't know anything because you just heard it from one person. And the next question you should ask, like who else should I be speaking with? They give you three other names. I think in tech, we all tend to be like this very, very kind of cool ecosystem of people who are willing to help as long as you don't ask too much of them. So then you speak with three other people and then they give you like other names and you sort of go around.

And ideally at some stage you feel like, you know, first person, you have no idea what they're talking about. You probably didn't even understand what they're saying. The fifth person, you might understand 50% or 50% is like new. At some stage, you're going to feel like, well, new stuff is 10% or

or 5% or 0%. I call it a spiral because it's going to go in circles around the target. And eventually you feel like, well, I'm hearing the same thing again and again. And you're like, well, if I heard it from three people, I didn't learn anything new. I'm sort of at the bullseye. Of course, at the level I am. So I'm never going to be like a deep learning specialist in the same way that true data scientists are. But as a product manager, I know it probably as well as I can, given that everybody I'd spoken with at the time was

not giving anything new at the level that I had desired at the time. I love that. Is there anything you've been studying recently that you've either used this method for or something else you're excited about learning that's new or on the cutting edge?

Usually I kind of do this for kind of use cases within our customer base. For example, if I wanted to sort of, if you wanted to go after a certain persona or a certain use case for the product. So we had this notion of can we do a better job for a specific persona within sales, people who are account managers.

So I would use a similar method. Like, hey, talk to one account manager, talk to an analyst or whatever the thing is. And eventually when you start hearing the same things, like what do they care about? Is it different than salespeople like selling new business or different than contact center sellers? When you start hearing the same thing, you're like, okay, I kind of got to where I need to be. Now I can make decisions.

I can always do another spiral and get one level deeper, which is, I don't know, do some user research, go all in. But at least at the sort of the conversation level, I've got it. I've got where I need. I love how simple this is, is you just start talking, just find somebody to talk to, ask about this. No pressure. And then just, OK, who else should I talk to? You just keep having conversations spiraling deeper and deeper into knowledge and wisdom.

Okay, one thing I wanted to touch on, which has always stuck with me about your approach initially when you were starting Gong, is how you found your initial ICP, who to go after. And it's really funny.

how narrow you got when you all decided here's who we're focusing on for our first dozen customers. So I have the list here. So when you decided here's who we're targeting, here's the list of constraints. We're going to target people selling their product in the U S in English over video conference, uh, using WebEx, which was the big one at the time selling software that is worth 1000 to a hundred thousand dollars. And, uh,

And there was only 5,000 companies in this bucket. Can you just talk about like why you found it was so important to get so narrow and just the power of getting really narrow, which is very counterintuitive to a lot of people where they're like, oh, we're just going to be for everyone. It's a huge market.

I think it's sort of the traditional, I call it the bowling alley or whatever you want to kind of form. Crossing the chasm. Yeah, the crossing the chasm kind of methodology, which you want to start narrow. You want to create this kind of small pond where people talk about each other and you can kind of light the fire in there. In my previous company, by the way, I did read Crossing the Chasm and I told myself, nah, I can do way better than that.

So we had one customer in, I think it was L'Oreal or I don't know, one of the cosmetics companies and American Express and Cisco, like different industries. And there was no way we could scale it because everybody had their own lingo, the way they thought about the technology and whatnot.

So by having a smaller set of kind of customers or I see kind of definition of customers, you can develop like much more focused. And then it's easier to light the fire because people move, right? At some stage, I think it was year one into the business, we heard from a company that they interviewed a salesperson. And the salesperson asked, are you using Gong? And they said, we are thinking about using Gong, but we're not. Like, well, I'm only going to work for companies that use Gong.

And that's the power of a small pond with companies that are like each other because you get this viral effect that is not common in B2B, but it's as close as you can because of those conversations. That other customer became a Gong customer literally because they interviewed a person who told them he's not going to come unless they bought Gong.

You can't do this if you have a wide market where people don't even talk to each other. And there is an assumption that you're not like burying yourself in this market. I love because today, like I said, at the very top of this conversation, you're just so ubiquitous. Like everybody seems to be using Gong. And I love that you started with something where there's like seven, I don't know, different constraints to narrow down who you're going after. And it's such a good example of the power of starting very focused and then expanding from that, which is what you've done.

Okay, last question before we get to our very exciting lightning round. We have a segment on this podcast called Fail Corner, where so many of these podcast conversations, everyone's always sharing all the successes, everything's always going great, nothing ever goes wrong. When in reality it does, things often go wrong. Can you share a story from your career or just the journey of Gong when things didn't go well, when there was maybe a failure? And if you learned something from that time, what you learned?

Yeah, I always kind of joke that in my previous company, we've done so many mistakes that if life limited you to a certain number of mistakes, I wouldn't have any left, I think. I still do mistakes, but just so many. So every one of them probably done twice or and then it's like some stages like, you know, third time's a charm.

