May 6, 2016
Streamlining Distributed Stream Processing with SuperChief
Ray discusses the experience of moving from Apache Storm to SuperChief. He talks about some of the challenges Librato faced with Storm and t...
All right, thanks for your time. The title is "Zero to 100," which is catchy, but really we're going to focus on the first 10 to 20 million ARR. But just in brief, we're really lucky to have Nicolas [Dessaigne] here, and I'll share why in a minute.
Those of you who might know of me from the Web, I am a VC investor now. Before that I was a co-founder and CEO twice. The last company was a company called EchoSign which was acquired for low nine figures by Adobe, July 15, 2011, at 1:06 p.m.
Then that and the rest of the business, I ran the Adobe Document business as about 100-and-some-odd-million dollar business when I left Adobe. Basically I made every single mistake you could possibly make.
At first I started, answering the questions on Quora. "How much will customers pay on a credit card? How do you do a great sales comp plan?"
I just started to do that and it took off. Then I wrote a blog. Then we got about a million views a month on this blog. I just found it very fun to share these learnings, and then what I really learned after sharing these learnings, and I should have known this with hindsight is, it turns out, everyone makes the same mistakes.
It turns out that, basically, if you are selling a product if it's not self-service, everything is sold and marketed the same for a given ACV. So products that are $4,000 a year, $40,000 a year, $400,000 a year, they're sort of kind of all sold the same way from a process standpoint.
If you take the product out of everything and just focus on the science of selling recurring products, it turns out these lessons work all across the spectrum.
I have had a lot of fun boiling down to my top 10 mistakes, many of which revolve around screwing up my first VP of Sales, which is a mistake about 70 percent of us make, and none of you should make it, because it's fatal. Nicolas isn't going to make it.
And then, especially since we're at Heavybit, I invited Nicolas to get here. He's all developer-centric. I was fortunate enough to be the lone institutional investor after Nicolas came out of Y Combinator like 10 months ago, or something like that.
And then they were almost at zero in revenue. They basically had almost no revenue. Then from there, two million in 12 months. But before that, maybe a year, a year and a half of getting to that first 10,000 a month in revenue. Quite a bit of change.
One and half years until boom.To see that kind of traction in 12 months is interesting because ultimately, I think, what defines an outlier, hypergrowth, a unicorn, whatever trite term you want to use, but it's really going from one to 10 million in six quarters or less.
I don't know whether Algolia will do it, but the early metrics kind of suggest it is, and what's fun is my ten mistakes, I think Nicolas has made fewer than 10. He's made many of them. But I thought that we could share like our mistakes and learnings on those journeys so you can make less of them.
So the first one, and I know a bunch of us in the room have made it, and then I'll ask Nicolas to share a story.
The first, the biggest sales mistake that first-time founders make is not hiring two to three sales reps.
The biggest mistake, right? You're so scared, maybe you've never really done sales before. I sort've done enterprise sales before, but I've never done high-velocity, inside sales, and you want that first one to be right.
You're nervous about money, and you don't know what you are doing. It's the worst mistake you can makebecause you don't do an A/B test. So, best case scenario, and we went through this with Rainforest four times until we got it right. We did an A/B/C/D test, right? And the D finally worked, but Rainforest never ran an A/B. And I'm going to ask Nicolas to speak next.
But best case if the first rep you hit knocks it out of the park. And the first rep I hired did knock it out of the park. He made almost a million dollars at Adobe's last year in W-2 comp. You know what the problem was? I didn't know why. Is he a magician? Does he have a special affinity for the product?
The first guy I hired was literally an actor. He was a thespian. So maybe it was his oratory skills. He was the greatest phone-talker in the world. And I'll actually tell you, two through nine didn't sell anything. And I'll tell you why next.
But A/B Test? I had no idea why. What was your experience?
Actually, maybe complementary to the A/B Test, we didn't do an A/B test. The first account exec we hired that was last summer, like summer 14. We didn't believe we had enough inbound to justify hiring several people.
Yeah, didn't seem like he had enough leads.
Yeah, that's right, let's first hire one. Then, that was back in Paris. The notice of people there is pretty long, so we finally waited for her to join a few months. She joined. She was not as great as we were expecting. We had to let her go one month later. And we were back to step zero.
We lost four months. Just because of this mistake. Had we A/B tested, of course, that wouldn't have happened.
We had the money in the bank to A/B Test. It was not an issue. And also, of course, the business continued to grow. By the time we let her go, we should have had two account execs anyway. It's always about always being hiring. Never stopping. Always having people in the pipe, ready to join.
Definitely a mistake there.
Yeah, I mean, certainly always be hiring reps as you soon as you have anything. So same mistake, different reason, same mistake. No A/B Test, no knowledge. And it will often seem to be getting like you don't have enough leads, but by the time you scale that first rep, you'll almost certainly have enough leads maybe for just two. I don't know.
Morgan, you disagree?
There's always going to be enough for two.
There's always enough for two.
Another good reason why one is just a terrible, terrible,
terrible mistake that everyone seems to naturally make
when they haven't done it before.
And then kind of a natural cadence just to tie it together. I know to some of you it may seem obvious, but it isn't. The optimal thing, really ultimately, is something kind of like this, which is in the beginning. Usually the founder, the CEO, is the VP of Sales. You have to be, you're the one out there selling it.
Usually the first couple customers, they can come from different places. They can come from PR. Maybe it comes from standing on stage at Heavybit. But it usually comes from hustle by the CEO. Then force yourself to get two people that can deliver.
They can hit some amount of revenue. Maybe they only pay for themselves. But paying for themselves is dramatically better than not paying for themselves. As soon as you have two that are good, you're ready for VP of Sales, because you're screwing around, because you don't know how to do it.
Nicolas and I were just having this conversation at lunch. Algolia is killing it, and how to sequence all of this hiring. But at the growth that Algolia is seeing, double-digit millions growing in the teens month over month, you're clearly ready for a big VP of Sales because you're going to need a ton of headcount to make this.
