The Future of Hyperautomation with Truly CEO Erol Toker

Erol Tocker: The long run value of humans is going to be creativity, right? Not solving problems, but knowing what problems are worth solving.

Sean Lee: Hey everyone, welcome to Operations, the show where we look under the hood of companies in hyper growth. My name is Sean Lee. How much of your time do you think you spend looking for or talking about benchmarks, trying to find those proven examples that you can copy, emulate, strive towards? I know I certainly do, but if we all fast forward, at what point do those benchmarks stop being aspirational and instead just become handcuffs? The way things have been done, so therefore, the way things must be? I've found that the people who question these limiting beliefs and look at problems through brand new lenses are typically pretty great people to learn from. They're the ones creating what will become the norm, the new benchmarks of the future. One of those people is Erol Toker. Erol, whose voice you heard at the top of the show, is the founder and CEO of Truly, a company that helps eliminate complex repetitive tasks through a methodology called hyper automation. In our conversation, Erol teaches me all about hyper automation, which I believe will be the new benchmark for operations teams. We talk about the long term value of humans, how operators can rethink the difference between low value and high value activities and what his company did when they asked themselves what would it take to not have SDRs. To start, let's hear from Erol what he means about the concept of hyper automation and how he discovered it in the first place.

Erol Tocker: So it's really interesting. Hyper automation is a concept that exists in the back office of enterprises and it hasn't really entered revenue. It's a brand new concept that we're sort of bringing into revenue. And what's weird about it is, you look at all these public companies, like fastest growing companies of all time, Slack, Twilio, fastest to$ 100 million in ARR. There's a company out there called UiPath that beat all of them. In just a couple years, they went from below$ 10 million ARR to$100 million ARR. So massive potential in that segment called robotic process automation. And robotic process automation is a little bit like what sequencers do. They just automate away really low value tasks that are repetitive. I click here, click here, I copy it from my spreadsheet, I paste it into an email. Sequencer takes care of that. Hyper automation is kind of this next generation evolution of that where basically what you're saying is, okay, there's the low value stuff we've automated out away, but there's the high value stuff that I still need a human for and what I'm relying on is human judgment. So I'm going into an account, I'm prospecting into it, and I need to get a feel of this account. So I look at how many VPs of marketing are there, or there's three, well that means something versus one. And how long have they been in there and how fast are they growing? So all we rely on humans for a lot of this intuition. What do I say? What's the most salient thing? Hyper automation is basically saying, how can I take the latest things that are happening in AI and try to replicate and emulate that judgment? And if I can do that, and this is where it kind of gets funny, right? Because a lot of people say, Oh, the robots are coming. They're never going to defeat humans. It's not about that. It's more if I could take the things that are high value that you don't have enough time in the day to do, like researching your accounts and doing this is still high value, but it's not the same as sitting down and really thinking about a customer's problem and kind of looking at it from tech. That's the creative stuff. So hyper automation is just taking that automation we've had for the last 15 years to just the next level with the latest AI techniques.

Sean Lee: This is one of those times when I'm so glad we have these moments in our show to stop and reflect on what we just heard. Erol just described the evolution from the robotic process automation introduced by the likes of UI path to this whole new concept of hyper automation. And my biggest takeaway is that we as an industry have evolved past just automating these low value activities. It's not just about logging emails or populating fields. Erol is using words like judgment and feel to describe the work that is hyper automation. These aren't low value activities. They're what he describes as high value. What that means is that our lens as operators, that radar we use to seek out inefficiencies, has got to change. So how do we learn to do that?

