Cisco buys Duo Security for $2.35bn

Scaling To A Billion and Beyond – Zack Urlocker, COO, Duo Security

In 2016, Zack Urlocker spoke at BoS USA about scaling to a $ billion and beyond. He had done it before at ZenDesk (COO), MySQL (VP Product) and more and was then COO of Duo Security. Today Cisco announced they had bought the Michigan based company for $2.35 billion. This is a talk you will want to watch. Again and again.

Slides, Video & Transcript below

Slides from Zack’s talk

Video

 

Transcript

Zack Urlocker: All right! So this is my first time here at this BoS conference and I was super excited about it when Mark said you get to pick your own walk on music. I got a few musical ideas, I tried to do a symphony orchestra – apologies for those who came early and had to hear that a million times before but here we go. Can we switch to the slides?

So the other cool thing is we get these awesome black t-shirts which as a rock guy I like that but I felt bad because it’s like I don’t think Boston is written that way. Note to self, idea for a social media spellchecker t-shirt app, powered by machine learning and big data. I think that could be a disruptive technology. A few glitches in the system.

The other thing I like about this conference, people are very friendly, so yesterday as I was walking back I met a guy, Tony Zuck, where are you? Hey! So Tony came all the way from Saskatoon and runs a digital agency ZU.com and my initials are ZU. So I thought that was pretty cool, ZU – so it’s like superpowers here. So Saskatoon is a great city, it’s wonderful that Tony was able to come out here cause otherwise we would all have to fly all the way to Saskatoon which would be a lot harder [laughing].

Who Am I?

Ok, so just a bit about me, I’m a software guy and I’ve been in the business for many years, I’m the Chief Operating Officer at Duo Security, that’s a venture backed cyber-security company, I will talk just briefly about that. I’m also an advisor for companies, DataStax, Hubspot, all these companies that start with P, I don’t know why. And I’ve had 3 billion dollar exits – Zendesk, MySQL and Active Software. I will talk a bit about scaling.

I’m also a reformed marathon runner – how many marathon runners here? A few of you, it’s a great sport [laughing] for the obsessive compulsive and you go out there with a few thousand people and don’t interact with anyone. So if you’re antisocial it’s even better. Can anyone guess where this photo was from? Boston, that’s right, I ran that a few years ago. That is the most fun you can have running a marathon. I wouldn’t say running marathons is necessarily a lot of fun but Boston is a lot of fun. So I ran marathons for 10 years and was trying to figure out what to do with my time and l learned base and built the best open source rock opera of 2016 with Rob and in fact, so we have a band, Electric Buddha maybe you’ve heard of us. Maybe not. We created a rock opera over the last year, we put it out as an open source offering under Creative Commons on funded on Kick starters – there’s a website rock-opera.com. Best $9 I spent so feel free to check that out. But we’re not here to talk about rock operas but about software. Okay.

Duo Security

So let’s go into a bit about Duo Security. It’s venture backed what we do is we protect users, devices and applications. So we have a cell phone app and it enables you to do 2-factor authentication – the reason that is important is that 95% of all breaches involve stolen credentials. Someone steals your password or devices that are out of date. If anyone here, if you have a Mac, Windows, whatever software you are running – keep it up to date. You would not believe how many breaches come from – that one might be crap I don’t know, I usually wait a couple days. One of the interesting things about Duo and some of the other businesses I’ve been in is it’s very high volume and a velocity based business. 75% of our customers get up and running on their own with no services or consulting in 24 hours.

From a business perspective it may be a bit more interesting. I joined the company a little over 2 years ago, I was an advisor for a while and we grew from $5m in ARR to about $50 million ARR in just under 2 years. Went from 35 to 325 employees during that time so it’s definitely a business that is scaling pretty rapidly.

Four Key Ideas

So, there’s 4 key ideas I will talk about this morning. It’s about looking for patterns. It’s about when to actually scale because people sometimes get it wrong. We will talk a bit about disruption because that can be a really powerful force and a little bit about velocity.

And yes, you need a product roadmap, just to be clear, but I was thinking about the comment yesterday you should or shouldn’t have roadmaps and you want to have a system that enables you to know where are you taking the business and where are you taking your product? What you need more is to understand your strategy on where you want to take things. That’s more important than what am I doing in the next 2 weeks, or however long your sprint cycle is. So that’s just my personal view, a lot of this you will have to decide which of these ideas make sense for your own business and which are in direct contradiction to 3 other speakers you heard here, and you have got to sort all that out.