My real point, and we can spend the entire session talking about it, is what you realize in the compounding way that recurring revenue works is, if you wait after those two reps to hire the VP of Sales, you're losing time. You won't grow as quickly. You won't get the good hires in.
You won't get as much revenue per lead out, and you'll just lose all that time of that compounding revenue. The great thing about recurring revenue, which probably most of your models are, it recurs. It takes a lot of time to get that going. And if you can tilt that curve up earlier, it has dramatic effects in the outlying years.
So don't hire the VP of Sales too early. Get two reps working. But in a perfect world, hire your VP, or head, or whatever it is, of sales, six seconds after you have two reps that can hit quota.
Otherwise, you're just wasting time. So, I screwed that up.
In the beginning, whatever you do, don't hire a rep you wouldn't buy from.
If you haven't done sales before, and probably most people in this room or watching on the Web haven't, it can seem confusing. What does a salesperson do? They're weird people. They bang their fists on the table, they're boiler room, they're cut from a different cloth.
And so people get really biased on a great resume. I want someone out of wherever, Heroku, Salesforce, Atlassian, pick whatever you want. And you were like, "I need that guy, and he's magically going to come into my company and make me millions."
First of all, forget about that we all over-weight domain expertise. We all make that mistake. You're gonna screw some of this stuff up. But the most important thing, in the beginning, is you don't hire a rep you wouldn't buy from.
That first guy I hired, the actor, I would buy from him. The next guys who my first VP of Sales, who was a disaster, hired, I wouldn't have bought from any of those next nine guys. And eventually, your sales team may look like a boiler room, OK?
And it may be that way, and it doesn't matter if you'd buy from those guys, but the problem in the beginning is that leads are precious. I don't think anyone in this room has more leads than they know what do with, and if they do, they're probably not very well qualified. You have more leads than you know what to do with?
We're often biased in favor of the resume or the smooth talk or what they look like, or more importantly, what they say they've done. "Oh, I hit 140% of quota."
I think, later, it takes a village, and it takes all different types of people. There are different ways to close deals. There are people who are natural born closers. There are people who eat nails for lunch.There are people who are great in the middle of the process. There are people that are very technical. There's people that are less technial.
Ultimately, there are different ways to scale the peak later, and a great VP of Sales that knows how to hire 20, 30, 50 reps will actually have a relatively diverse group of reps, different ways to close.
But in the beginning, when it is you, and you don't know how to do it, I don't think you want anybody that you wouldn't be comfortable giving this precious lead to. Otherwise, you're doing the job, only it's worse because you have to oversee everything they do.
The worse thing, whenever I see a sales rep leave early, like the Algolia example, or some of the early folks at Rainforest, I always ask, "Would you buy from them?" And the answer is almost always no.
Probably we didn't test her enough or so. Like the first rep, I remember we were very happy with the way she interacted with us in the beginning, but it was wrong because we didn't dive in it, and she was just very good at saying what we were expecting her to say.
But she was doing the same thing with the customers. And that's not good, because that was not the whole product. We need to intune, especially when we are selling to technical people.
You need to be very authentic about the product, and what it is able to do, and so on.
And so in the end, you need to be able to give them your trust. You can not monitor them all the time. You need to trust them to do the best for the company. You need to be able to hire from them, but to trust them.
I think as odd as it sounds, once you truly have a brand, maybe it's not as important for the sales team to be authentic. They just need to know how to close. But before you have a brand, they're looking to every person as an ambassador of a multi-year journey with this product.
It's not the product just that I'm buying today, it's the product I'm going to be buying next year and the year after that, and all the improvements, and the journey I'm going on. If they don't trust the sales rep, it is not going to work.
And more importantly, you as a founder, if you don't trust the rep with your lead, it's going to be a totally dilutive and distractive use of your time. So, anyhow, hard learning. I think we've both been through it. Rainforest has been through it, others have been through it.
I think we'll move onto the next point to accentuate what really happens if you haven't done sales before. There is so much good stuff on the Internet today, someone can come in and talk about ARR, and quotas, and territories and all the amazing metrics they hit, and you hear this a lot from marketing candidates, too.
They can't always do it. But at least if they can talk the talk, and they have at least 18 months of experience selling something vaguely similar, and you would buy from them, then you have a chance. Because it's just you.
And the third one, Algolia is a fun case study here because, actually, Nicolas did much better than I did on this. I don't know if anyone went to the SaaStr annual conference I had, but Aaron Levy and I talked about how we both made this mistake.
This is not seeing the pattern early enough. And what I've learned, I wish I'd known this myself, I would've done far better. What I've learned as probably as early as 500K ARR, 40K a month in MRR, you'll actually have a pattern.
You'll have some amount of customers, 50 customers, 40 customers, I don't know. A hundred. And you start to get a pattern. If nothing else, you'll probably get a revenue pattern. And usually I call this small, medium, or large.
I don't know what the sizes of small, and medium, and large are, and I don't know what your ratios are, but you'll usually get a group. I was talking with this a little bit with Iron.io guys, before this, they saw this similar. Nicolas saw this early. And you have to understand that pattern.
Instead, one thing you can try and do is change it. Change the way it works. "I just really want to be a self-service product. I hate sales, so I want to focus for us on 'pro' and all the small customers." But if you do that, you'd never get anywhere.
If you go back to Box, Box started all at the bottom. And now freemium was less than one percent of their revenue at 250 million revenue. So you've got to find this pattern. And what I've learned is this pattern, unless you change something dramatically with epic amounts of marketing or sales dollars, you're going to have roughly the same pie chart ratio at 1 million as you do at 10 million.
Now those customers may get bigger. I'm guessing at 10 million. Nicolas has their first six-figure customers. And maybe at 10 million, they can close a 500K customer, right? Knock on wood, some right here.
But if we just redefine what "small, medium, and large" means, I bet the ratios are going to be similar. So the real lesson is, force rank this stuff and double down on your time here. If your small is a huge segment, then do all that and stop screwing around, talking about big deals all day if you don't have any, right?