So the one I just gave you is like probably the worst is, you know, crossing the chasm. You start a company, you have this like technology. I was thinking, let's go horizontal. And that technology was whatever web integration, something eventually ended up being an e-commerce content syndication or content management SaaS software, which is the right way to go because you want to specialize in a certain market. But initially just going all in was like just ridiculously not smart and

And the other thing we did together was like, that was, the previous company started year 2000. So that was like the bubble, one of those very, very nice bubbles. So we're like, you know what? We actually got three customers, admittedly in totally three different segments. Now let's go and scale. Now we only, we get like, we need like one salesperson, one SE and C kind of do what's now called product market fit. I don't know that the term even existed there.

and we're like nah you know what investors told us you gotta hire more people so we hired under 20 sales people all of them failing miserably uh because hey we didn't have a true product market feed but even what's worse we didn't have a true focused icp with like a very very repeatable uh product market fit so if you sort of

hear me talk about sort of how we started Gong. Amit is the CEO and he kind of drove it out of that business strategy. But sort of me being sort of a co-pilot there, it definitely bringing the same. I'm not going to make that mistake again. I might do new and fun ones, but not that same mistake again. Awesome. Thank you for sharing that. With that, we reached our very exciting lightning round. Are you ready? Sure. Let's do it. First question. What are two or three books that you find yourself recommending most to other people?

There is a set of books. I think one that is sort of the starter one is, I think it's called right now, The Ideal Executive. People don't really know it's sort of a management book, how to run a team and whatnot. I think the original version, funnily enough, I think it was called Mismanagement. But nobody wants to buy a book called Mismanagement. You'd much rather buy a book that's called Ideal Executive because you, of course, are not mismanaging. You're the ideal executive altogether. So you're just reinforcing yourself.

But jokes aside, it basically kind of gives you the... It tries to sort of define people by four characteristics. I think misnamed, but like, are you an administrator? Can you like... Are you... He calls it a producer, basically get the job done. Integrated, which kind of brings people together. And the fourth one, it's basically kind of change agent, you know, kind of do a lot of mess and change stuff. Usually entrepreneurs kind of include that part, of course.

And basically his claim is like, nobody does the whole four. Maybe you're good at one, maybe okay at the other. And personally, I'm horrible in administration. So I obviously acknowledge that and I try to sort of compliment myself. So I think there's two things in it. Firstly, just those, I thought there was like the four good ways of looking at people as a manager, as a leader, of course.

That's one. But I think even if you disagree with those four, just understanding that you want to look at the people in the organization, yourself included, and it's through the prism of key characteristics, and you can select a different framework, helps you a lot with

creating high velocity discussions with others because I can talk with somebody and say, hey, you're a P. It's like, well, I'm not a P, I'm an I, whatever the thing is. And that makes a discussion that is like much, much faster and more comprehensive than just like trying to explain this from scratch. Like, hey, you tend to do this and you might want to do this and you might want to strengthen that. So

I'd recommend starting from this, but there's probably other methodologies you can pick and maybe kind of some of the listeners here have already had one. But that's one I like because I kind of found it useful. And it's called The Ideal Executive. I think so. I'm pretty sure. Yeah. Great. Any other books before we move on?

That one's going to probably work. I like crucial conversation. That's going to be the beaten path. It's like how to conduct conversations with people in your organization. I think it's never bad to sort of re-immerse yourself into how to speak properly with other people.

We have an episode coming up where we're going to share scripts and phrases to use to have better hard conversations. Ah, that's good. Yeah, I'm excited for that. Slash scared. Okay, next question. Do you have a recent movie or TV show you've really enjoyed? I didn't have TV, like broadcast TV for many, many years. So nowadays there's Netflix so you can find stuff. But in my case, TV and movie tend to be pretty esoteric, fringe.

So I've recently watched this British TV series called Slow Horses with Gary Oldman. And it's a really kind of fun, you know, sort of funny spy thing, which I found amusing and intelligent at the same time. So some kind of comedies tend to be pretty kind of lowest comedy number there. That one seems still fun and witty at the same time. So that's my latest that I kind of really kind of enjoyed watching even the third season. So.

I love Slow Horses. It's like, I don't think it's that obscure. I think it's like one of the ones Apple promotes often. I will say this last season was not my favorite, but the other two were awesome. Yeah, I 100% agree, which I said, even the third season was okay, but the first two were really, really good. Yeah, it's super fun. It took me like three tries to actually get into the show initially because people kept telling me it's so good. And I started watching and it's just like, who's this old messed up guy just complaining endlessly? But you got to keep watching.

Okay. Do you have a favorite product you've recently discovered that you really like?