90% of Startups Self-Destruct

One of the scary thing about software companies is that 90% of them self-destruct and even if you get venture funding, 90-95% of venture backed companies don’t go anywhere. They never really have a positive return for the employees. If you got in super-early as an investor you probably do okay. That’s that rock opera thing again. There’s a couple reasons, what happens is you think of your business – this comes from the start-up genome research project and if you think of the different stages of your business when you’re first starting you’re trying to discover what are the needs in the marketplace and how do I match that to a problem that I can solve. So there’s a discovery stage when you’re figuring out which problem needs solving that people will pay me more and other speakers talked about that, it’s an really early stage and as a start-up you try to think, how do I find the right thing to build and make it into a repeatable model? There’s a whole long time where you’re trying to figure out is this legit? Can I validate this in the marketplace? Are there enough customers who care about my social media t-shirts spellchecker application that they are willing to pay me to continue to build this application? And the challenge is that if you pour a lot of money and you’re in the wrong phase and you are still validating, you have 10-20 customers and say great, now I’m ready to scale, you may have a very long horizon in which you’re still validating it and you tend to burn a lot of cash.

I see this with a lot with some of the start-ups that I’m working with where you see a few positive signs but it’s not getting traction yet. So the key thing is to really look for a repeatable model and you really have to be honest with yourself about this. I joined software companies where we never really got that traction and you got the first 10 or 20 customers and as you drill in that was the founder’s brother who bought this product, and this was someone we talked into a custom solution and they had no idea what he was buying. So you have to discount those and be honest with yourself. But the key thing is, before you scale, you have to make it efficient. If you scale and throw a lot of resources on it before it’s efficient you’re pouring water into a leaky sieve. So it becomes very challenging and for Silicon Valley companies they don’t seem to mind that. You see a lot of companies where they are burning cash like crazy and they are like we’re gonna make it up in volume, we’re going to have market share. Even some of the larger companies out there, when you see some of the crazy cash burn of companies like Uber, Lift, etc. you’re like ok, how many billion dollars are you gonna lose before you actually make money? Even some of the high profile IPOs that occurred in the last few years, if you look at the unit economics of a company like Box – Box is a good company, and the founders are nice guys but if you look at the economics from their S1, you look at that and say they are losing a lot of money for every single customer they add and yet they’re scaling and there was a time when you said we’re pursuing growth at all costs and that means we will burn a ton of money doing that. Luckily the market has gone more rational and you need to have a positive return on investment. So maybe there’s a little bit of a return to sanity here.

Find Repeatable Patterns

So I often talk about repeatable patterns, the key thing to look for is, what is driving adoption of your product? And patterns doesn’t necessarily mean like 100% of your customers, it may mean to me anything between 12% and 15%, or 18%, if you see a particular kind of buyer and their use case is this. Pay attention to those things and say how do I find more of that?

What you’re particularly looking for is patterns that suggest industry, title of the person buying, they are particular use cases – what got them to look for your product and decide that today they want to go implement this? For example, Duo Security we see patterns all the time where we need to get PCI compliance so we will implement some security and looking at 2 factor authentication. Great. We can have content around PCI compliance and there’s a thing that happened earlier this year, the Panama papers, anyone familiar with that? The leaks, it was a law firm and had some data that somebody hacked in and got their password and the law firm is like. oh shit! We should have put some security in place. Meanwhile every other law firm is like, “Hello Duo? What have you got?”

We look for patterns that suggest, okay, law firms care about these issues, let’s make sure we have the right content for them. So you should look at this in terms of who are you competing with? We know, when I was at ZenDesk there were certain competitors we ran into. There were not many, 2 or 3 – we were looking, okay we need to have a playbook that is optimised for when someone is coming in and looking at us versus them. It’s this other thing and not being the general abstract like all things, there is a specific buyer, a guy or gal and they are looking, at I need to solve this problem and looking at alternative a, b and c. Why should I care about your product? Optimise the hack out of that pattern because every buyer is an individual and they come from a certain frame of mind and if you can tune your communication to that person and make sure you are speaking to their relevant stories, giving them examples from the industry, speaking to their issues and it matters greatly if you’re talking to an IT admin or a CIO or a business executive, you want to optimise every piece of that. Look at – you really want to do this full cycle marketing where you can look at the lean sources and the content and say which campaigns or pages or emails get me which type of prospect and what do they care about?