If you're like me and your big deals, which back then were probably $1,000 a month, but our big deals, what we call enterprise, is a majority of your pie, spend a commensurate amount of time there.
And Nicolas went from a $19-a-month product to a product that's hundreds of thousands of dollars a year by doing, if nothing else, not making mistake number three.
At the same time, we kept the two good markets. So we kept the long-tail developers as customers, and then went enterprise. For us it's two completely different, good markets that are complementary, and for a long time we struggled, "OK, what should we do? Long-tail or enterprise?" And we finally decided to do both, but with different techniques, different tactics, different approaches.
The long-tail would provide awareness, because the developers, they speak among themselves. They may become our customers with their company. And then enterprises, they provides credibility. It's not completely black and white, sometimes you can do several things, but you are right in the sense that what we discovered is it could work in a specific kind of customer. The size, the market, we would double down on this one. And we still have a lot of improvement to do.
But to some extent, though, I think what you've done is with the long-tail with the small-tail. They're marketing.
From a revenue standpoint, the bet is on the bigger customer. That's the learning.
Today we are like about 60% of revenue, enterprise customers, and only 5% of all customers.
Let's step back. Let's dig in because I think it will be interesting to folks. We should have done this at the beginning. I screwed up. So, tell us in 90 seconds what Algolia does. But then let's compare how the two segments use the product, because it's interesting.
- Sure. So Algolia, we are a hosted search API, so we basically help developers to deliver a great search inside their websites, mobile apps, everywhere. You are all used to Google-like experience. We want our customers to provide that inside inside their own websites and mobile apps.
You've probably used us maybe on the Hacker News, Product Hunt, Altispring a lot, Vivo. I'm stressing the local customers.
Yeah, interesting to this, so Algolia starts off as a $19 mobile SDK, right? Becomes a web API. And we may have our own views of it. You're the CEO and founder, I'm just a bystander. The value that these customers get is different.
And so, I would argue the big reason some of your customers are paying you six figures now is because you can actually replace developers building search with something that's not even 100 times faster in some cases than Elastic.
But actually, I don't need people. A solution versus a tool.
I don't think we are yet there. Early, enterprise customers, they probably don't choose us for the price to get to pay less than developers. But the thing is, it's quite difficult to hire these people anyway. Creating a good engineering team, able to build a search internally, is crazy difficult.
At the same time we provide them a product that enables them to do things that are actually better than what they would be able to do themselves. And we can prove that. The thing is, for these big customers, part of who are SaaS process, is actually to demo the product.
So we would get our data, build a demonstration, usually in a day or two, and send them the demonstration that is already 10 times, 100 times better than what they may have spent six months, one year, to make nternally. And that's really what make us win the deal. So it's not really the money they would spare on developers.
They may not have the developers. They don't have the capacity to build.
And if they have them, they may have something better to do with their time. Then it depends on customers of course. Some like big web companies, say, Twitter, Facebook. They have huge developer teams, so that's competitive integration.
Mistake number three is especially, we'll go on to the next point, but especially if you're biased. If you don't like sales people, if you don't like sales, if you want to sit in front of your computer all day, oftentimes we're too biased towards the end of the market.
And then one point out here that I've learned is related to it is the pie. The second one is not going upmarket faster. The simple fact of the matter is, it's basically the same amount of work to close any customer at the end of the day. So if you can close a customer for $100,000 a year, it's better than a customer that is $10,000 a year.
In some cases it's less work, the bigger the customer gets, because they often have people to do the implementation, and they often know how to ask the right questions.
So don't go upmarket if you can't. But go upmarket faster.
I think that it sounds great, but here is, if I can boil it down to one mistake I made. And in fact, one of my co-founders quit over this point, he walked out the door. When we closed our first six-figure customer and he said, "Well, we'll never get another one," and "I don't want to do this."
What I learned is any customer you can get one of, you can get 10 of. No one has heard of any of us, in this room, has anyone ever heard of any company that's here? I mean, not really.
So if you can just get one, do you really think there aren't another nine somewhere on this entire planet, Earth, to buy your product. And everyone gets this.
"OK, I'm doing OK. My product's $5,000 a year, and then Google comes in and wants to pay me $300,000 a year." We don't want to do it. It's distracting, it's dilutive. If they really want you to build something that has nothing to do with your product, it's distracting and dilutive.
If they want you to build something that's on your 2017 product roadmap today, and they're willing to pay you a half million dollars a year, and you already were planning to build it someday, I say fracking build it now!
You're just learning that the market will pay you if you reorganize your product road map, right? Because one is 10, right? It took us 18 months to close Google. Then it took us six months to close Facebook. And then 30 days to close Twitter.
So the net, net one, and I want to hear Nicolas's story, is I don't believe there are anomalies in customers. There are anomalies in many things. If you get one, if you can reproduce that code base and those features. And it's a big customer, bet on it.
And now as a VC, one of the first questions I ask when I meet with startups, I probably asked Nicolas in the very beginning, "what's your biggest customer?"
"Well, my average customer is $1122 a year, but I have one that's $20,000 a year." Then I know you can get 10, for sure. And I know you can pretty much get 100. And I'll just build my own model. Because I know this. But you think it's one, since I've lived it, I know that you can get 10.
Did you have an anomaly that you had to fight over, or anything? Or maybe this was one you had to struggle with?
Probably one of the mistakes we haven't done. We went upmarket. The thing is you need to be, iterating. You cannot go from zero to 100 in one day. If they come inbound, of course you want, sure. But if you are like struggling to reach out to customers, just pick the ones that is slightly bigger than your biggest one, and continue like that.
And quickly enough, you will realize that you have customers are already 10 times bigger than the ones you had before. And it's really working. As you get new customers, it becomes so much easier to get a bigger one, and bigger one, and bigger one.
And then the last thing that I'll say on this, and I wish I'd known this. This one customer, this Google that pays you 300K or whatever this big one is, that's not really off your roadmap. It's just a pain in the neck, and a lot of people don't want to do it. It's not just their not an anomaly, they're also the future. It's a gift.