Erol Tocker: The way that I frame it, the way that I think about it is before, if you think of ops just in general, right? There wasn't a ton of ops. It was originally just sales ops and sales ops was like under sales. And the job of sales ops was to make sales efficient. And so what is making sales efficient? It's like whatever the VP of sales says, right? That's what we're going to do. That's the job. And so what's the VP of sales say is like, well, it goes to the reps and says, Well, what's taking up time? And so that's one way of looking at it. I think what's happening with Rev Ops, and I think what you guys are doing at Drift, for example, that's having really big impact is flipping that and saying, Well, it's not about the sales reps and making them efficient. It's about making the customers buying journey efficient and pleasant. And then as a result we become efficient because we've cut all the fat out of the equation. So I think when we're talking about what is a low value activity that we're automating, the first thing you have to do is think about, well, low value from what lens, low value for whom. And then once you know that, then you can jump into and start having the conversation on what's a low value activity or what's a high value activity.

Sean Lee: So can you give me an example of one of those ones? If you're viewing it through the customer lens, can you give me an example of one of those activities that hyper automation might make different?

Erol Tocker: Let's say that I want to go into an account and I don't really know what that account looks like or how it works, and I have a flow, so I have to go and find the people and guess the emails and find the accounts. And some salesperson might say, that takes a lot of time. And I'd say to you, Hey Sean, that's not selling time, that's not sales. I'm doing non- selling activities. And then you'd go and buy a Zoom info or an app Apollo or something, get the contact data, give it to me, and I'll just drop in my emails and just essentially span the account. And so I see spam here in the sense that some of those are high value going to the right people, but some of it is spam, which means that they're getting unsolicited outreach they don't want. And so from an internal view, that's like how you would say, Hey, we're making sales really efficient. Bill or Jane over there is no longer doing this really inefficient activity. They have all the selling time. But at the same time, I'm doing untold damage to that account. It's a bad experience for the customer. Maybe Bill is the wrong, or maybe the person that you're getting is the wrong person today and six months later they get promoted and they have this really bad impression of you and in the long run, and maybe the rep who sent that isn't even there. And so somebody inherits this account, it's like, why won't these people talk to me? And I think that's this myopic more sales operations, inward focused approach. So with hyper automation you'd say, Oh man, I got to go on LinkedIn, I have to go look at these people. I have to try to figure out the personality. It's like how do they all fit together? And it's a lot of unstructured data. So it's not like I can just sit there and look at, like some people put in a headline, some people don't. So it's not like if this than that. You kind of have to apply a lot of nuance to that analysis. And even how you find that analysis, you may have to leave LinkedIn. It's not just like, Oh yeah, I'm going to go scrape LinkedIn. I may need to actually go to some other network like Twitter and go find more information there and start stitching it together in different ways. So to me that is an example of a high value automation with a customer experience in mind versus a low value automation that's really centered on the seller. Now both impact the seller. It's like are you looking at it in a time horizon of a day or are you looking at it a time horizon of three months, six months?

Sean Lee: So not only do we as operators need to make the distinction between high value and low value activities, we need to view any improvements that we're making through that external customer lens first. Enabling someone to enroll 500 people in the same sequence doesn't do anybody any good. So seek out the better alternative to that. Erol isn't saying that that's easy. Think of it this way. Erol pointed out to me that this elevated or this hyper automated way of doing account research is exactly how you'd teach the perfect SDR to do it, but just the SDR doesn't actually have to do it. The AI workflow does it for them. And if that sounds complicated, it is, but maybe not as complicated as you think.

Erol Tocker: I was talking to somebody today about how all this works and we were looking at a very small part of the problem and we were having a conversation code, no code. And I said, what's funny, is we originally prototyped this in a no code tool and no code is part of it, but we've written 15,000 lines of code around it and then it's a very small part of it. And the person was like, Okay, that's really complicated. But if you think about it, 15, 000 lines of code versus your 50 billion neural connections, millions of dollars taking you from kindergarten to whatever, I think people really underestimate just how much value add humans are adding in the part of the sales process. It's getting to know you and finding targeting. The long run, to me is, it's so clear. The long run GPT3 is out, AI's already creating init. All that stuff. The long run value of humans is going to be creativity, not solving problems, but knowing what problems are worth solving. So when I think of, when I go into an organization, I don't even, when you think of the efficiency pitch, I only focus on one thing, which is how much time are your reps speaking to other people? That's the only metric that I think matters. And the reason for that is if the true value that a person can provide is exploration and creativity and stringing a problem together, if that is the maximum value that a person can provide, probably most of that is going to happen in spoken form or collaborative form. So it doesn't really matter to me if you are, like we've automated away all your activities. If you have two hours of speaking time a day, I don't care how efficient you think you are or how much from an internal lens, I made sales 300% more efficient, doesn't matter because the maximum value that they can provide is two hours of the day. That's the utmost maximum value that they can provide. And so I think that's how I think about the shift with Rev Ops is it's like how do we rethink what our resources are and how do we align them to customers in different creative sort of ways to maximize that airtime.

