I have now spoken at the Business of Software conference for 3 years in a row. It is my favorite conference both as a speaker, and as an attendee. The reason is simple -- the content is great and the attendees are awesome. The following is a full video of my 60 minute session. If you enjoyed my talk from Business of Software 2009 or Business of Software 2008, you'll likely enjoy this too. I think it may be my best one so far. And, thanks to Neil Davidson, Joel Spolsky and Mark Littewood, I'm thrilled that I will be once again speaking at the Business of Software conference in Boston. If you're in the business of software, you should attend. It's a great use of your time.
The following is a full transcript of the session, for those that like to scan/read instead of watching video.
Dharmesh Shah at Business of Software 2010
Intro by Joel Spolsky:
Our next speaker before lunch is Dharmesh Shah . Many of you will know him from his awesome blog, OnStartups.com, which now has a Q&A section called answers.onstartups.com which part of the stock exchange network so it’s all one big beautiful overlapping thing. He’s the co-founder and CTO, I think, of HubSpot, which he told me, where he manages nobody, but provides a great deal of leadership. And this is, I think, his third company. So, please welcome Dharmesh Shah.
How’s everybody doing? So you just got done watching Seth and David, who are both exceptional speakers; very organized, they have this cohesive train of thought, they have a message, they have structure. I have none of those things.
So, I’m a hacker by trade. By the way, who’s heard me speak before at a business or software conference? Watched the video? Ok a bunch of you. I’ve gotten a little bit better, but not very different. The material is all different, by the way. I’ll try and speak a little bit slower. Well, actually there is… we’ll see. We’ll see how that goes.
We’re going to talk about building software businesses. I put this disclaimer up every time I speak, which is I write code every night; I’m a hacker, I don’t speak professionally for a living. So, be gentle. If you have questions, by the way, along the way, feel free to ask them. One of the upsides to me not having like a consistent string of thought is that you can’t interrupt it. So, just raise your hand and try and get my attention jump up and down or something if you want I’ll just double click on one of the slides and we can dive in.
I picked things out of the last year’s worth of experience based on what I thought would be interesting and that’s not obvious. At least it was not obvious to me over the course of the last year.
Looking back, just one quick thing –uh, two quick things: So, last year I launched my book about marketing. You guys helped propel it; it did really well. It’s still doing very well; number 7 out of - not that I’m counting how many marketing books - I actually wrote a script that does that and tracks my rankings hourly so I don’t have to do it myself, (Laughter) but anyway… thank you for all of you that bought it.
And, I’m expecting my first child in January. My new startup venture; hopefully it will require less capital (Applause).
As Joel (?) mentioned, I write a blog; lots of you read it, I think. And HubSpot is the company – it’s my third startup, and a lot of the stuff I’m going to talk about actually comes out of HubSpot for a couple of reasons. Not- we didn’t invent this stuff, and one of the dangers of speaking and talking about software businesses and startups generally is most people, including me, extrapolate from a data point of one. Which is “oh yeah well, this is how it worked” and “boy isn’t that great” and I’ll share with you that data point of one in terms of stuff that worked for us but at least it’s better than extrapolating from a data point of zero. Right, so we’ve actually tried this stuff and kind of know what’s worked for us or not and you use what works for you.
I put this slide up here a) To brag (Laughter). This is the customer growth chart and if you notice like back in ’07, we were kind of flattish, just getting started. That’s when I first spoke at Business of Software; this is my third year in a row and clearly my presence at Business of Software is doing really well for the business (Laughter), so I hope to get invited back. And this is customer growth, by the way, the revenue growth ramp is steeper, because revenue per customer is growing and we’ll talk about that as far as metrics. I don’t want to talk about HubSpot the Company and what we sell because I’m not salesperson; I don’t sell things. We’re going to talk about the business.
One thing that’s interesting about HubSpot is of the first nine people on the exec team, all of which are still there, or still here – they’re MBAs, including me, including my co-founder, including our VP of Engineering, including our VP of Sales, all the way on through. And so one would think that we were like doomed to failure from the beginning, with that many MBAs sitting in a single startup. And we’ve managed, so far, not to screw it up and we’ll talk about that.
All right, and as I mentioned, all this stuff does not come from us. HubSpot – one of the things I love about working at the company, one of the things I love about the team is we are students of the game. So we know we haven’t quite got it figured out and we know there’s going to be plenty of opportunity to screw it up so we like to talk to as many smart people that will give us the time. These are just some of them. There are lots of smart companies in the world; these are folks that I’ve actually like consumed copious amounts of alcohol with or sat for multiple hours and gotten the chance to tap their brain. And I’ll talk about Drew from Dropbox because he’s the one I consumed alcohol with most recently, and was influential. By the way, how many people use Dropbox? I’m just curious.
So, one of the things that’s interesting to me – this is one of the things that really appeals to me – this is my favorite conference, I’ve said that publicly many many times, is because it’s about the software business. And the operative word there being both software (obviously) and the fact that it’s a business.
And that we are trying to build businesses, not do interesting projects – these are businesses, and it’s ok to talk about revenues and profits and margins and customers and product and those kinds of things as a business. So I’m going to spend a very brief amount of time because I get this question every now and then about venture capital. HubSpot has raised 33 million dollars of venture capital in our last 4 years. And it’s my first company raising venture capital. I’ve been a self-funded bootstrap startup entrepreneur and blogged a ton about why you shouldn’t raise venture capital. Not because it’s evil, which I don’t think it is. But I don’t think it’s necessary for most of you. So if you’re a first time entrepreneur, I’m going to leave this message and we can definitely chat about it over a beer tonight or something like that in terms of when is the time to raise it or not raise it. But the key here is that it’s not necessary.
Too many entrepreneurs get started, like “I’m building a software business; here’s what I want to do, and the first thing I’m going to do is go write a business plan or a PowerPoint deck or do something and then go off and try and raise capital”. That’s a mistake. Because as soon as you try and do that, essentially you shift your mode from solving a customer’s problem or an industry’s problem to solving the investor’s problem. And those are not the same problem. Investors have a very different set of problems that they are trying to solve. So, if you can avoid it, don’t raise capital too early on, especially if you are a first time entrepreneur. We can talk for an hour about just that. And whatever capital you raise, make sure it fits.