And whenever anyone kind of half complains to me about this scenario in my startup, "Well, you know, we're only doing $200,000 a year, and we got this $200,000 customer, but it's so much work." I say this is the biggest gift in the world, right. Because they'll teach you how to get 10.
You don't know how to get one. But when you get one, you'll know what to do to get ten, and it's like the cheapest, it's paid education. It's being paid, it's paid college. So do this one, it's the future, right?
OK, fifth one, and Nicolas and I were just talking about it. Just working with Algolia, I didn't understand this concept that I call a mini-brand. Obviously we all understand the brands and technology. We know Google and Dropbox, and pick whoever you want. That's product we've all heard of.
You know, in my relatively brief tenure as a corporate vice president at Adobe, all I was allowed to talk about was the brand. Right, just the brand. Adobe, in its certain segments, has over 95% revenue market share, so when you have 95 revenue, all you talk about is the brand.
I wasn't allowed to tweet. I wasn't allowed to blog. But it's not silly. It's because it's the brand. And the brand does work. Like when you have a billion in revenue, the brand works.
What I didn't understand at all, and I didn't see it happening, and I wish I had, because I would have tripled down, was the mini-brand. And when a mini-brand usually happens, it's typically when you maybe get about 100 customers or so, and it generally takes at least 18 months. Maybe it could happen in 12.
But it's when just enough early adopters in your little niche, in little part of your industry that you do well, small object search for tech-centric developers, not the whole world, or whatever it is, third generation VPs of engineering building a Q18 from scratch.
That's kind of a niche for Rainforest, but those people start to hear about you from their peers and their friends, and that's how you, from my case, how you get to Google, to Facebook, to Twitter.
It's not the brand, but it's the acceleration of the mini-brand.
It's usually somewhere between 100 customers and maybe two million in revenue, and it's a phase where it gets easier.
This is why it's worth fighting over-investing
in customer success. And I did this piece,
I did it at SaaStr, I've done it on the blog
at what second timers doing, and all of us
that have done it before.
You know one thing that we're way doing, over-investing in customer success. Way more customer success than sales, because we know this happens. And we know as soon as we get this mission, mini-brand, in our little niche, our little segment, it all starts to build on itself. And you're seeing that a little bit just now, right? A year ago, who knew of Algolia?
Nobody. Well, that's the nice part of the thing, is that when you start to have people recognizing you like at a conference or anywhere, "Oh, I love Algolia" because whatever it is, they saw us. They used us, that also feels good. It's really in terms of motivation, it's great.
And for your new reps, it's even better because they, your new reps, when you hire them, probably they didn't know you before. They start to sell the product and they realize that actually the brand is already known. And it's a big boost in motivation.
And can you see it in the leads yet? Are you starting to see leads from the brand?
It's difficult to define where they come from. Definitely the biggest chunk of our leads come from word of mouth, so it's kind of the brand. But it was also the case one year ago, some website like Product Hunt, Hacker News are helping a lot because that helps us to get our name out in the community we are targeting. So it's progressive, it just happens more and more.
Yep, a corollary of this, what I've learned is, it almost doesn't matter how long it takes you to get to mini-brand. Some of us can get from our first million in 12 months from launch. Some of us can do crazy things. Some of us take years, and years, and years to get there. But it's hard to have that real compounding effect until you have a mini-brand and people start to hear about you.
That's when I start to judge you. When you get to that million and a half or two million in revenue, you start you want to see that growth in the teens, to be great, to be hot. But honestly, I don't even care in some ways how long it takes companies to get there differently. But once you get there, it just gets easier.
So this is your first bit of cavalry. This is when it gets easier. And my huge mistake is once I had a mini-brand around two million, I didn't exploit it. I didn't market the hell out of it. I didn't ask my customers for referrals. I mean we got reviews and stuff. This was years ago in Internet time, but I didn't leverage the fact that your mini-brand will create more customers.
I didn't triple down on customer success. I didn't triple down on asking them for more customers, and so I squandered time given all the hard work I did to build the mini-brand. OK, one one, and I don't think Nicolas can help on this one, thank God, but maybe it will come.
So my, this is probably mistake number one, even though it is six on the list, is right after I got to that mini-brand, we actually had, there was little bit of a recession in '08 or '09. Probably nobody remembers, but things were a little dramatic. VCs like Sequoia sent out this R.I.P. memo.
It wasn't like today where all deals are like at three billion or higher, pre-money Series A. And so a couple of things happened at the same time. I hired my first VP of Sales from a big tech company. The economy went into a tailspin. And, like who knows what happened?
Actually, Salesforce itself didn't hire anybody for a year. So there's a lot of drama going on in the economy. Our revenue didn't flatten, but our growth flattened. We stopped growing even though I had millions of dollars in the bank. I had this fancy VP of Sales from Salesforce. He hired 10 reps.
And what actually happened was our revenue per lead fell in half. OK, so our leads kept going up, which I'll show you. But our revenue per lead fell in half. And this is sort of inexcusable. This is like alarm bells should go up.
I didn't do this analysis on the right until months into this process. And what I realized is my VP of Marketing, who was great, who I hired at 20K MRR, and it wasn't a day too early, which I wrote about, she hit her lead commit number every month. She hit the number of leads that she committed to deliver the sales team every month.
But the revenue per that lead went down 50% 60 days after I hired my VP of Sales. And so if you take the growth in that and the 50% decline in revenue per lead, you can see a kind of flat revenue growth for an entire 12 months.
And so, I can explain to you why. The simple answer is, he wasn't good enough. He wasn't a good fit for our stage of company. And so he hired the wrong reps. He hired reps I wouldn't believe in. If we had more time, I could spend 60 minutes on that.
But here's the short mistake that I made. I should have fired him in 35 to 40 days. Because look at all the time I wasted. And the truth is, when you hire, and even though Morgan's under the gun, he is agreeing with me here.
And so, and it's even worse, because when this happens, not only is it terrible when you have huge arguments. Whose fault is it? Is it the global economy meltdown? What's happening, right? There's no excuse.