At MySQL, we used to think a lot about who is our user? It’s a DB administrator and a certain size company. What do they care about? So you have to have that domain understanding of the problems, that that particular person has, how do you solve that problem and how do you solve it better than anyone else? You’d be amazed how many companies – I sit down with them and go like let’s talk about your customers and they have this irony that happens as businesses become larger and easier to create and people buy on the web. And there’s a company that says, well we have 1000 customers. Well let’s pick one. “I’ve never actually talked to one, we just look at the 1000 of them”. And this is this terrible mistake you can make which is if you think of them in aggregate, you never understand their issues. So if you get good are understanding their issues by having face to face communications, then you can create a communication platform content that works.

[Audience question]

So the question is how do you think of these different personas? Whether it’s a buyer influence or manager, you get to talk to every single one of them and what I find is there’s lots of people involved in a purchase and enterprise sales in particular because you can’t tell who is making the decision, you just have to have a lot of meetings. I found a useful technique, before you go and sell somebody something, go out and schedule 6-10 meetings with customers individually, not a focus group or anything like that and I’d go and have a sheet of paper 8.5/11 with 5-8 questions on it. I would ask them the same questions in every situation and I would scribble notes on them to try to understand in security, database or whatever, it doesn’t matter. “Tell me about your last project. What worked and didn’t work? What were your frustrations and the surprises?” If you go and talk to customers, they will tell you everything and they know very much what their problems are and they’ll tell you all the stuff that went wrong so then you go like, ok so now I know their problems and I hear them, hopefully from different perspectives.

So classic situation when we’re at MySQL, we go to a company and we’d say if there’s a CIO and some IT guys and programmers say, ok we’re Open Source, and you guys use any Open Source? The CIOS are like absolutely not! We don’t do anything like that and a couple developers are like well, actually, our entry system or whatever it’s like that runs Open Source and this one and this one and so sometimes you get misinformation. So we learn how to communicate to those people and understand what was that person’s issues that he cared about and how do you get them all moving in the right direction?

I found that customers are pretty frank and will tell you about their problems but it’s very hard, particularly if you’re a founder of a company, to get in one of those meetings and not start saying let me tell you about my compression algorithm and how it works. You don’t wanna do that but understand their problems. Even if they say what do you guys have? Just say we will come back and have another meeting and tell you about that. So if you do that, you go to 8-10 customers and have that sheet of paper and you ask 80% consistent questions and then you spread them on the desk and look at them, you will see patterns.

I have been surprised, my career was in product management. I went out and we had assumptions about what we had to build. Then you go to the customer and you go, you know that thing we thought we had to build? Nobody cares about it because they already got a solution, they’ve got their own version control system we thought we’re gonna have to write one of those, we have to integrate with that. Really be careful that you don’t go in and say here’s all my assumptions and let me tell you why this is a great solution. Maybe we’ll come back to that.

Optimise for Specific Patterns

You want to look for these patterns and a key thing is don’t scale until you figured out what the patterns are. In particular as a system, if you think of your business as a system, you want to put more input at the top, marketing dollars, ad programs, whatever and you put sales people and what you have to see is a correlation between how much resources you put in and how much comes out the other end. Meaning, if you hire more sales people did your revenue go up? Okay, if so you’re ready to scale. And the classic mistake I see all the time from founders is, particularly venture backed – we just raised several million and the VC said I need to hire 20 sales people so we hired 20 sales people and I’m like, let me guess, 10 of them didn’t sell anything. He was like, yeah, how did you know? Well, cause if you want to scale 20 sales people productively, first you have to scale 3 of them, then maybe 6 and then 12. Why would you hire 20 people? And I understand the reason. They got a spreadsheet and were like if we don’t hire all those people we’ll never hit our targets. If you hire all those people you won’t hit it, so that’s not really good argument. So look for that predictability that says ok, if I spend money on this, I can get more and until you have that and a very clear understanding of the correlation between your spend and what you get, then don’t scale. Don’t try to grow it faster because of some artificial pressure.

Once you are scaling, then you want to look for ways to improve the performance around specific patterns – meaning we know people will migrate from this product to this one if they are in this use case. When Laura was talking about her product the other day, it’s like okay we know which users have a limitation, if they think this way great! I can go find people like that and show them our stuff and get them to buy so you need those patterns.

Continuously Experiment

One of the things that’s important is to create a culture of continuous experimentation. That means building a learning culture and you have to be ok with some of these experiments will fail. You have to make it ok in the organisation that people feel like, I will run an experiment and by definition some of those experiments are going to fail. You may wanna limit how much you invest until you prove it but you have to have this ability to do a lot of experimentation.