I assume you have a silverware caddy in your dishwasher, right? Oh yeah. To put like forks and knives. Yeah. Cutlery. So that's my favorite product as of lately. And I'll tell you why. It's a funny story. I lost mine and you can ask yourself, how can you, you know, freaking lose like one of those baskets. And it was in the, in the dishwasher, of course. And for some reason I couldn't find it. That's like, you have to be really kind of out of your mind to not find it. Anyway,

So I go to some Amazon or eBay or wherever, I just buy a new one. And then of course, a day later, I find it's like a fence somewhere within the dishwasher. Now I have two. So this is my latest invention. If you have two of those baskets, you can put one of them in the sink and you can just like continuously load your cutlery or silverware while the thing is working or you haven't vacated it. So it kind of changed how we organize our kitchen with something that probably costs 10 bucks per

No product manager has ever thought about offering two of those with your dishwasher. I don't think they even tried to upset any of that.

And I told it to some people and actually ended up buying a second one and suddenly it was successful, which is the most ridiculous thing. It's like, you know, spend 10, 15 bucks, get something organized in a completely obscure and unintentional way. I'll give you an even crazier idea that a previous guest suggests. Rory Sutherland has this pitch that you should have two dishwashers. Everyone should have two dishwashers because

One is your clean and one is dirty. And you just take your plates and things out of the clean one, use it and put it straight into the dirty dishwasher. And why are we just putting things away constantly? Just like go from one to the other and one to the other.

So there you go. Similar idea. Similar idea. Like 15 bucks, so a little bit maybe cheaper. Exactly. Houses are not designed for two dishwashers. Okay. Two more questions. Do you have a favorite life motto that you often come back to find useful in work or in life? One that I use is going to sound funny, but it's actually real and I use it and I actually believe in it.

I'm not sure if you know, for philosophy, there's like razors, like Occam's razor, which is basically- Oh, there's other razors. Yeah, there's so many razors. And there's one that I think came in in some Murphy book or whatnot, and it's called Hanum's razor. You can look it up, Wikipedia, wherever. And it basically says, it goes like, never attribute to malice, that which is adequately explained by stupidity. Mm-hmm.

So it obviously sounds funny and it's trying to be funny, but it's so helpful because so often do we attribute people's behavior, thinking of company, customer, I don't know, personalized sometimes to malice. Like, oh, this person's not returning my calls because X, or this person hasn't given me feedback or has given me feedback because of X. And it's sort of like we all, I think there's a saying, it's like always assume well or with intent or whatever.

And this is sort of the more funny way to sort of say that. Yes, the person is. And again, stupidity is only a funny way to do it, maybe inappropriate. But it's basically, yes, they just didn't know. They didn't care. They didn't think about it. They weren't trained, whatever the thing is. And if you take this model in your day to day life, at least I find that it's it's so true and so often true. That is funny, but inspiring. I really love that quote. I think of it often when somebody is doing something that's annoying me.

Final question. You mentioned that you are from Israel, you live in Israel. You mentioned delicious food. Hummus is one example. Is there another Israeli food that you think people are sleeping on that you think people should try when they have a chance? Israeli food has become a little bit, you know, kind of got a little bit to be in fashion lately. So people are coming from the U.S. to visit us in the office are like, oh, Israeli food is so good. And I'm like, what do you mean? It's the same food we've had for like 20 or 30 years. I think the taste changed because you kind of eat more healthy and less oily these days.

Most of the Israeli food is sort of Arabic in nature or Turkish. So there's great falafel, great hummus, pita bread, Turkish delights of sorts. So a lot of those. And some very, very obscure. And if you come to Israel, I'll show you around some very less known food that only kind of special guests get to taste. What's that? I can't say it here because... Oh, you can't say it here. You have to say it now.

There is a thing called sabich, for example. Nobody knows of it. People claim it came from the people who came from Iraq, but my wife's father came from Iraq. He's like, we've never seen this before. It's sort of pita bread filled with hummus and eggplant and eggs and maybe something else. I have no idea. Tahini, maybe. I don't know.

It tastes good, but it's such a weird combination and it's become a little bit of a thing. And nobody knows what the origin is. I think it's some somebody made a mistake and gave it a name. And now it's like ubiquitous. You are making me hungry. Elon, this was amazing. Two final questions. Where can folks find you if they want to reach out and learn more? Maybe ask some follow up questions. And how can listeners be useful to you? I'm

Pretty available on LinkedIn is probably the best way. I tend to kind of read my inbox and LinkedIn and respond when I can. And then useful to me, I mean, if you want to come work for Kong, check out our careers page, of course. The product team is mostly based in Tel Aviv and Dublin, Ireland. So yeah,

maybe a little bit remote for most people, but there's sometimes folks in the US and sometimes non-product. Of course, Rog, we're hiring quite a few people these days, so we'd love to at least give us a chance. Awesome. Elon, thank you so much for being here. Thanks for inviting me. Bye, everyone. Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.

Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at Lenny's podcast dot com. See you in the next episode.