Sean Lee: So you and I talked about this a little bit and I mean it seems like you're kind of throwing away many if not all of the traditional measures of what makes an ops team good and also what the right way to measure the output of a sales organization is as well. So can you talk about this evolution that you've gone on from that kind of more traditional lens that you're talking about to the way you're thinking about it now, which is I think at its core talk time per customer per day.

Erol Tocker: So first of all, I'm not redefining anything. I think the game has already been redefined by the customer. And I think again, the company you're at played a huge part in it. The whole premise of sales operations and making internal metrics, the whole premise of that is that we have leveraging control over the situation. 30 years ago, we had control through information asymmetry. So it was like I'm in the Chicago market, I'm the only person who sells the widget, I'm going to call in this prospect. Or maybe there's five, but what are the odds that five people are going to call in the prospect at the same time? That was the first information, that was the first leverage. Then the second leverage came from, I can just achieve, I can automate these low value tasks, I can achieve it. And today there's kind of five versions of every app. And even though they're different, they all make the same claims. The thing as a buyer is all I can go off is the claims and the cost of exploring those claims with you. And I don't always buy the best product, I just do my best to predict what that ROI is going to be. And so now, it's like you can bet me all you want. If I can book a meeting off of Drift and I get to that first account executive and they get me into the product release, that's what I'm going to do. I'm just going to do the easiest thing.

Sean Lee: The long run value of humans is knowing which problems are worth solving. The long run value of humans is knowing which problems are worth solving. Erol has me convinced that our creativity is not only our most important asset, but the means through which we provide value in this new world. I talk about this with my team all the time. If you're just a requirements gatherer and you go off to do exactly what someone else thinks is best, you're missing out on the opportunity to truly shape and contribute to the outcomes at your company. When Erol, we have to rethink what our resources are. He's talking about us too. So how did he arrive here? This wasn't always what Truly did exactly. And any good entrepreneur, he solved his own problems and realized others had them too.

Erol Tocker: Our company came out of a pivot and in the middle of that pivot, one of the things we had an opportunity to do was kind of look at everything that we've done up till then. So we just gotten profitable, we're kind of looking around, we're like, all right, we did good, but not the best. We're going to pivot the business. What's been working and what hasn't been working. And one of the things that we found was we wrote down all the things that we hated doing. And what we realized was all the things we hate doing, other people hate, like the customer hates as well. So here's a good example. One of the biggest problems we had was quoting. We hate quoting. Why? Because after I give you, right after we do the negotiation discussion, I generate the quote, 19 seats. We go through the CPQ tool. We fill everything out, we generate this thing, we get somebody to double check that we didn't screw it up. Great. Send it to the customer. Oh, I want 20 seats. So we go and do that. You send it again and the customer gets back to you and says, Hey, yeah, actually admin licenses, I kind of forgot about that. And you're like, Come on man. And then you send it and then you have to follow up six times. Let me ask you this, how would we fix that with sales operations? What's the view to that problem that I just described?

Sean Lee: I mean, at that point it's all process, right? It's back and forth with the customer and really you have to have a rep to help make those changes.

Erol Tocker: But let's just say you have a really adamant VP of sales and is like, I want you to automate this with tech. What would you do?