When I was at MySQL, we had an expression bad news is good news. Meaning if you find out a problem in the company like that’s good. Cause then you can go and fix it. And it’s amazing in large companies, I can guarantee you when there’s a publicly held company and they miss their numbers, it may be a surprise to the executives, or if you have a valued employee who quits or a customer who churns, it may be a surprise at the executive level but there’s someone in the organisation who knew that. Yes, of course we missed our numbers. The product is too hard or the leads were crap or we shipped too late or… they know. The frontline people know the reasons why things are failing. So you want to tap into that expertise and figure out, now what do we do about it? So bad news is good news. The other side of that expression is good news is no news. Yay, we hit our number! Ok, so what? Yay, we shipped on time! Having good accomplishments – you should be generous with praise, but good results don’t tell you anything. Positive conditioning is great. So bad news is good news, good news is no news. No news is bad news.

So if you’re not hearing anything from your frontline employees, then it kinda means like okay, they don’t care anymore and they’re just not telling you what’s going on anymore. So you actually have to really cultivate this ability to say it’s okay, we will try something, it may not work and if it doesn’t work, we’re all ok with that.

Beware The False Dawn

So the other thing in scaling to be careful of is a false dawn. Does anyone know what that is? There’s some term for it, zodiac something but it’s a couple of hours before daylight. I’ve never been up that early so I don’t really know about it but you see light on the horizon and you think it’s sunrise and it’s no, it’s just a reflection! So the thing to be careful of is when you see a couple of successes, and you say hey, we sold this one and this one. Great! Now let’s scale 10x or 100x – be very careful of falsely assuming that things are repeatable when they might not be.

So as an example you close that $1 million deal with a large enterprise and you say great, now I am hiring an enterprise organisation. Look out, we’re gonna take on Oracle. Every company I ever worked for has made this mistake. We’re gonna go to the enterprise next, we’re a mid-market – now we’re going to the enterprise. We hired a bunch of people and it’s still a year away. So be very careful! In particular you wanna avoid some of these early signs of, you close $1m deal or multi million deal but it came with a whole bunch of requirements that you were never planning on doing and it’s classic enterprise sales where, last day of the course, sales guy comes in and good news, we got this $3 million deal. The only problem is they need our SaaS thing to run on premise and it has to be PICs compliant and have to integrate with their IBM system over here and translate it in to 14 languages but it’s an awesome deal! You look at that and you’re like okay, great… My advice is when those deals come in, unless those were things you were planning on doing in the next 6 months anyways, just walk away and don’t do them. Ok, if you have to make payroll and this is your difference of making payroll, maybe do the deal but it’s a deal with the devil. You want to avoid doing these one offs assuming they are scalable and then ending up in this situation when you keep building things that are inherently non-repeatable.

The discipline that I was trying to have is sell what’s on the truck and you have to train your salespeople. If it’s a deal and comes with a bunch of strings attached, we won’t take it. If you see those requests and it’s going to happen for the next 30 customers, go ahead and do it. Personally I would be very careful of growing head count faster than your revenue growth. The good news it’s really hard to hire good talent so there’s an inherent latency – you want to hire a bunch of people and you can never hire them as soon as you can but be very careful if you end up hiring headcount in advance of your revenue forecast.

Signs You’re Not Scaling

So a couple signs that you’re not scaling – every deal is customised, they are all different. Organisationally one of the things you see is the executives and the company are making every decision – when I first joined ZenDesk, Mikkel, he’s a great founder and an interesting guy, we had a line outside his door of his office and he said I don’t understand why people come to me and asking me to make decision – it’s because you let them. You have to say I’m not deciding on these things, you decide. I will talk about it. Another thing, it’s where people aren’t hitting these numbers, and a lot aren’t, and everybody is working 60-80 hours a week and going crazy! These are all signs that you’re not scaling and if you’re in that situation just take a step back and say ok, again go back to what are the patterns? And take a lot of stuff off the table and say we will focus on these 3 patterns. This class of buyer with this problem – we will be the best at it. And for engineers and founders who are particularly more on the technical side, you always have this feeling like you know, it’s software so we could do anything. What do you guys do? Do you do this? Well we could… What you actually need to do is be laser-focused and you are way, way better off saying we do this. We do this one thing and do it really well. Don’t talk about the things you might do or could do, just find customers who have that exact problem. And it might seem that then we’re limiting our business? It’s probably not the case and what will happen as you will be more effective in having a narrow focus and having everybody in the company know we solve this problem better than anything else. MySQL was the world’s best web database, or at least that’s how we talked about it, or the most popular Open Source database. We didn’t try to do all things that a database could do. In some ways, technically – to be clear, in every technical way, MySQL was not the best database. Oracle is a much better database for enterprise applications, but we knew the one thing we were really good at.