There is a reason HubSpot has raised it. Quick question for you guys: How much capital do SaaS businesses raise prior to going public? We’ve looked at all the publicly traded software services businesses, so Constant Contact, LogMeIn, SuccessFactors, NetSuite – all of them. These are software business – hosted software, multi-tenant, doing smart things – they went public so they’re obviously relatively successful. Back in the day, when I did HubSpot it’s like, “OK, well we raised a million dollars of angel capital and I have no idea how we’re going to spend a million dollars of capital, because it’s just the two of us… why does software require that much… like, ok great we’ll just make this last two years and we’ll see if the business goes.” And we raised another 5, another 12, and then all the way up to 33. The answer is 42 million dollars median raise pre-IPO. Not average, but median – across all public and traded SaaS companies. So all these SaaS companies raised a ton of money.
And the question is: well, why would they do that? And the answer is for a SaaS company specifically, where you have a subscription-based model, it takes a fair amount of capital to acquire customers. And the faster you’re growing, so the ones that raise the most and the fastest was salesforce.com and their growth curve is steeper than HubSpot’s. It’s actually the only publicly traded company that’s steeper than HubSpot’s, which I’m proud of.
So it takes money to grow these SaaS businesses because when you’re doing a subscription based model you’re incurring all your acquisition costs to get the customers (sales and marketing, etc.) upfront. And then they pay you over time, usually on a monthly basis. Even if it’s an annual basis. Usually the lifespan for a SaaS customer should be 4-5 years if you’re doing things right. So the reality is you spend all the money upfront and then the revenue comes in over time so essentially what you are doing as a SaaS company is a form of customer financing. And, oddly the faster you’re growing, the more customers you’re signing up, the more capital you will consume, fueling that growth. So that’s one non-obvious thing that occurred to us along the way.
There’s an investment firm out on the west coast called Pacific Crest that did some research that says across about 78 privately held software companies, how much money does it take to get to specific points of growth? And then how much money do they expect to take before they break even? This is about right. I’m sure lots of you have gotten to better levels of revenue without raising this kind of money but this feels about right as a rule of thumb. It’s a great report, by the way. We can talk more about other numbers from that.
So I’m going to go through some very Obasic metrics. Some of this stuff you know, or have heard from me, so I’m going to be very quick about it. It sets the model for some of the things I want to talk about. One is your Cost of Customer Acquisition. And this is your total Smarketing (sales and marketing dollars) divided by the number of customers you sold. And this is all in- so this is salesperson’s salaries, marketing people’s salaries, marketing programs, all advertising, AdWords, the website, all of it - and divide by the number of customers sold. So let’s say you spent $50,000 and you got 1000 customers, your acquisition cost was $50.00. So that’s an important number to know in terms of what it costs you to acquire a customer.
Another important number, which we’ll talk about, is the overall sales velocity. How many new customers are you selling month over month? And what’s the second derivative on that? How’s the growth going? We’ll talk about the way it plays in.
The other number we need to know is the Lifetime Value of a customer. ARPU stands for Average Revenue Per User. Essentially how much money do you make per customer? Let’s say we’re just talking about monthly business. It applies equally well if you are talking annually. So let’s say you’re charging $100 per month and your average customer stays 50 months. Fifty months worth of customer times the $100 of revenue, which equals $5000. So that Lifetime Value is important and the obvious one, which we are not going to get into, is that your lifetime value should exceed your cost of customer acquisition. So if the total money you will ever make on a customer does not exceed the amount of money it took you to get that customer, something is fundamentally wrong.
Now the non-obvious stuff. One is the retention rate. So if you look at what causes a customer to have a lifetime value of x, there are two factors that go into it: How do you charge them per period, and how long do they stay? The “how long do they stay” part of the equation is around your retention rate. Of the customers that sign up, how often do they cancel? What’s the cancellation rate or “turn rate”? And this one’s actually kind of subtle, because different people measure retention rates in different ways. So one way you can measure your retention rate, or the inverse of which is the turn rate is this: Let’s say you turn rate is 3%. That means if you started your month with 100 customers, of those 100 customers, 3 of them will cancel, on average. That’s a customer cancellation rate.
Another way to measure it is revenue cancellation rate, because different customers might pay you different amounts of money. So you might have going into it 100 customers, and the number of ones that cancel are the ones that are paying you the most, let’s say. In which case the cancellation rates are higher than 3% even though 3% of your customers cancelled so on a revenue basis, your turn rate might be higher.
And the last but very important piece is what we are calling, like many people are, is called “discretionary turns.” Discretionary turns essentially is the customers that cancelled that actually had an option to cancel in the first place. So let’s say you go into that month with 100 customers. But only 80 of them actually had the opportunity to cancel because some of them are in 6-month contracts or 12 month contracts or whatever. So just the fact that you kept some of those, you shouldn’t get real credit for because they didn’t have the option to leave. They didn’t have the option to walk out the door. So the discretionary turn is really what you should be watching because that’s the signal that tells you “are you doing something useful? Do people value it? Are they happy or not? So those are numbers that are extremely good to know. And one of the non-obvious things we learned at HubSpot is the absence of turn or cancellation is not the same as the presence of delightedness. So even if they had the option to cancel – they could have hit that discretionary cancel button – “I am out the door, this stuff sucks” – just because they didn’t do it, it means they’ll continue to pay you, which is great, but it does not necessarily mean they are delighted, or happy. And your long term success – low turn rates are absolutely important, but you need happy customers otherwise life gets really hard very quickly because they happy customers are the ones you get the reference and word of mouth. I’m not going to talk about marketing but it’s extremely important so you should be measuring that stuff in terms of how everything from classic customer surveys
You could do classic customer SAP surveys, you could do Net Promotor score, which is a simple two question survey, you could put something in the product like Wufoo does and say “Are you happy?” right inside the product and people could tell you, “Yes, I’m happy”, “No, I’m not happy” and measure that just to kind of get a sense. Just because they’re using the product does not necessarily mean they are happy.