Sean Lee: That's a good question. I mean, I think you try to make it customer facing where they could make some of those tweaks themselves on seats maybe.

Erol Tocker: Okay, so that is...

Sean Lee: Input.

Erol Tocker: That is incredibly futuristic, right? 99% of people that I talk to when I ask this question will say, CBQ sucks, right? Replace the CBQ.

Sean Lee: Yeah, yeah.

Erol Tocker: And then maybe we should import that person into a sequence to send them a daily reminder after the contract set. That's how, and I'm glad that you say that by the way, because again, it makes sense why you're an influencer in the community and all that stuff and why you work at Drift. Like 99% of people immediately say CPQ. And so what we thought was exactly what you said is, this is a shopping checkout cart. Why am I checking out somebody else's cart? It's like on Amazon, you go check out your own cart. And then what we realize is when you start working backwards is our pricing isn't clear. We're making stuff up based on the wrong. We got somebody through a funnel who's not a perfect fit, but we want the money so we go through this. And if you add up all the costs that's associated with that nonsense plus the new CPQ upgrade, that's going to solve everything, but it isn't. If you add up all that cost, this is a very unprofitable, inefficient organization. So I think my viewpoint, our view about Rev Ops came from that. We were sitting there trying to look at our own business and figure out why are renewals so terrible? Why is it so, why do I have to chase after this person? Really? We sold them. NPS is high, da da da da da. Why is that? And the answer is always somewhere really deep up funnel. And so the thing that solves these, and so if you're trying to solve every problem bit by bit, bit bit, you miss the big picture. And that's my attitude towards low value activity automation. It's like all right, we're making it easier to chase the person down for the renewal. We're making it easier to, and that's like, that's not the problem. The problem was like, hey, we weren't paying attention to the account when the new VP came in. That was the problem.

Sean Lee: Erol's progression here is very interesting, but probably not unique. We've all had those painful moments in our own processes and we just want them to be easier, but he's doing something about it. So what does the end of this rainbow look like? He's right that there are a million little things you could solve. So does that require a million little versions of hyper automation? Is that even possible? It turns out that over the course of two years, the results in Truly's own business speak for themselves.

Erol Tocker: So our MQ to SQL rate went up from 15% to 85%. The amount of time that we spend on demos, like before the demo, have to get your water and prep, it went from one and a half hours to two minutes. We don't negotiate anymore. We don't negotiate pricing. Upsells don't require recontracting. It's all consumption based pricing. Customers do their own CPQ, always. Basically we went and bought some off the shelf technology that's not very great. It's actually designed for SMBs, to sell the SMVs, but we just threw it out there and people just now check themselves out and it's just a stress for free environment. So that's the outcome. That's like the promise land or that's how we define the promise land for us. And that's what we were really happy with.

Sean Lee: That's amazing. So part of this absolutely a mindset shift, focusing on the customer facing components of this. To your point, there's definitely a drift angle here where it's like you are not just acknowledging that the power is in the hands of the buyer, but actually giving them more power, which is really interesting. Help me make this concrete for people. An hour and a half of prep, down to two minutes. Connect the dots for me on what a typical Truly rep was doing before to make up that hour and a half and what is I guess done on their behalf today such that their prep was only needs to be two minutes.