Disruption Changes the Game

So let’s talk a bit about disruption. Anyone heard that term? Everybody, ok. So this comes from a book The Innovator’s Dilemma by Clayton Christiansen and it’s a really good book but lots of people buy the book, very few read the book. Watch the video, it’s more efficient, or we’ll distil it down to 3-4 bullet points.

Disruption is really an interesting thing. You could have a successful business without being disruptive but if you are truly disruptive it’s like having the wind at your back. The funny thing about being disruptive is it actually means you have an inferior solution, when judged by the traditional criteria. Meaning in an established market you have something that is decidedly worse for most people. When the IBM PC came out and it was competing against mini computers, how did people regard the PC? Yeah. It’s a toy and joke. 64k? Are you out of your mind? DOS? You call that an operating system? It wasn’t considered a serious solution, by the traditional criteria.

When Linux came out or MySQL or AirBnB – I think AirBnB is a super interesting business but I tried to explain this to my wife – you’re gonna stay in somebody’s house? Like do they clean the sheets? How do you know it’s not an axe murderer? By traditional criteria, you’d say that’s kind of a weird thing to do but if you’re really successful with it, it does actually change the game.

The iPhone was a huge game changer in the cell phone industry. The first version, how was it as a phone? Yeah, it was a shit phone, right? I mean like it couldn’t – the mic quality, calls were dropping all the time, the connectivity was terrible. Everything about it was terrible. And classically, Motorola, Palm, Microsoft all said, that thing is never gonna fly. It doesn’t even have a keyboard! There’s some famous quotes where people really misjudge that phone because they were thinking of it from the traditional criteria of what a phone was and how they would interact with it. It changed the game.

Are You a Disruptor?

The question is – I love the show HBO Silicon Valley, but it’s a bit too close to the truth to really be funny. I’ve been in those meetings. So how do you know if you’re a disruptor? There has to be an existing market, people have to be buying the thing in the category in which you’re trying to exist and spending money. You can create a new business and a new market, it just means it’s not disruptive. So an example, Duo we’re in the security space, security is a $60 billion market, there’s $2 billion around authentication alone. So there’s an existing market.

Secondly, there are underserved customers. If you think of the market as a pyramid, it means top of the market – what do we call it? Ok, enterprise, mid-market and SMB. So typically there are people at the bottom of the market who are under-served and they say I’m not buying anything. They choose not to buy. To really be disruptive you disrupt from the bottom up and you expand the potential market by appealing to the people who are consciously staying out of the market. They aren’t interested in the same things that the top of the market has.

So to be disruptive there’s an existing market, means there are incumbent sellers that are typically selling at the top. You go and sell at the bottom. So Hubspot is a great example, lots of marketing companies, Hubspot comes and dominates in the SMB and graduates to the mid-market.

Third thing is you compete on different criteria. The things the under-served care about are different than the big companies. Hint, don’t listen to the big companies, go out and talk to your target market and ask them why they don’t buy stuff! In security this is easy – why don’t you buy security? I don’t think we will be targeted. We don’t have any security expertise or how to implement it. We don’t know what’s important, we’re busy. Lots of reasons people stay out of buying security products.

The 4th one is the hardest one, which is the new business , to be disruptive the new business that you’re creating has to be inherently unattractive to the existing vendors. And Clayton makes a big distinction between disruptive innovation, and sustainable or sustained meaning incremental, improvements. And you wanna ask yourself will the incumbents, the competitors I’m facing, do what I’m doing? Why won’t they do what I’m doing?

Relational Database Market

So let’s go through a mini example. We’ll talk about the relational database market, this is David, the co-founder of MySQL he is on a boat in the harbour in Stockholm. Give away the software free, make it up in volume. Yes, that’s a winning strategy. Guess whose that boat is behind David? It’s Larry Ellison’s yacht [laughter]. There’s an old joke on Wall Street but where are the customer’s yachts?

So Oracle is a great company and very efficient but we’re trying to disrupt these guys. Database market is $30-$40bn dollars. It’s an oligopoly and this is going back a few years when I did this. So dominated by Oracle, IBM DB2, and the database market at the time that MySQL came up you could only describe it as moribund – this is a market that nobody cared anymore and we had Ingres, Cybase and Informix and they were all gone or less relevant than they used to be. It was dominated by a small number of players. So when we started going in there, it was like you’re going to the database market? That’s over, Oracle won! You remember all those billboards on 101? Why would you compete with a company that’s 10,000 larger? Well, because we are gonna disrupt it.