One of the things that was fascinating at HubSpot is all these little tensions that exist within the company. We’re 180 people now; we’ve grown really fast - I think it was about 100 last time I spoke here. I’m going to talk about one of them. There are all these little conflicting forces where you pull on something and something else gives.
So we’re going to look at the three primary things we just talked about: Sales Velocity: How Many Customers Are You Selling Every Month, Acquisition Costs Per Customer Basis and Lifetime Value.
In most cases, if you try and prove one, something else is going to suffer. All other things being relatively equal. So for instance, let’s say we came in and said, “We’re not selling enough customers. We want to drive sales velocity up.” And we’ve done this at HubSpot - we’ve done all sorts of crazy things at HubSpot, and so we can go to the sales team and say, “Oh, by the way, we are going to increase quota and goals by x percent.” We’ll try that and see what happens. And as it turns out, it does work. But multiple bad things can happen. So let’s say we do that and say “Ok, we’re going to sell 10% more” and let’s say we want to keep the Acquisition Cost the same. We’re not going to spend any more for leads - we’re not going to get any more better prospects in the door, better leads in the door. What will happen, invariably, is Lifetime Value will go down because you will sign unhappy customers. Because the sales people are going to be more aggressive, they’re going to do more arm twisting, and essentially you’re going to get less happy customers as an outcome of that process.
Let’s say for instance in that same example, we want to grow sales velocity but we also want to keep the same number of happy customers. We want to leave the lifetime value the same. We don’t want to degenerate that. What happens then is your acquisition costs go up. Because what happens is sales people say “Oh, we need more growth - that means we need more leads coming in at the top of the funnel so we can cherry pick and sell the best customers”. If you want to have happy customers, just give me 500 people to talk to instead of 400 people to talk to and then I can sell more. But the reality is those additional leads are going to cost you money so your acquisition cost goes up. So the moral of the story here is that in most cases, if you’re not very very careful, you will pull on something in the system - there’s this great big system dynamic, where you turn this knob or dial over here and something unexpected happens somewhere else. So, the trick is finding that balance and the balance is very tricky in that there are always little mini subsystems within there. This is just a sales/marketing vs customer happiness vs acquisition cost, but we can take any 2 or 3 variables within the company and what we found is that they’re interrelated somehow. We do something over here and something else happens over there. So you might be asking yourself, “What do we do? I want to improve sales, and I want happier customers, and we want to drive our acquisition cost down.” The cliche but right answer is - invest in not just the product - invest in the experience. And there’s a very good reason for this, and it’s semi-obvious, which is if you invest in the experience, everybody wins. It’s easier for the sales people to sell a better experience. Customers are happier, your Lifetime Value goes up, because they’re going to stay longer, because the experience is better. Acquisition costs go down.
We have 60 salespeople at HubSpot, out of 180. Very sales and marketing driven organization. That’s probably what fueled that growth. We woke up about a month and a half ago, we did some surveys and some soul searching and we said, “Well, this kind of sucks”. And it sucked not because we were not selling - we like our sales people just fine. We want to build a multi billion-dollar business. The path that we were headed on was going to create a 250 million dollar business, a 500 million dollar business, not something exceptional that our grandkids were going to talk about some day. So we made a radical change.
Up until last month, HubSpot would hire 3 new sales reps a month to drive the sales velocity, to drive the sales growth - which required marketing spending to drive more leads and produce content. We put that to zero and said “we are not going to grow sales and marketing anymore. Period. We’re going to take all the cash we would have spent on sales and marketing, and we’re going to pour it into product and into experience. And we think that’s the right thing to do. We had to convince a 60 person sales team that we’re not growing sales/marketing. We’re not going to grow marketing the company, the board, the investors - a bunch of people. We’re going to take a hit next year in terms of revenue projection. The idea is invest in the product because it works and because it’s the one thing that essentially you don’t get those tradeoffs in terms of other things within the system breaking. The other thing that was not obvious to us was there is a difference between making customers happy and making happy customers.
So, the making customers happy. I had a company meeting about a week ago where we announced some of this new news and what I said was that, we don’t want to focus on making customers happy, because what that connotes is, “well, ok, sales and marketing does this over here, and then some number of customers shows up - we have 3200 of them now, and then it’s product and engineering and support and all those folks that are then ‘responsible’ for making the customers we have happy.” That is not the right answer. The right answer is the entire purpose of the business is to produce happy customers. That’s the output. It’s not something you do after the fact and say, “Oh, we’ve got these customers - now what do we do to make them happy and keep them happy?”. It starts from first exposure - it starts from marketing, all the way on through sales and everything. And so everybody in the company needs to be thinking about, “Is what I am doing contributing to the raw goal of the company?” - which is to manufacture happy customers. Don’t think about customer happiness as a subsequent thing that product and support worry about. Everybody should be thinking about producing happy customers.
At HubSpot, we have a bunch of these little business hacks - tricky little things that we do that we look back on and say “wow, that was kind of smart of us - that was clever”. This is on our top 5 list. We have a number at HubSpot called the Customer Happiness Index (CHI). And we’ve had it since month 3, so we came up with it relatively early on. It’s a geeky way to measure how happy customers are. It’s a number from 0-100 that measures the probability that any given customer, given the option to cancel, will still be a customer next month. So we can go through all 3200 customers and we measure the happiness index. And the reason this is important is not because it’s important to know, and that is important, obviously, the number is immensely valuable, and we can talk about all the ways that HubSpot uses this one number. But the biggest value of CHI, is its simplicity. The fact that it’s a single number that ties to not profits, profits would be great, right, you want financial numbers. But it’s hard to hug a dollar. It’s hard to get employees all riled up and say, “Yes, I want the P&L to go up by 7% next month or next quarter.” Yes, they need to care at some point and we’re all, in a positive way, red-blooded capitalists. But the nice thing about the Customer Happiness Index is we can look across groups and say “Oh, by the way - you’re the 60 sales reps, you’re the new people, you’re the ones that have been around for a while, and we measure the average customer happiness index for every single sales rep. Ok, are you selling happy customers?” If you’re not selling happy customers, you’re doing something wrong. Your job is to produce happy customers at the sales point all the way through.