And so, I think it might be on the next slide. When I hired my second VP of Sales, my real VP of Sales, we doubled revenue in 90 days. He doubled revenue in, actually it was faster. And I can explain to you why, but you already know the math in the head. I just restored the discompetence that we had before.
He just came in and he did exactly what you could guess. So think about it, how do double sales in 90 days? He really doubled the sales in 60 days. This was a dire conversation. This was like Lehman Brothers, the world is ending.
I hired Brendan and I say, "Good news is you've got the job, I really want you. I'll pay you anything I can afford. Bad news is, if you don't double sales in the first 180 days, we're completely out of business. It's hopeless."
This is just math, this is not my opinion. We have 3.8 million left in the bank. The venture capitalists have gone golfing again like they did in 2001, and we just have to make it work, and he did it in 60 days. Let me tell you why, and then to illustrate the action item. It obviously wasn't new prospects or customers because he didn't have time to generate new leads in 60 days. So obviously it wasn't that.
But what he did is a couple of things. He upgraded the team, So he took those nine people out that the guy hired before that were terrible, and he brought in three pretty good people. He had been the VP of Sales at LinkedIn before, so he brought in three of his good people.
I don't even know that they were the best, but they were good. Did they know what the product did? They had no idea what the product did. Did they know how the API worked? They didn't even know what an API was. But they knew how to talk to a lead already in the system, speak to power, understand a business process change, and close the fracking deal, all right?
So we just took the same body of leads he was handed, brought in three good people, kept two people out of the nine. One didn't make it. He kept the one SDR, who Sam Blond is now the SVP of Zenefits that will do one to 100 million in 24 months, so he was the best guy.
He was only as SDR at the time. But he got rid of all the losers. Get rid of the losers, and what happens to your revenue per lead, it goes up. Because the losers are sucking, you're taking your precious leads, and they're halving the revenue. Just get rid of them.
And there's a few other things that we could chat about. So my next point, again this is a fun topic, here's the real takeaway, and it's a horrible thing to say. Seventy percent, and this is not really statistically validated, but I know it's about right.
Seventy percent of first-time founders, the first VP of Sales doesn't make it 12 months, and they make the bad hire.
And we can talk about how to make the hire right. You can back into it through this. But my real point is, since you may make the mistake, you'll know in one sales cycle.
If they can't come in and take your seven crummy leads or your 58 leads or whatever, and do a better job with them then what's happening before, this is not a VP of Sales. Right, and don't especially don't listen to VCs. Don't listen to board members. If the resume is good, they always want to give this guy more time.
"Give Sarah more time. She came from Netsweep. She came from wherever. She came from New Relic." Don't give her any more time. It's hopeless, right? I mean, Morgan, you saw this. How long did it take to have an impact? Pat yourself on the back. A sales cycle, a sales cycle.
If your sales cycle is a year, you might have to give them 12 months. But it's a sales cycle. And you'll know qualitatively in half a sales cycle, you'll know, right? So anyhow, that was a pretty bad mistake. Don't ever make that one.
Pipeline's for losers. When you're the Chief Revenue Officer at Salesforce, it's 6.8 billion in recurring revenue. Pipeline is critical, right. It's all figured out, it's a process, right.
I will tell you categorically, and now Morgan is going to hate me for saying this, but I will tell you when I know when I made a loser VP of Sales, when in the board meeting or whatever he talks about pipeline. I don't want to hear about pipeline.
I want to hear about how we're going to hit the revenue for this quarter and this month. What's your MRR goal for this month, for this year, for this quarter? How are we going to hit it? I don't care whether you do it with magic, or whether you do it with love, or kumbaya, or whether you bash walnuts with the thing, I just don't care. The great ones just deliver.
And you know what the losers do? They miss their number, but they always have pipeline. Like you're doing like half a million in revenue, and the losers have like seven million dollars in pipeline. They always do. And it's always more than an order of magnitude, right? And they'll never close any of it, because they're losers. Because they know.
So, I'm not saying you don't internally need to kind of have a pipeline, of course you do, and it's nice to have a dashboard, but when a VP of Sales presents that as their core metric, you know it's just hopeless, right?
And this one I screwed up eight ways till Sunday, but actually people get this one even worse now. Since Nicolas didn't make the VP of Sales, I want to take it, this one is too much time on prospects versus existing customers. And let's chat about this just for a minute.
There are really only three type of customers. The ones that simply don't grow fast enough. Companies, they eventually die. The ones that grow faster than they understand why they're growing. Those are fun. And then there's the ones that basically are growing as fast as they can possibly, they don't leave anything on the table. They're doing the best they can.
And when you're in the middle, or in the first category, you tend to spend all your time on trying to close that caught deal this month. We got to get whoever it is. We got to get Aetna. We got to get Google. We got to get whoever or we're gonna die, and that's all true.
But what tends to happen is founders, we spend way too much time on these new deals.
And when the truth is, when we talk about mini-brand and all these other things, frankly, it's far more important to make your existing customers happy. And why is that?
It's not just because we kind of intuitively know this at the qualitative level. It turns out, it's true at a quantitative level, because if you look at this chart, it's a little hard to understand, but a lot of the folks now on the customer success space have actually adopted in an attempt to explain it better for me.
But this is a revenue pattern for a recurring revenue customer. And our customers tend to last years, right? Your Algolia customers are going to last like seven to 10 years! You have net negative churn.
By revenue, net negative churn, and all enterprise customers, I define enterprise as just big. They all should have net negative churn by revenue. You're going to lose some by customer count, but they all should buy more and grow, or your product sucks. I mean, there's exceptions, but we have to bifurcate churn.
There's a lot of churn at the low end, but you should have net negative churn at the high end. So what those customers will do is they'll do two things. The part on the right is in year two and year three, this may not exactly happen annually, they'll buy more.
So that 100K deal next year becomes 150, and the year after that becomes 200, but it's better than that. Because not only do they all buy more as a cohort, but then they leave. And they leave from Google and go to Twitter, and you know what they do? They have the same problem.