Erol Tocker: So the first thing we did was to basically say we're not going to have decks anymore. The website is the deck. So because basically we're going to show people, we're going to walk people through the website and what's there, We're not going to have this thing where marketing said something and then they came in through that and then sales is now showing a different deck and there's an misalignment, whatever. So that was the first action we took. And as we did that, what we realized was before you had to prepare your demo environment and get ready and get focused. Because if you click the wrong thing, the wrong thing's going to happen. And today what we do is we show videos. We show videos on the demo. So we have videos that are 60 seconds long and we're still not done. Those videos are now going on the website to make it easier to buy. But what we realized is, as ludicrous as it sounds, like why would I get on the phone with you to watch a video? Well, it's like that's all we had. But what we found is if we can break down the things that they want to see into pieces and show them on the demo and see how they react and copilot and ask them, So did anything stand out to you in that? What should we dive into on the call? You get all this valuable market research out of it. And so what happens is those videos improve and eventually what you find is it's the perfect demo. You need to, the wording is perfect, the length is perfect, everything's perfect. So we get on calls. And the other thing is we know that those videos work because they're aligned to the marketing, they all map back to the website. So here's a good example. When somebody complains about sales enablement, I'm like, you don't have a sales enablement problem. You're letting random people into your funnel. The reps don't know how to handle that and they're telling you they need another deck. The solution to that is not where sales enablement. Stop letting in the path unfit customers into your funnel and whatever you train them on is going to work just fine. Or enable them to hang up the phone, enable them to get off that call so that you don't create this downstream. We need ahead of enablement, we need more training, which takes away from the talk time. You don't need any of that. You just need a really clear funnel that each piece kind of works with the other pieces.

Sean Lee: The thing that struck me throughout my conversation with Erol is that the rest of us are victim to our own limiting beliefs about what Rev Ops is or what is possible through technology. And Erol goes into these conversations with his customers without the handcuffs of these limiting beliefs. In fact, he's made it incredibly simple on himself and his customers by following some role models of his own.

Erol Tocker: A hyper automation conversation is really, really simple because again, we focus on the journey. Most people just don't even know like 99.9% of what we're doing is possible. So we take them through a funnel. And by the way, so in that funnel, one of the biggest projects we did here was we took our product and we broke it up into 15 different products, just based on the AWS model. We were like, we spend$500, 000 a year on this thing and we've never really spoken to a sales rep except for when we were begging for them to accelerate the contract signature so we wouldn't be penalized for being out of contract and lose our distance. How do we that? And so our journey is much more oriented to how do I make this really easy for you? Hey, Sean, at Drift, you got our marketing, you get it. David might not get it. There's all these... They don't get it. I'm just like, how do I help you, Sean, get it and show one more person what it is? And it's never like, hey, show the full solution. And the idea, it's like how do we use the product as a vehicle to plant an idea inside that organization and then build out the product in a way where there's natural hooks into additional use cases where the product starts to tell the story itself. And I think that is more of a Rev Ops concept more than a hyper automation concept, which is at the core of that is just what is the point of, what is the purpose of a business. The purpose of a business, I think, and by the way, this is not an easy question. We asked ourselves this question in the pivot. We were actually, do we close shop? What is our criteria for deciding to stay open? And what we landed on was the purpose of a business is to drive lasting change. As much lasting change as possible. And the opposite of a massive impact, but non lasting change. So non lasting impact is like oh my gosh it's a fad. So sequences became the thing, we've spanned the crap out of people and now nobody paying attention to email. That's a fad. Transformation, a lasting impact is more like we've completely changed the way that we think about this problem. Like with Drift, with you guys, and I promise I'm not sponsored by Drift, but I think it's a great company, right? They do a really good job of saying, look, this is how it was before. We need to flip the lens and rethink how we think about this problem. And so the only way to make lasting change is not ACV. It's not number of logos, it's not this, it's not that. It's almost like inception. How do we create an experience that leads this organization to change the way it thinks about the problem? And then you work backwards from there. And it sounds I think a little abstract, but from a product manager perspective, if you think about it, this is all product managers do, this is literally what they do. They look at user behavior. They try to figure out what are they doing today, how do we change that? Is it a nudge? Is it this? Is it AI? I don't know. But how do we get them to change what they're doing? And I think we're just applying that way of thinking to revenue. And I think that's the lens.

Sean Lee: And to your point about what you want the lasting impact of Truly to be, right, how people think about those problems makes total sense to me. I think the other thing that you mentioned briefly, and I think this probably ties back more to the hyper automation side, is helping people to even know what's possible. I think plenty of ops folks, myself included, did not come to ops from some sort of highly technical skills set. Nobody majored in college. So what have you learned about making that more tangible for people to say, Hey, yeah, that problem you hae, you might not even know this, you don't know what you don't know, but there is a different more technical way to solve this that actually also solves your business problem.