We look at marketing channels. When we sell through inbound lead generation through our blog, if we get a customer through that channel vs that channel, does that produce happier customers or less happy customers? We look at our support team. If someone talks to these three people vs those three people, are they more or less happier? And the idea is to have this single number that you can adjust over time because what goes into the number changes probably about every 4-6 months.
The three that it boiled down to for us was:
1. Frequency of use. If they used the product every day, every week they were more happy than if they were just using the product every two months.
2. Breadth of use. How much of the product are they using? Is it just one feature or are they using seven things?
3. Sticky features. This was the key one. There are certain features you will find that have an exceptionally high correlation to happiness that says if people use this feature, even if they log in only once a week, they behave as if they’re logging in 17 times a day. They have the same pattern. There’s something about that particular feature that causes them to stick. And the weird thing is you won’t know what that feature is. Your product management group won’t know - you won’t know until you actually sit down and measure it. My advice is, even if it’s coarse, you’re logging all this data anyway - try and measure it and try to get better over time. It works wonders across the company. If I could point back to one thing that says what made HubSpot tick, this is way way way high on the list. So, I encourage you to do that.
The other hack that we did. My first start up, we had this at all our management meetings and founder meetings. We had an empty chair in the room - the designated chair for the customer. And we would literally pretend there was a customer in that chair when making decisions. And so, ok, if we came up with something - like David said “if the nickel stands right on it’s edge and you have to make a choice, how do you decide which way you’re going to fall?” Because we make these decisions all the time - they’re non-trivial decisions. Do we raise prices or lower prices? Do we do this feature or that feature? I found it very helpful in the first start up to actually have that chair, as if we had a customer here who was savvy, who we trusted, who let’s say had a stake in the company, what would they say? What would their vote be?
And so at HubSpot, we took this one step further. (Laughter) We actually have a stuffed bear; her name is Molly. Our marketing persona is called Marketing Mary and we have Owner Ollie and she is Molly. She goes to all board meetings, all management meetings, and any other meeting that someone wants to invite her to. She is required to attend board meetings and management meetings. Not a single meeting will go by where someone will not say, “That’s bullshit. What would Molly say?” It sounds cheesy but if you think about it, you can have really smart people and debate things for two hours and it becomes crystal clear sometimes that yeah, there’s nothing in it for the customer. Molly would vote no on this particular thing. If we’re evenly split, let’s go with the customer. So that hack - turns out it works. It’s a very simple thing to do. We like it. I’m not suggesting that you give customers some sort of veto right. You’re in the software business. Part of what you deal with is that the customers are often very very good at identifying their problem, not necessarily good at identifying solutions for those problems. So, I’m not suggesting they have veto right, but they - just like everyone else on your management team - they are going to be wrong, they are going to be opinionated. Pretend like they are actually there. What would they say? At least let them be heard.
The other thing that we’ve learned is this. Most things include batteries these days but I grew up in a time when that was not the case. So you would get a gift on Christmas Day, and you’d open it and you’d be all all excited and the thing didn’t have batteries included. So that moment of happiness that was possible just doesn’t happen. It’s like, “wow, crap, that kind of sucks”. And I’m from India and so we don’t buy batteries. It’s like, the answer in an Indian family is always that there are always batteries somewhere. You just have to look hard enough to find them. We don’t buy new batteries, which defies the laws of physics because someone has to buy them at some point but anyway... My mom’s answer is “there are batteries in the house somewhere; don’t buy new batteries.” But the lesson here from a software business perspective goes into this notion of services. So I may, deep down inside, however many layers you look, I am a product guy. I just am. And there’s a very simple reason for that. The margins are better. I can write a piece of software one time, sell it to 1000, 10,000, 1,000,000 people and my cost of delivery is relatively low. You guys all get this, right? It’s simple! That’s why we don’t like services. If I had a choice between the two, I’d sell product because the margins are great. One thing we have learned is that I think services sometimes get a bad reputation. I’m going to give you some anecdotal evidence here. So, let’s assume right now that HubSpot is measuring, which we are, CHI for customers, so we know what their happiness is. We don’t have to wait for them to cancel; we can actually tell you with relatively good precision, not only how many customers are going to cancel next month based on the patterns of using the software, but which ones they are. Like, “Oh, you have a CHI of 22. You don’t know this yet, but you’re unhappy and you’re going to cancel 14.2 days from now” or something like that. (Laughter)
Services represent 7% of HubSpot’s revenue. And it’s a break-even unit within the company. So we don’t try to make profits, and we don’t try and lose money. We essentially solve it for break-even. But a very interesting thing happens when you look at services from a software business perspective. Here’s what happened at HubSpot. So we’re like ok, well we don’t really like services all that much - it drives the margin down because on some portion of our business, although it’s relatively small, we make zero margin and that, if you average out, is our gross margin for the entire company goes down as a result of having some portion of our business as 0%. As it turns out, that was the wrong way to look at it. And here’s why.
The way we looked at it was - “Well we charge this much - we charge I think it’s $125/hr for our services people”. Right, so we don’t give away services for free and they help train onboard customers and bring them up to speed. And a customer will spend between 4-8 hours with a service person at HubSpot. It’s paid support, essentially. And, at first we thought “oh, it’s a break even business” because we can look at the $125, we know what our fully-loaded costs are, we know what the employee salaries are for the people that work on that team, and we’re like “Oh, it’s break even”. As it turns out, it’s not break even. Because what it doesn’t capture, the “Oh, this is the revenue coming in the door,” this did not capture the increase in Lifetime Value. Because we did the research. We sat down and looked at our data and said, “Oh, you know we’re going to spend a month or two where we don’t provide any kind of on boarding services”. We’ve got all this documentation, but let’s see what happens, because before, every single customer that signed on at HubSpot would have some human help them for between 4 and 8 hours and it was sold as part of the product. And so we retracted that for a while and said “Let’s see what happens if we don’t have services.” As it turns out, absence of services led to much unhappier customers. That seems reasonable. It’s like, “Ok, well lots of people don’t know what they are doing, this inbound marketing stuff is new, the product’s not that easy but we try”. And so it turns out though we measured the actual happiness index lift, which we know is correlated to turn, which we know how to calculate the economic value. With every hour that we spend on services, we know the increase in Lifetime Value. That says these customers are now going to say 7.2 months longer as a result of their increase in happiness as a result of us having spent the time. We have the data. And so then we took it the other way. It’s like, “Oh, we’ve been spending 4 hours per customer. What happens if we increase it to 8 hours?” Are customers even happier? The answer is yes. The more time we spend it seems that they are even happier. And obviously there is a limit to this. I’m not suggesting that you spend an entire lifetime with the customer. There is a point at which that system breaks. But the larger message here is question the assumptions around people.