Not always, but sometimes, and they bring in the same vendor. And they tell their friends. And they go to your customer conference, and they tell your friends, and you get this second order revenue. What happens is, using these types of SaaS metrics, we way underestimate our customer lifetime value. It's usually two to three "X" higher than it looks based on a simple calculation.
And so if you think about it, one, you should have far more customer success people than you think making these folks happy. But just as importantly as CEO or founders, you should be there! Get out of Heavybit. Get off your Mac. Get on a jet, which most people are lying when they say they want to do. Some people, they actually don't mind, like poor Nicolas who spends half the year on different continents.
Most of us don't enjoy it, but you got to go and visit all you best customers. And you've got to visit all of your customers that are your top 20% of your revenue.
And at the SaaStr annual conference that I did a couple months ago, Nicolas was there, I asked all the CEOs, "How much time do you spend with your existing customers?"
Aaron Levy is now more than half. He spends half his time on the road now. Two years ago it was none. It's half. But half of that is, the majority is with existing customers. Not prospects. Is that because he thinks it's more fun? No. It's because he knows about this second order revenue.
We went through all the folks there. David Ulevitch from OpenDNS, which is a fascinating company. Regularly closing million dollar deals. He's always there with his existing customers on a jet. Never went on a jet before four years ago. This shit works. And so stop agonizing over every last deal and get on a plane and visit all your customers.
I'll tell you the last anecdote. It's been four years since the acquisition of my last company EchoSign and we've lost, well the company has done OK, but lost a lot of customers since I left. But we didn't lose a single big customer when I was there. Never. And it wasn't because the product was perfect. It wasn't because we had 100% uptime. It wasn't because of anything.
It was because I freaking went there, and the customers believe in you as a founder, and they believe in your journey. They stand you, they'll complain. They'll threaten to quit. I mean Groupon, every time I met with Groupon, oh my God, "We're gonna leave tomorrow."
But I was always there in Chicago or wherever, and you'll just never lose them. So, anyhow, final corollary, which I only figured this out a few months ago, but it's obvious from the math.
It turns out, the faster you grow, the more time you should spend with your existing customers.
First it seems counterintuitive, but it's true from the math. Because if you're growing so quickly, this body of customers that you have to spend time is almost inherently going to be neglected. The hypergrowth is happening. So the slower you are growing, I guess you have no choice. You've got to be in on every deal.
But the faster you're going, throttle back. Don't get sucked by your VP of Sales or sales rep or whoever it is, and just get on a jet. Maybe enough on that. What are your learnings? I think you spend a lot of time with the customers.
I think I know what I will do tomorrow. Probably not enough. Definitely not enough. We're spending a lot of time with our customers but, as you say, we didn't tell the team fast enough, and probably that we struggled to spend enough time with every single customer. So we are feeling the pain that you are describing.
How many of your biggest customers have you met with personally?
Probably half of them.
Half. Not all of them.
Yeah. You got to go visit the other half. in the next couple weeks.
And the other thing is that, in our company-wide, we are already seeing, even if we are very young, second-order revenue. Not only upsales, but people changing companies or whatever. They use us on the new project. Already today, and the product has been available for one and a half years. It's pretty short.
Our update is happening, so we already know that the customer success is our best marketing. We are really investing a lot on that, but well, we are hiring. I mean, it's very difficult even when we want to hire people there, it's difficult to find the good ones.
And that's probably one of the drawbacks, is that when you don't have enough people, you may concentrate on new customers. We never lost any enterprise customers yet, but these things kind of do happen.
That second order, when they leave and go to the other customer. When you as a founder have visited them, the odds of that go up dramatically.
Because again, they're not buying the product today, they're buying a multi-year commitment, and if they meet the CEO, that's the journey they want to be on.
That's why Aaron Levy is spending more time on the road now. After 10 years and 250 million that he had to before. I'll tell you lastly, we're running out of time, but you think it's just that I remember.
So we talked a lot about Groupon, which was probably my biggest customer that complained the most and threatened to cancel every 30 days. But I remember the first onsite we did at Groupon, right when Groupon was exploding. It had gone from like zero to 2000 reps in 24 months, I mean it was insane.
We were in there in a meeting with all of the champions, and CEO, and hearing all their complaints and all the darts thrown at me. And who's in other the conference room right across there? It's Mark Benioff, right, meeting with Andrew Mason and his staff.
I'm like, "What is Mark Benioff doing?" He just fueled up the jet from Kona. He didn't come from San Francisco, he came from Kona, where he lives a third of the year. I don't know, can you do that non-stop? What kind of Gulfstream do you need, do you know? Anyone like a G-something expert in the room?
But he'd flown from Kona, in the middle of a horrible winter to the worst place to be in winter in Chicago to meet with Groupon after Groupon had already agreed to pay him 30 million dollars a year from Salesforce. Why is he there?
I can't tell. I didn't ask him, but I know it's this reason. I know he knows it's more important for him to spend that time with that customer when it's closed, than it was to try and get Groupon through the door. He has only so much time.
His Salesforce was probably doing 2 billion back then, that he showed up from there. And that was kind of my insight that you gotta spend more time there. And then kind of the corollary for this if you don't do it.
Here's the learning, now we'll move on to the next one, is actually if you don't spend enough time with your big customers, you actually won't lose them until the third year.
So, in the beginning, when everything seems dramatic, and everything is week to week, maybe you don't care. But here's what happens with bigger customers. It usually takes them a little while to make a buying decision. And then even if it's not formal, it's budgeted.
Even if they don't deploy your solution for a while, the next year comes up, and it's already budgeted, so they always renew. I mean ,we had these customers, we had these financial service customers, like New York Stock Exchange, Nasdaq, Dow Jones, who deployed us with Salesforce, and then they would take two years to deploy Salesforce.
So were like, "Oh, we're gonna lose them." But they always renew, because it is budgeted, and the second year comes up. It's usually not until that second year when they start to measure the ROI and whether it really works, that they'll really invest the time to switch vendors, because it's so much work to choose a vendor, to qualify a vendor, to test it, to put it into production.