Erol Tocker: Yeah, that's an awesome question. So I think that there's kind of two things happening. So there's my thinking, but realistically there's two things happening in the market that are impacting at the same time. The first one is if you look at Rev Ops, just as the economy's getting worse, there's a lot more money and resources going into it. People are seeing as more strategic and there's just more higher level of talent coming in. It used to be kind of almost an IT function, the ops functions. And I think you'd get these real gem people who do amazing things, but still in really an IT kind of view. Now these people are VPs. They work directly with, they often report directly to the CEO. So that's one thing, like actually you're getting people who used to be product people or McKinzie consultants or whatever coming in or bringing that experience. I think the second part is just a much more bottoms up motion. It's just community driven and I think that's about meeting people where they are. One thing we really try to do today is to not, like people have to be ready for change. That's really, really important. If not your top, if you're in a company that isn't ready for change, what good is it me telling you about all these technologies and things if that's not what's going to make you happy and get you promoted and save your job. So we've just been really focusing on social to drive awareness. And one thing that's been really cool is just three weeks ago I put a post on LinkedIn and I was like, Hey, I'm starting a new Slack community called Automation Heads. It's literally for people, we are not going to talk about people in process. We're only going to talk about technology. And it's just a safe space for people to ask outlandish questions and pitch people on outlandish ideas of like, I'm going to use GPT3 to do this, or what do you think of Elon Musk's robot and what's that going to do? Just like that. And we got 100 people on social signing up for this thing.

Sean Lee: That's awesome.

Erol Tocker: And that really surprised me. And I think that shows that there is appetite, but it's appetite that wasn't there two years ago. And so I think it's just going to pick up from here.

Sean Lee: Yeah, I agree with you and I think the community based part of it, to show hopefully one small example, this is where people are going to find the answers. If I'm looking for a new answer to a problem, yeah, maybe I'm going to Google some stuff first, but then the second place I'm going is to all the different ops Slack communities that I'm a part of and I get an answer 100% of the time. It never has failed me. And so from that perspective, it's a very reliable source of intel to help me to do my job. Especially in a role that you don't usually have a lot of peers. And so the folks who can take advantage of these types of communities and these types of resources the best, chances are, are the ones that are going to do the best in their roles internally because they don't have five or 10 people to they're left and they're right that do the same job of them that they can bounce ideas off of. Whether it's from your Slack community or just something you guys have been working on internally recently, for the operators who are listening, give me one workflow automation, whatever, that people might not know is possible that you guys have recently been working on that might help people in their gigs.

Erol Tocker: So we've been working on this for six months. This may be blasphemous, but we sort of set out a project to basically say what would it take to not have SDRs who don't do anything but talk to people. By the way, we don't have SDRs. We used to have SDRs and we used to have sales people outbounding. And we basically going back to our new metric, which is we only want to hire people who talk to people and help them in creative ways. It was like, well, how do we do that? And as of a couple weeks ago, it works pretty well. So it will take the workflows and it will take our old SDR playbook and we give it to it and it will go on LinkedIn and a whole bunch of different data sources, pull the data together, figure out based on our criteria, who the right people are in the account, analyze what they talk about online to try to map that to interest as a first guess in terms of what's going out. And then guess and validate the email down to the level of if you weird stuff, how would you validate an email if you didn't know? It's like copy paste the into Gmail and see if you get the little avatar. Right? So that has been really life changing.

Sean Lee: That sounds it.

Erol Tocker: And it's life changing, not because that stuff sucks or is annoying or expensive, it's just life changing knowing why certain campaigns are working and not working. That's the thing. Before... Have you heard of sprint testing as a concept?

Sean Lee: Yeah.