Question from the Audience: Inaudible.
We tried that too. And yes and no. So we tried it and we almost decided to do it. The question is: is it so valuable that you should just give it away for free because of the increase in Lifetime Value? So, when we have these consulting sessions, we have 4 hour sessions that a very very smart human on our end talks to the customer and gets them up to speed on marketing and the software and things like that. What we learned was our hit rate of actually scheduling time with the customer went down when it was free. Because they didn’t value it. If they paid $500 for 4 hours of consulting, by God they were going to consume their 4 hours of consulting. So the way we look at it, is that anything we can do to guarantee ourselves that they will actually take a benefit of the services, we are all for that. It’s worked very well for us. By the way, I’m the type of person - I don’t like humans all that much. If there was a way to build a multi-billion dollar business out of my basement, not hire anybody ever, I would do that. As it turns out, that’s really hard (Laughter). I’ve tried. It’s really, really hard. It takes people.
Once you have this Customer Happiness Index, by the way, then you can run lots and lots of different experiments - some of which we talked about. You can also say, “Well, what happens when we go to, instead of having this one on one consulting, what if we go to like some online chat-based, video-based, content-based model? What impact does that have and we can kind of calculate.” It’s like “Oh, it’s going to cost us $50,000 to produce this video or do this thing.” And we can actually measure true ROI because we know relatively quickly within a week or two, what the impact was on happiness, which is once again, correlated very very well with Lifetime Value.
Question from Audience: Inaudible
Very Positive. So, it’s interesting. I was dubious, by the way on this particular experiment. I thought there is now way that humans are going to sit and watch something vs actually talking to a human. The sample size is not large enough for us to have confidence in it yet, but the early data suggests that the Customer Happiness Index for those that go through that One to Many process within one customer segment is actually marginally higher. And we tried to dig into why that is because it makes no intuitive sense. It’s like, “Why wouldn’t you want to talk to a human if you could?” But so far, it’s positive. It’s working out. Only within a certain segment, but it’s working. So we’re going to invest more. The other thing that works is this. Let’s say you have this CHI and you knew exactly which customers were going to cancel, and you knew that if you could keep those customers – their Lifetime Value is, let’s say in HubSpot’s case $20,000-$30,000 – call them. And say, “Hey, I think you’re unhappy. What can I do to make you more happy?” And that is essentially the script. We have an entire team of people at HubSpot called Customer Success Managers that do nothing but run the report, look at the bottom lowest 2 customers and call them and see if they can convert them. Like, “It’s clearly our fault. What did we do and what can we do to help?” And we’ll do things like give away another 4 hours – we’ll do all sorts of stuff to make a customer happy because we know what the value is. And the question you would ask if I weren’t talking so fast is, “Oh yeah, but those customers that you save- do they ultimately end up cancelling anyway?” And the answer is yes, about 1/3 will cancel because there is something intrinsic about that relationship that just wasn’t working, and just us calling them and making them happier is sometimes just temporary but in other cases it works so it is still profitable.
Question from the Audience: Inaudible
Yes. Great Question. The Question is, “by not investing in the sales team, doesn’t the overall experience get worse because you have fewer sales people?” We have a great consultative, relatively sales process. My answer is that nothing at HubSpot or in any company, is forever. Everything is one grand experiment that we happen to be this far along in the process at a particular point in time. So our decision to stop hiring sales reps and investing it all in product, will endure for at least 6-9 months. That’s the current plan. And then well look at the numbers. We have certain goals that we want to get to in terms of customer happiness. And once we get there, then we are kind of like “Ok, tweak the engine. The car is doing relatively well now. It’s not shifting as much anymore. Let’s go faster.” And then go back into sales and marketing mode. So it’s that balance. And it’s like every quarter there’s like a little thing like, “Oh, we did this over here and this broke over there. Let’s go fix that.” So it’s this kind of iterative optimization process. So I fully expect that we will go back into getting more salespeople and invest more there.
Question from the Audience: Inaudible
Because we’re geeky. And because we don’t like multiple variables. We don’t like two experiments at the same time that might impact each other because we’re so smart that then people pick whichever experiment and data that they like that kind of validates their thinking. And we like to point to it - it’s like, “Oh, if this is the only thing we changed, and life sucks,” clearly there’s got to be some correlation. It’s not always definitive but …
Question from the Audience: Inaudible
So, the question is - HubSpot’s got a bunch of cash in the bank. We’ve raised a bunch of capital. And doesn’t that give us the opportunity to run more of these tests and do more of these experiments? And the answer is yes, that’s one of the upsides to having capital. It’s hard as a bootstrap to run a lot of tests, because you’re looking for revenue – you’re looking for customers. You can’t say, “Well, we’re going to turn off the revenue spigot for two months and see what happens, because we’re curious and we’ll see” As it turns out, employees don’t like that too much sometimes. (Laughter) I’ve done that before in my prior startup. So yes, the cash does help there. So my message to you on the capital raising is that if you have relatively clear, precise visibility into your funnel – into your overall business process and what the economic drivers look like…
For instance, we can tell you - here’s how much we pay a sales rep, here’s the month that we will become cash flow break even because we know how much they sell, we know what the customers are worth. And as the data grows, we become increasingly confident. It’s a very predictable business. So if you look at that curve, that curve, the sales curve I showed you? The reason it’s so smooth is not an accident. It’s because we solved for that curve. It’s like, “Ok, we want revenue to be ‘x’ next month, and the month after that, and the month after that.” And we tweak the business to get to that curve. And so, if you have a business where you know that pouring a dollar into the business from a sales and marketing or whatever perspective, yields $2.00, $3.00, $4.00 of lifetime value – raise money. Just enough to run the experiments that you need to because you should be pouring more water into that machine. You’re producing cash. Investors love that. (Laughter). If you can put a dollar in and get three dollars out, that’s a perfectly good reason to raise money, by the way.