And so, at the low end of the market, your $10 a month customers, $20 a month, you'll lose them on a credit card. They'll churn out in a month or 90 days or nine months, but really it's three years. So if you're playing the long game, you've got to do this because you won't see enterprise churn for three years, and then it becomes catastrophic if you don't invest in it.
Actually, one exception for us is that these big logos people would go to their website to check if all the searches are working. If the demonstration is not good, it's actually a logo we don't want to put on a website. Of course, sometimes we do anyway. Always with 12 logos.
But that insight for us is very, very important to invest time with them to make sure the demonstration is correct because that will be a showcase for us. Because it's public, because it shows our facing. Usually it depends on the project, but often it is.
I do want to say it again just so I know I said it. You know the flipside is not firing a bad VP of Sales in a sales cycle. Whatever you do, you'll know it. You'll know it in 50 days. I've been through this myself. I've been through it with like 10 other companies. If you don't see numbers better in one sales cycle, just move on, as painful as it is.
And, the one thing I didn't say before, let me emphasize. Here's the other mistake we make. One is, we kind of know it, but all of our advisors and investors say, "Give it some time, because Bob was from Twilio and he was so great."
But the other mistake that we make, because I see this again and again, and I'm like, "Listen, I know Linda is not really working. But I don't have anybody else. Should I keep a warm body in this chair?" I can tell you the answer always is no.
You will find, if you are any good and scrappy, you will do better without Linda, even if you have to manage your sales yourself, even if you don't want to. Even if you have to go back in time. You will do better.
And, if nothing else, it will torture you to find her replacement. I guarantee you this. So whatever you do, and I went through this, when the time finally came, and I was way too late to fire that first VP of Sales, I changed my mind. I'm like, "Oh my God. There's these 12 guys, and they're terrible. I don't know if there was three, I didn't know what to do with 12 people.
Luckily he quit the next day, and Brendan reached out to me on Facebook, one day after that, so thank the Lord for that. Although it did take me 18 months to recruit him, which is a different topic. But I should have just dealt with it. I should have fired him, and actually just fired all the reps, and had it be me and three guys. We would have done better.
So whatever you do, don't go into that "Linda's not good but I'm going to keep her anyway because I don't have anybody." Just suck it up and fire her if it doesn't make a difference in one sales cycle.
Then just a couple thoughts to wrap up, and then we can do questions. Two kind of higher level reminders. First of all, there's two stages. I know this stuff is hard. Recurring revenue is hard because it's so hard to get that ball rolling.
If your product is $100 a month, and you close 100 customers, man, that's only 10,000 bucks a month. How much does it cost to have a developer at Heavybit full time? It's probably more than $10,000 a month, I think. Just being a little facetious.
It is so much work to get that ball rolling, but I wish I had known there were two inflection points where it gets easier. And it actually gets hard in between them. The first one is when you get this mini-brand. When you get to this million and a half, and two million in revenue.
The problem is you don't have enough people, and you're not spending enough time with your customers, but it does get a little bit qualitatively easier when people have heard of you, because you get air cover. You go into a deal, and you don't have to explain yourself completely from nothing. You have proof points, you have logo customers, you have all this, and it does get qualitatively easier even if the number isn't easier.
And then the second one is once you get to that 2 million, here's what I wish I would have known. Do whatever you can to get to 10 million as fast as you can. And the reason is it gets easier. Because until you're at about 10 million, unless you're like massively funded, until you're 10 million, it's rare you have fat on the model.
You don't have a whole extra sales team to go experiment on a new product line. You don't have four extra engineers to throw at a product extension. It's even worse. If someone leaves, if your VP of Engineering leaves, your Director of Sales, it's fatal. It's horrible, you have no redundancy. But once you get to 10 million, you will have redundancy in your system one way or the other.
And it's just because of that, I can't tell you how much easier life got around 10 million revenue for me. I wish I had known that, because it seemed like this recurring revenue is also a pie-eating contest. The dials go back to zero at the start of every month, and you've got to roll that ball out.
But as founders, it does get easier at two and 10. It gets easier at two and 10. But it's hard in-between, because there's so much going on between two and 10, and you don't have enough people. So try and just skip all that time.
I'm being a little facetious, but what I mean is, raise a few extra bucks. Hire those extra people. Do that extra, get on that jet one extra time because if you can get to two to 10 a quarter earlier, six months earlier, your life will be geometrically better, and I wish I had understood this, because it does get better.
And secondly, we're all among friends here, right? Bear in mind, as you go up that hill to 10 million, you'll get a lot better, right? Everyone gets better, right? We all get better as CEOs. You might not know how to sell, you might be a mediocre engineer, you might not be particularly good at any of these things, but as founders you'll get profoundly better at selling your product, managing your company, hiring people.
It's rare that you meet, I don't think I've ever met a company at 10 million ARR or growing 10 percent more a month. They may have a kooky CEO or something, but they're great! They're great if they're founders. It makes you great, right? If you saw Drew Houston in the press the other day, they were quoting him early, "I would have sold Dropbox if someone had given me a million dollars." He's come a long way.
Right, all of us do. Nicolas, you probably feel a slight bit better as a CEO, right? They all get better, so you'll get better.
I guess my point is, we all feel a little bit inadequate in certain areas. Just sweat it, bring the best talent in under you as quickly as possible, but you will get better.
So this is another way it naturally gets better. I wish I'd known that.
And then just one last point. This is more of a strategic mistake that I made. I sort of knew it. It was crystallized, and if anyone was at the SaaStr annual, when I had Parker and Sam up on stage and they told the story of when the plan for last year was to go one to 10 million in 12 months, and I was actually with Sam when Zenefits was doing $30,000 a month, and we were planning what the organization was going to look like at 10 million, and how I was going to ask Parker for twice as many sales reps than was in the budget, and all of it seemed very ambitious.