Erol Tocker: Right. So it's like sprint testing, it's like how do use sprint test with ads? Sprint testing is easy because it's the same ad and exactly where it's being served to who And you have that. With email, you don't really have that because there's so much manual stuff going on and you don't really know if there's deviation. So our current project is trying to optimize our email funnel with bursts of only 30 emails. So what we're trying to do is with no more than 30 emails, get to a point where we've optimized our emails on one sort of domain where we can make all kinds of mistakes and then take that to a high value, an untouched domain. And interestingly, the goal of the experiment is to turn off all tracking on the real scale thing. So it's like do lots of measured experiments small and then get to a level of confidence where you're so confident in what you've done because the underlying data's good where you're like, I don't need to measure opens anymore. I'm just going to send it. And there's a little trick that you know what we're going to be posting. We have a content series on LinkedIn. We're going to post there on how to track clicks without having click tracking. So we're going to rely on meetings booked and actual content read as the metric of mass outbound rather than opens and clicks and other stuff that turns out that doesn't really say anything anymore.

Sean Lee: We're definitely going to have to have Erol back on the show to hear the results of his SDR experiment. But before we go, at the end of each show, we're going to ask each guest the same lightning round of questions. Ready? Here we go. Best book you've read in the last six months.

Erol Tocker: Oh my god. Breakthrough Advertising.

Sean Lee: What is that?

Erol Tocker: So if you go back to original copywriters, David Ogilvy. Breakthrough Advertising is a book by one of those greats, Eugene Schwartz. And basically he breaks down, he's basically like here's the mathematical theory of markets and how different products and different parts of maturity all boil down to certain playbooks. And the guy is just a genius. His book legally costs$ 500 to acquire. It's a very expensive book. But if you're very skilled, you can find it online.

Sean Lee: Nice. All right, cool. I'm going to put you, I think, safely in the ops category here. So these next two will be just fine. Favorite part about working in ops.

Erol Tocker: Solving problems that involve people. There're never... It's not one plus one equals two. There's usually a part of human psychology that I enjoy very much.

Sean Lee: Least favorite part about working in ops.

Erol Tocker: The people.

Sean Lee: I laid that one up there for you. Someone who impacted you getting to the job you have today?

Erol Tocker: Well, I mean, I was really lucky to be part of a team just right out of college. I got lucky and I picked a team that got acquired by Google. A lot of those people went on to do really great things. So it was the hardest job I probably will still ever have, even compared to an entrepreneur. And I just do everything I can to replicate that here. Just a group of people who do nothing but push each other to do things that they didn't know they could do.

Sean Lee: That's awesome. Last one. One piece of advice for people who want to have your job someday.

Erol Tocker: Don't start too early. I think 24 was too early to start a company. I think probably like 28 is just the time when you've just had enough runway. When I started this, I didn't know what an MSA was. So if you don't know what the MSA is, don't start a company. How about that? That'll be my advice. At least in the B2B space.

Sean Lee: Yeah.

Erol Tocker: So, yeah. Don't start too early.

Sean Lee: Do you think that there was something about extra time that better equipped you to be the founder you are now?

Erol Tocker: I think, Peter Teil says this in his book. Every startup is different and there'll never be another one. So there's no point being scientific about it. I think there's a lot of survivor bias. Of course, like my story, the core, the crux of my story is business one didn't work and then we got lucky enough to be profitable enough to take a step back and ask ourselves questions that most people can't ask, either because they've got bankrupt or because they've scaled so fast that every problem is somebody else's problem. I think that's my fortune. But I wish you better fortune. I hope your first idea works better than mine did.

Sean Lee: Thank you so much to Erol for joining us on this week's episode of Operations. If you liked what you heard, make sure you're subscribed to our show. A new episode comes out every other Friday. Also, this is our 90th episode, so if you've missed any of the other 89, now's a good chance to go back and listen to some of our earlier episodes. Also, if you learn something today, make sure you leave us a review on Apple Podcast or wherever you get your podcasts. Six star reviews only. All right. That's going to do it for me. Thanks so much for listening. We'll see you next time.