Cathy ____ gave a brilliant speech at a presentation last year in San Francisco, and the one thing she said which resonated with me because I think it happens to be true, is that we shouldn’t think about being producers of ‘x’, we should be thinking about being producers of brilliant users of ‘x’. So, let’s use HubSpot as the example. So I would say, “HubSpot – we don’t make marketing software. We make marketing superstars”. Her example was if you’re a digital camera person, you don’t make the best digital camera, you make better photographers. And it seems squishy. It seems like, “Well, yeah but don’t we really just produce software and sell that?” Yes you do, but as it turns out, as a kind of motivating, aligning, sounds philosophical… it works. Because it’s that kind of customer alignment – the software is the vehicle. Seth said this - “It’s not about the code. The code is not the point.” The customers are the point. The market is the point, essentially. Like, “What are you doing to produce better users of ‘x’”?
So we want to produce a million better marketers, essentially. That’s our mission – that’s what we want to do. Software is how we get there, but that’s what we want to do.
The other thing that we’ve learned is this notion of “simple. So, when we started the company, our whole idea was to have this kind of simple, easy and integrated platform for marketers, because there were much better products and each are individual. For example, we have a content management system. Wordpress is much better. We have an analytics tool. Google Analytics is much better. For everything that HubSpot does – and we have like 19 features – there are venture backed and in some cases publicly traded, phenomenally successful, great companies doing just that. And you would think it is completely idiotic for any entrepreneur worth their salt to ever do something like that. It’s like, “Why would you do that? Why would you have something that competes with Wordpress and Google Analytics?” All together of all things! And the answer, actually, is inspired by Apple.
___________, from Harvard Business School writes about this a lot. But the idea is that if you’re going after massive opportunities, you should not be trying to take market away from your competition. If the opportunity is big enough there are enough non consumers out there that you should be selling to. Look for the blue ocean. Pick your cliché or phrase of choice, based on which business school book you read. (Laughter) But the concept is actually very simple. So, what Apple did was – when Apple released the iPod, they did not say, “We’re going to create the best mp3 device with the highest gigabytes of storage, with the best cost performance ratio”. They said, “We’re going to go after the non consuming mere mortals who are not enjoying digital music but should be.” And Apple has been brilliant at this for most of their entire history as long as Jobs has been in charge. Apple asked themselves a question: “What do we have to do get those millions of people that are sitting on the sidelines that should be enjoying digital music, because it’s a better way to enjoy music – what do we have to do to get them into the game?”
And as it turns out, more often than not, when you ask yourself the question, “What do you have to do?”, the answer is not “build a better ‘x”. The answer is “build a simpler ‘x’”. What’s keeping people out is not some feature. What’s keeping people out is because it’s too hard to get into the game. So Apple said, “Oh, we’re going to build a simple device. We’re going to have partnerships with the content producers. We’re going to have this network essentially. We’re going to have this way you can download song, we’re going to take out all the copyright and we’re going to do this. And it’s going to allow millions of people to enjoy digital music.”
And so, as you go back and you think about your businesses, I think that in just about every industry there are always opportunities – blue oceans of un-served customers that should be enjoying whatever you have to offer but aren’t yet. And that’s a much happier market to work in. It’s frustrating sometimes, but it’s a much much happier market to work in. So I would suggest that you think about that, like “What can we do to simplify?” Like, Southwest didn’t go after people that were flying on Delta. There are all sorts of examples of great successful companies that essentially went after a broad, entirely new market and took new people and got them into consuming whatever it was they were offering.
The other thing we’ve learned at HubSpot. This is a culture hack. Transparency. A short story: About a little over a year ago, my co-founder and I were chatting like we do ever now and then. And we said, “Ok, we’re growing – I think we were about 70-80 people, and we should see whether the employees are happy. We do customer surveys all the time. Let’s check in with the employees”. We had never done that before. We’d never asked anybody at the company, “Are you happy?”
And so we did. And we said, “Ok, well let’s go do the employee survey and let’s ask people.” We did a Net Promoter Score which is essentially 2 questions: 1) on a scale of 1-10, would you recommend HubSpot to a colleague? Would you recommend that they work here? And 2) Why? Which is the qualitative, subjective side to it. So we learned two things:
1. Employees were happy. They were exceptionally happy. And one of the things that bothered us, quite candidly, is that our employees were happier than our customers. Hence, this “take dollars out of sales and marketing and put them into product and create happier customers”. But the employees were really happy and the answer to the question of what made them happy was – the other employees. Like, “I love the people that I work with. They’re really really smart, they care”. All those kinds of things that sound cliché. And that’s not the point of the story. The point of the story was that a little lightbulb goes off with my co-founder and I and we were like, “Oh! We have a culture!” That was the first time the word “culture” had ever been mentioned in the history of the company – ever. It had never been spoken before. It’s like, “Oh we have a culture and it sounds like it’s pretty good. It’s working, people like working here and we’re producing revenue.” And our whole job, essentially, is to try to not screw it up – like HubSpot’s doing well, let’s try and manage not to screw it up. And one of those ways not to screw it up, is – we’ve got this culture. We probably should try and do something to preserve it. We’re like “What is it? I don’t know!” We had never dealt with this before. And since I’m the one that likes people the least, I was the designated person to figure it out (Laughter) in terms of employee culture and human interaction.