But what they did was, five months into the year, they took a break and they said, "Well, how can we get to 20?" And it was done in a protective way. No one was going to get fired, but they doubled the plan. Almost none of us are going to go from one to 20 in 12 months. But once I hired Brendan, that real, second VP of Sales, we doubled my revenue per lead and doubled everything.
We made the plan every quarter in every year. We beat the plan, not every month, but every quarter and every year. And that was great. But once I had those resources, what I should have done was double down. I should have not increased at 10% or 20%. I should have doubled, and maybe not literally doubled the plan, but I should have taken a great team and challenged them to do something even better.
And you can't do that until you have a real team. You need a real management team together, where you can really do that. But I think this is one of the greatest thought exercises you can do once you're even at a million or two million in revenue. But you need a management team.
How can we double the plan?
How can we grow, maybe double, we can call 50%.
How can we grow 50% faster than plan this year?
No one is going to get fired.
And how do you do that?
And one way to do it, if you don't know,
is what I call "imagine a world" exercise.
Get together with your team and imagine a world where capital doesn't matter. OK, it is 2015, and you know once you get to that Series C round of 700 million, maybe capital doesn't matter. But it's really more of an exercise. Imagine a world and take every leader on your team: sales, marketing, engineering, product, customer success, and imagine money didn't matter.
You could hire anybody you want, but only one hitch. It as to be ROI positive. You can hire 100 engineers if you want, they just have to build enough features to pay for themselves, measured over, say, 36 months. Not our typical ROI of 6 months, but the features have to eventually.
But you can hire as many sales reps as you want as long as they at least cover their cost. They don't even have to be profitable. You can do any marketing program you want as long as it returns one to one in 24 months.
What you'll find is, if you hire great people, for this exercise, they won't overspend. A great VP of Marketing will never tell you, "I'll take 50 million," because they know they can't process that money.
In fact, what they'll do, let me see if it's in the next slide, they'll actually come up with an amazing plan for you, because they'll only spend on what works.
The great thing is, know this, they've done it before. They know they're not going to hire 50 mediocre reps in Tunisia to try and sell your $500,000-a-year business process solution. They know what they are doing. But you give them a little bit of extra fuel and let their imagination go. They come up with amazing ideas.
I try to do this exercise, I've only been doing this part of it for so long. I try to do it after the start of every year, after people have a break, after the planning process. And you need a management team. But do this, and then if your team can tell you a way to grow 50% faster if cash didn't matter, then, as founders, as CEOs, your job is to go make it happen.
Whether it is to raise more money, whether it is just to take more risks, whether it is to let them spend and hope that they do it, I know it may sound scary, but if they believe they can do this at some crazy level that's even vaguely creative, let them do it.
Because the advantages today of growing 50% faster, not only how that compounds from an ARR level, how it works for evaluations and the value you're billing the shareholders, it can be an order of magnitude more valuable.
Do this exercise, and you can watch this video on stage on SaaStr.com with Zenefits, but if you have great people, it seems to kind of always work. And as CEOs and founders, we always get more growth out of it. Our job is just to fund it. And it doesn't always require more money, right?
You've done that yet this year? Probably don't have the full team built out to do this, right? You and Julien have already done it? You've already raised the plan. No, not really.
We were having a dinner discussion earlier about should sales and engineering and product sit together, and some other topics. What I'll tell you is, it was kind of a small point on this large one. Let me just tell you as you're looking to two million and 10 million, these next phases, when it's good, it is really good.
And a couple corollaries of that. What you'll find is that great teams have basically no turnover, at least from up until like 10 or 20 million. And we probably sort of kind of understand this in engineering, but let me just tell you, it is true in sales, too. The best sales team actually had no turnover once they're good.
If I'm a sales rep, and I'm working for a great VP of Sales, and I'm making a huge amount of money, and I'm growing my career, and I'm getting responsibility, and I have a chance to get promoted, why would I leave for another company? I wouldn't. Great sales teams stay together forever. And these great teams feed on themselves. And they energize the rest of the company.
If you have a great sales team that'll feed on the great engineering team and vice versa. So, the net, net learning from all of this is whatever you do, spend 20% of your fracking time recruiting.
Always. Force yourself. Did you do 20%? Close enough, right? Could be more. Especially with your bi-continental thing. You're spending a huge amount of time recruiting. And I know that we all say that we do this 20%, but I think we lie a little bit. I think we say we're a little thinner, maybe a little taller than we actually are. Twenty percent.
I think that there is nothing more important once you hit a million in revenue for sure, and probably a half million, than recruiting your management team. There's nothing more important.
Stop screwing around hiring individual reps. Don't even try to hire individual engineers or developers. You need leads, you need managers, you need a team. Spend all of your time recruiting them. I did a terrible job recruiting my second VP of Sales because I got him passively, but it took me 18 months to get him.
Fortunately, it was less than one week after the first guy quit, screwed up, so again, praise the stars for that. But I should've, if I had spent 20% of my time recruiting, I probably could have gotten him earlier.
The single best VP of Engineering I tried to recruit five times. It was always the wrong time. But if I had spent all of my time recruiting him, I would have closed, no, I think I would have had twice as much revenue. So force yourself. Get on a jet, spend 25% of your time with your existing customers, and spend 25% of your time recruiting once you have even half a million in revenue and drop the other crap. Drop it. Whatever it is drop it.
Maybe you're going to lose the deal this month. Maybe some feature is gonna slip. But here is the thing. If your product is sellable at 500K in ARR, it's sellable at 5 million. All you need, you actually don't need another feature.
A good friend of mine who works with me now was SVP of product at Salesforce for many years as it went up the scale, and he's a product guy. And Nicolas met with all of his companies in January. He's like, "Do you need to build a single feature to hit the plan this year?" And the answer is always "no."
I mean, it's yes, but if you ask people honestly, like, "Do we actually need a feature?" Once you hit a million in revenue, it's always sort of "no." I'm not saying drop a feature. I mean, we're all product people, we love it. But don't sacrifice anything for getting the people on your team. It's all that matters once you have any traction.
All right, thank you. Thanks a lot.