Disrupting the Database Market

So this is just an example on how things work. The incumbents, the traditional database players, they focus on customers who had mission critical applications, that would run on expensive hardware, big Sun servers, big HP servers, proprietary systems, basically selling very high prices, so typical Oracle license is $40k or $50k per CPU and stuff like that and they have an expensive direct sales force – Armani wearing, Ferrari driving guys who make millions every year. And a big focus on complexity, features. like we’ll just take care of everything for you. It’s a great but complicated product. And a lot of the focus on differentiation was all about TCP benchmarks. Nobody actually runs TCP benchmarks but that’s what they focused on. MySQL, we were like the complete opposite in every single way. So at the time, this is 2003-2005, we did web applications. CIOs, they didn’t even know they had web apps. They were like, we run the data centre stuff, there’s some guy with tattoos and pierced body parts, he runs our website, we don’t know anything about it, it’s just some static pages. Would that matter? At the time it was a completely underserved market. Web developers – they couldn’t even spell DBMS. But we had a lot of under-served users, basically web developers who learned HTML and were like how do I get a bunch of customer information there? You need a database! What’s a database? Just go download MySQL and you will be fine. We grew a market on that. It was the definition of underserved and was completely ignored by all the vendors, Oracle and Microsoft, they didn’t care about web apps. At the time nobody did, it was just this obscure corner of the world.

We ran Commodity X86 hardware and then the most disruptive thing, which was a historical accident, was it’s free. It’s under an open source license. How many people here have used MySQL in your businesses? How many people paid? 1 or 2, it’s typically 1%-2% conversion rate of people that actually pay. Which is a very tough business model, but at scale it actually works. Meaning, like we would get 30k downloads of MySQL per day. Some people probably downloaded it twice, but it’s high volume so if you’re at that 30k and you get 1%-2% that actually pay. I built the business model around it so there was some value to paying – but every single element of what we did was in some ways the opposite of what the big guys were doing.

If you think about it as a market, we never aimed for the top of the market. We didn’t want to convince the stock exchange they should run MySQL. Instead, we were focused on web developers and we made that really work and we focused on certain patterns there of web developers and OEM software developers. Back when people used to ship software – there’s a class of applications that run on appliances or you would embed a database in there. We did really well but understood clearly the patterns between the buyers and how they would use it. It was fairly easy to go so how we could – we were a better fit for those apps than Oracle. We always said, when people were like can you migrate this thing from Oracle or DB2 to MySQL? I was like no, you should just keep that running and keep it where it is. We can integrate with it and have replication, we can run backups etc. but just keep that thing running because no DBA ever wants to – there’s no upside to migrating your database. If you migrate it to something else and it works you keep your job. And the opposite, if it doesn’t work, what the hell were you doing? It was an interesting time in the market.

We were successful and grew that to $100bn in revenue, we sold the company to Sun for $1 billion and a few months later – Oracle had been trying to buy us at the time but Sun came in and moved much faster. We thought Sun was in good shape, maybe they weren’t in good shape after all. Then Oracle came back 6 months later and tried to buy MySQL out under Sun and then 18 months later Oracle bought all of Sun. I don’t think they bought Sun just to get MySQL but it was a nice side benefit and the MySQL business runs inside of Oracle and it’s actually continuing to grow and the technology is done very well, which is one of the outcomes when you’re disruptive. Sometimes the big guys, it’s hard for them to create disruption but it’s easy for them to see disruption in the market and think how do you own that? I see a disruptive force, how do I aim that at Microsoft or my competitors? I think they’ve done a good job with it.

Are You Being Disrupted?

The other side of that coin is to consider are you being disruptive? How do you know if you are being disruptive? Everything becomes about price and a commodity offering and whenever you talk about your differentiation the customers say, we really don’t care about that, we want the cheapest thing. CPUs have gone through that, storage is heading in that direction. Your traditional place that you have in a big business, they don’t work anymore. You launch new products and nothing happens and then you see this weird phenomenon where the things you think are a toy, they’re taking off. I don’t understand why people buy that thing, it’s a toy! Our stuff is way better. Sun could never get their head around the fact that x86 boxes from everybody were outselling Sun. They were like we have redundant powers suppliers and redundant this, and if you go to a big company, and they have racks of x86s, machine dies who cares? You got fault tolerant power supplies? We don’t care. The machine dies, we remove it and put another one, we got 1000s of them. So the whole mind-set is different.