So went and asked people and one of the things – this has worked really well – and we’re big on this one – is transparency. So we identified our cultural attributes and I’m not going to bore you with all of them. One of them is transparency and it’s on the list of “things”. We are transparent to a fault. Especially within the employee base. So every employee at HubSpot – and we used ______, by the way. We have like 4,000 pages and tens of thousands of comments. Every day, every employee is on the Wiki at HubSpot. But the information that’s available on the Wiki is how much capital we raised, what price we raised it at, what the right price was, how much cash is in the bank – literally, the bank balance of the company, how much money we burned last month. How long we will last. How long is the runway? Everything is there. Everything we present to board meetings. All the things investors know, every single employee up and down the chain knows that at HubSpot. The only thing we don’t share is salary information of the employees. Absent that, just about everything is open.
Question from the Audience: Inaudible
We like employees to use their discretion. We’re not thinking like a public company yet, so we probably share more stuff than we should which is ok. We’ll probably have to change that some day. But there’s no formal policy that says, “Here’s the stuff you share, here’s the stuff that you don’t share”. Except I think the board meetings. I think there’s something like if 5 years from now, we go public, the board meeting minutes go into the IPO docs or something like that but I think other than that most of it’s relatively fair game. But they use their discretion. It’s like, “Ok, if you think it will help, and it’s germane to the conversation then it’s ok.”
That’s been very helpful for us for a couple of reasons. One is a quote that our VP of Engineering has used in the past, which is “Light is the best disinfectant.” As it turns out, management, broadly, including me, does much fewer stupider things when those stupid things are publicly accessible to everyone. Really stupid stuff happens behind closed doors. You make all sorts of decisions, like “Oh, we’re going to do this, we’re going to change this”. And we tried this, by the way. You would think we would learn by now. My co-founder and I will put an article out there, “We met the management team, we had this long all-day offsite or whatever, and here’s what we’re looking at”. And we’ll have 8 pages of comments within 3 hours. Like, “That’s the stupidest thing we ever heard” kind of comments, essentially.
Question from the Audience: Inaudible
Yes, I like to jokingly say when we initially introduced transparency as one of the things, it was because not being transparent… I call my co-founder lazy for not being able to make stuff up. And transparency actually consumes much less calories. Because then you don’t have to decide, “what do we disclose or not disclose?” Just put it all out there – we’ll work it out.
Along the same lines in terms of culture, one of the hacks that we put in last year, which is controversial….
So we had this meeting of the exec team at HubSpot and we were 80-ish people and we didn’t have a number of things at HubSpot. We don’t have a director of HR or anybody with the word “HR” or any kind of creative title. Nobody responsible for employee happiness at Hubspot. Yet, we were voted best company to work for in Boston. We beat Google this year. So something’s working.
One of the things we did last year was a culture hack. So, we had this meeting and our CFO comes to us and says, “Guys, we need a vacation form and a system to track because nobody knows how much vacation time they have. Whatever it needs to be is fine. We can keep it simple”. So we just looked at each other and were like, “Well, why do we need to do that?” And he said, “Well, you know, because employees need to know”. And so we decided our vacation policy at HubSpot literally is we have no vacation policy at HubSpot. We don’t track it, nobody approves it. You take the time you need and that’s it. And our hope is in both directions that people don’t abuse the system. And so the common argument that comes back is “My God, that will never work” Ok, if you’ve got people that you’re worried are going to abuse the no vacation policy policy, you hired wrong. Fix it. My hope is that doesn’t happen.
Who watches Mad Men? Great show. Those who are not, you are missing out. You should watch it. Go back to Season 1 and work your way through. So Mad Men is set in the 60’s about a Madison Avenue firm. It’s obviously a bit of a caricature of those times, but we can’t believe they ran businesses that way. Drinking vodka in the afternoon and mistreatment of women in the workforce, and just all sorts of stuff. And we look back on it and it’s like, “My God they were idiots. How could they not have known?” My question is – 10 years from now, if someone did a documentary on businesses as we are running them right now, what will people make fun of, then? What are the things we are doing now that are going to look idiotic 10 or 20 years from now? Because we’re hoping to build a business that outlasts us, essentially. That’s the goal. That will be around after we’re dead. And so we want people to point back and say, “Well at least they questioned it, if nothing else.” And we still do stupid things. Some of them deliberate and conscious. So that’s the no vacation policy policy. And we’re still alive, by the way. We’re still making our revenue numbers, no one has abused it. I’m still standing. I still have a job.
Question from the Audience: Inaudible
Very good question. Because we haven’t been able to come up with “What’s the upside”. So, like on the employee transparency – I can relatively pretty well argue why it’s important for all the employees to know essentially what’s going on. Because it makes them better decision makers, makes them feel more bought in. I’m not sure that, as a red-blooded capitalist where the upside of sharing it publicly is worth it. But we do take our customer numbers and essentially our revenue numbers are out there. We have our big customer conference tomorrow so we’re pretty transparent but just not to the same degree. But that might change. We’ll see.
Let’s talk about Free a little bit, in terms of Freemium. It’s all the rage in the software business. And if you’re doing a Freemium model, by the way, don’t forget the “meium” part – it’s not just about the “free”. (Laughter)
I want to just say one quick thing about Freemium. HubSpot doesn’t have a classic Freemium model, but we have lots of free stuff – content free tools, etc. but it’s not the core product – yet. But in talking to a bunch of entrepreneurs, which I do a fair amount, one of the things that troubles me is that one of the traps that we fall into as business people is that once you have a Freemium, when you’re making decisions, you start talking about these kinds of “triggers” and “traps” and “What can I do to cause a higher percentage of people to move from ‘free’ to ‘pay’?” And that’s a legitimate question to be asking – “What can I do to make more money?” Like, “Oh, only .2% of our customers are paying and the industry average is 2.2%. How do we make that better? And that’s ok. What’s not ok, though, is the way in which you frame those decisions. So it should not be “Where can I lay these little triggers, these little wires in the jungle – oh, I was doing this and now I tripped – and now I’m a paying customer, I have to be”. The idea is not to trip the user base. The idea is to create or manufacture the value in a way that your customer triumph results in them switching. And this applies to upgrades as well.