It’s interesting when you see companies being disrupted, the management team is like, we need a miracle. All those companies down there are being disrupted, some have already been disrupted, not to pick at anyone but if you read Yahoo and I’m sure Marissa Meyers is a great story but you read the interview, yes we build a new business around mavens, analytics, something, something, and what do they buy? Tumbler for $1 billion, it’s gonna be awesome! They buy 20-30 other companies, that’s the definition of a company that is looking for a miracle. They did find it because they had some investments in China that actually worked out.

The other thing about companies being disrupted, instead of building disruptive strategies, what do they do? Every single one of these companies they redesigned their logo [laughing]. Disruptive! Like the Yahoo! logo is different than the one before. Could you tell? It’s crazy – I used to have Microsoft up here but actually it’s done a good job over the last years. So even good companies can get out of this, but it’s very hard because big companies, they are focused on preserving their existing business, and the best thing you can do as a small company is understand how those big companies operate, and someone was talking about competing against YouTube – and if you understand their playbook and you have an offering they cannot do because it would ruin their business… We knew Oracle, Microsoft, Cybase etc, what are they gonna do? They won’t make their product free. They can’t do that. It would decimate their revenues, so what do they do? They come up with Oracle Express. Microsoft thought that was a great idea and had a SQL Server Express and everyone has their Express database and what is that? Well it’s free – has a few limitations, there’s a famous Oracle website, Oracle FAQ and they run on MySQL, which we thought was awesome! They are like half pregnant, like we want to disrupt ourselves but we can’t because we would destroy our business.

Velocity Trumps Everything

My personal view is that velocity trumps everything. So when you’re building a business, trying to get – everything is about getting speed and you have to look at all – you can run a business lots of different ways but I found that speed is really good so you want to make your product easy to understand, to buy, to use, to consume, easy to get into production. Every decision you look at say how easy is it?

For a while we had a trial process at ZenDesk and there were 39 steps involved in getting in to do a trial. 39 steps? It’s a good book, it’s a funny musical, you don’t want 39 stops in the signup process, that was a joke [laughing], 39 Steps? I thought maybe some English people would get that. But you have to say how do we make this super easy? And if you’re an executive in the company, go do the trial signup process, do it routinely and say how easy is it? When you bring in new employees, find out how easy is it? Make it easy and it means tell – make it a short, concise story, don’t tell them everything about how the product works, just enough that people need to get going.

Optimise the heck out of that, because the more velocity you have, the more robust the business you’re actually able to build.

When in doubt, simplify. Any time you look like we could have 8 versions of the product, add the custom this, custom this – if it looks like a Chinese menu, simplify. Maybe you have 2 or 3 versions of the product, but don’t have 12, don’t have a bunch of add-ons or make customers understand your technology before they buy. They don’t care about how the technology works in general – most cases they just want to know does it solve my problem? Just be upfront about those issues and look for ways to make the machine move faster.

What Increases Efficiency?

Couple ways to increase the efficiency, so where you can deliver more value you can increase your ASPs if there are certain features that matter to certain classes of customers, that’s a good way to drive your prices up. If you do better pattern matching within your segments you can improve your win or conversion rates. You need to study them and understand why do some deals close faster than others and think about okay, how do I optimise every single phase of that process for buying and acquiring customers. Create a culture of a lot of testing. I talked a bit long so I gotta move fast –

Build a Culture That Scales

Last thing is about building a culture that scales. When you’re hiring, even if you’re growing rapidly try to hire people who have headroom for a couple years. If this is the biggest job they’ve done it’s hard for them to grow. You probably want a mix of people but having someone who has grown in that path already is a good thing. I look for people who are consciously competent meaning they understand what it is they’ve done that has enabled them to succeed. And sales, this is a tough thing because sometimes this guy is a great sales person but they can’t teach other people what the techniques are. So we are looking for people who are aware of how they operate and can teach that to other people in the organisation. You’re looking for a track record of growth and learning, things they didn’t necessarily know inside the business. If you’re building a disruptive company be very careful of hiring anyone with too much incumbent DNA. So when you’re hiring – MySQL, we had I think 1-2 people in the entire company who ever worked at Oracle because if you’ve been in Oracle for a long time you just think of the world in a certain way and it’s hard to turn that upside down.

So when you’re disruptive, look for people who have been disruptive in other cases. One of the key things that I do is I try to push decision making down in the organisation. And this is maybe the toughest thing for founders to do because if you’re a founder, it’s your idea and baby and you’re smart, probably the smartest person in the room and naturally you want to control the decisions. And a lot of executives get in this habit when they are like, I will control everything and make every decision and then when something bad happens, shit hits the fan, suddenly no employees know actually how to make decisions because you’ve never given them that opportunity.

So I generally avoid making any decisions that are not fundamentally strategic – I organise an offsite with the team and will get some help from an admin and say do you want to do it this or that way? You can decide – but what food do you want? I don’t care what the food is, you have good judgment, make the decision. You have to train the organisation. It’s a trivial example, but you want to train your organisation to make decisions. And accept that some of their decisions will be wrong. You want to scope what decisions they can make and which ones are truly strategic. The decision to open source your software is a bit hard to undo so you might not delegate that to your junior intern developer but a lot of the decisions, you need to push things down. The people that are close to the problems, they know what to do. But you have to force people to make those decisions and that’s how you get velocity. People then feel like I feel ownership in the company and able to make those decisions and if you ever feel like your job as an executive or founder, you’re just approving stuff all the time. That doesn’t add value, giving guidelines, so it’s about how to make decisions that’s great. Determining the strategy where you want to take things that’s super, making sure the organisation is listening to customers, making the decision about do we ship this feature first or that one second or with the QA good enough? Hire people who can make decisions and then let them make decisions, force managers to make decisions and if you have managers who after a while they aren’t making decisions, they can’t be managers. Their view is I pass the information from up above down below and vice versa – we have email to do that, we don’t need you to do that.

The other thing is diversity of view point helps get better decisions so a lot of decisions and that’s diversity in every single way so ethnic diversity, background, people that come from different companies. Dharmesh described Hubspot a lot of execs coming from this traditional MBA background and they are very quant focused – that left them a blind spot and you want to think about how do I want to complement the skills I have within the organisation and add a different perspective? Often we do management off-sites, like we will put together a group of people from all kinds of diverse areas, finance, marketing, security – to talk about product strategy, and people who are not the product experts because sometimes they will say something that’s like, I never thought of it that way. That’s how you get better decisions and the irony is the tougher the decision, the more likely the executives will want to control it and say we need to decide. Actually sometimes you open that up and encourage participation from others, you will end up with better decisions.

Find Your Own Path

Ok lastly is when you’re building a business, you have to find your own path so it’s not about following a playbook from Atlassian or MySQL, you have to find your own path and the key to that is to focus on your customers. Competitors come and go. At ZenDesk we had 12 different ZenDesk clones some of those still exist today, MySQL there’s a lot of databases. We focus on solving the problems of our customers, don’t worry about the competition – you can’t really influence them anyways. And don’t worry too much about what category the analysts put me in – just figure out how to tell my story so that my customers understand what I’m doing.

Know More Than Your Board

Ok, we’re running out of time but I think this is the last slide. How many people here have a board? How many people think you get conflicting views from your board? Probably a lot. I’m on a couple of boards and I decided to do that because I want to be the helpful board member. Cause I’ve been in lots of board meetings and sometimes there’s someone that is helpful and sometimes less so. And there’s a certain art to managing your board. You want to get their input on things but at the end of the day as a CEO, and even if you’ve had investment from venture firms, believe me they do not want to run your business. And what you have to do is really get their input on the issues but own the decisions, own the strategy and don’t worry too much about trying to please your board cause if your board talks you into doing something that you think is wrong and you do it and you screw up the company that’s when you’ll get fired! No one will remember that was the board’s idea – that wasn’t my idea! Like, you’re accountable. Don’t worry about how things look, don’t worry about the optics, I hate it when people talk about the optics of a decision. Like, how will this affect morale if we do a lay-off? How will you be in business if you don’t cut your burn? Worry about those issues. Don’t worry about how things look. Worry about how things are. Get their input. It will be contradictory at times and just say here is what I have decided. And your board wants you, if you’re a CEO or founder of a company, they want you to make those decisions. They’re tough decisions and sometimes you can get a friendly person on your board who can be your counsel or whatever, talk to people from your board but let them know how you’re thinking about things. And you know the business way better than the board members do cause you’re living it all the time. And if you’re in doubt about a strategic decision, go talk to customers so you understand that better than they do cause they think about your business a few hours month, a few hours a quarter at best, and you’re living it all the time.

Ok, I think that’s about everything. Blah, blah, blah, be positive. Make a difference. Ok, thank you!