This is going to sound philosophical but there is a difference between laying little trip wires in the sand, like “Aha, we got you! You were this and now you’re that and you need this feature”. The company that I think has got this is – and I love the company - salesforce.com. salesforce.com – brilliantly successful. My hope is that we don’t have to this to be brilliantly successful. That remains to be seen - but they are diabolical about those traps and trip wires. So you can start as a salesforce customer paying $500 a year or something crazy like that and then all of a sudden you wake up one day and there’s this big ass non-linear shift. It was not an accident, by the way. As you were meandering through the woods of salesforce usage, and then you fall over this one little thing, and now all of a sudden you’re paying $20,000. It’s just crazy. They’re diabolical. My suggestion to you is you don’t have to be that. You should be thoughtful in terms of where you know you’re creating value and kind of associate the price the customer pays for the conversion to paid.
So, I’m not a “brand” guy, but I like this quote I heard at a conference at MIT of some sort and it was a table away from me so I can’t attribute the source, but it’s about a brand is what people say about you after you leave the room, essentially. That’s like the layperson’s attempt to describe brand. And where it helps is it makes everything else easier. Like Dropbox. I asked Drew, by the way. He was in town last week. He’s one of 5 people - I will cancel dinner with my wife in order to have dinner with Drew from Dropbox. (Laughter) And my wife knows it and she encourages it because she’s a user and she’s like, “Well ask him why they haven’t built this feature yet – why can’t you share folders?” Which I did, and they’re coming, by the way, for the record. (Laughter) You didn’t hear that from me - technically that’s still confidential. (Laughter) So I asked Drew, and I’ve asked him this before, which is “Drew, I totally get that the reason Dropbox is so phenomenally successful is the product.” Because he’s told me this 100 times. “Dharmesh, it’s the product. Build a product.” So I asked him this time, because I’m slightly smarter now, “Drew (and I’m very crafty in the way I structured this question) next to the product, what’s the next most important thing that contributed to your success?” And his answer was, “Don’t screw the customer”. That’s it. When they sit down and make decisions, Drew says, they do not make decisions where the company benefits and the customer suffers. That’s it. Don’t do those things, essentially. After you get the product right, then secondly, essentially, solve for that customer happiness. It sounds squishy but it works.
How am I doing, by the way? (Applause)
So, the “be a good egg” thing is around what I call the “path of truth and justice”. So at HubSpot we believe we are on the path to truth and justice because we believe in this inbound marketing and stopping SPAM and direct mail and killing trees and harming kittens and all those things. (Laughter) And it’s true. So we really believe we’re making the world a better place. But I’m going to pause it to you that today versus the 80’s or 90’s- if you were in the software business, it was sort of ok (not ok as in ethically ok, but profitable) to be somewhat evil. Like Oracle (Laughter), a bunch of relatively successful software companies. My apologies if someone from Oracle is in the room. But – supremely successful software company – and it worked and the reality is you had all sorts of kind of misbehaviors because the market didn’t punish bad behavior that well because we didn’t have an opportunity to. So you had drive-by sales, essentially, like “Oh, yeah, here’s a $600,000 _____implementation or whatever that never saw the light of day” and very aggressive sales, lots of bad things. And my argument to you now is that if you’re really looking to build the next multi-billion dollar business, bad behavior gets talked about now and it will crush you. Becaue of the Internet. People can search on “HubSpot sucks” in 30 seconds and find out how many people think HubSpot sucks. You can’t hide anymore. Customers don’t have to go to a conference to find out, “Oh yeah we tried that product from that company and it kind of sucked”. It’s all open right now. SO I think the red blooded capitalist, right thing to do, is to put yourself on the path of truth and justice and solve something where you can say, “Well I’m actually solving a problem. My customers are happy. I care about that. And my hope is that that leads to a multi billion dollar company some day”. My answer is that maximizes your odds. If it were me, that’s what I would do.
Question from the audience: Inaudible.
Yes, this question is about salesforce. What about salesforce? Of all the things that keep me up at night at HubSpot, high on my list of questions that keep me up at night at HubSpot is the “What about salesforece?” Not from a competitive perspective. So if you look at the practices of salesforce – I talk abou their pricing and it’s clever and diabolical and whatever. (Laughter) I met with at least half a dozen people from the original executive team at salesforce. Sat down and talked to them and said, “Oh, yeah, what did you guys do about this?” They are very transparent as it turns out. Very open company. But there’s a bunch of stuff at salesforce that I just fundamentally don’t agree with in terms of corporate culture and aggressive sales tactics and all that kind of stuff. They built a great product and a big business. I wish them all the success because they did a great damn job and laid the groundwork for a lot of us. My hope is the successful software businesses of the next two decades will be kinder, gentler, and more focused on customers. So I’m going to close on this note, which is, as a former bootstrap entrepreneur (this is my third one), is… and there’s been lots of discussion around “Should we take venture capital? Should we dream big?” My response to you is that dream big and execute small. What I mean by that is this: If you are starting a company, for those of you who are entrepreneurs in the room – if this is your first one, it will not be your last. It’s a genetic flaw and you’re going to do it over and over again. (Laughter). And the way to think about this is that your big dream does not have to be embodied in your current company. For instance, it’s completely ok to say, “Oh, I’m going to do this bootstrap startup. Someone is going to offer me 15, 20, 25 million dollars. I’m going to take it.” And there are lots of people who argue, like “Oh, you didn’t dream big enough and you should have held on and it was growing and why not…”
Take the cash. (Laughter) Take the cash and here’s why: Because the cash funds the next dream. And it’s ok. You will have plenty of ideas, I promise. Your issue is not that you don’t have ideas. Your issue is you have too many. And it’s hard to tell the crappy ones from the potentially good ones or the great ones. I’m going to close, take a breath and I’ll take a couple of questions until Neil pulls me off the stage. Thanks for your time, by the way. Always fun.
Ok, two questions.
Question from the Audience: Inaudible
The questions is this: We have this customer success group that calls customers that are low CHI, so we how do we track the data. We’ve got about 18 people out of MIT on the team – we’re very data geeky. So everything that happens, essentially we try and capture so we can analyze turn and things like that so we know that, “Ok, so the happiness of someone that we kind of “re-happified” or “re-delighted” even though they weren’t, has this kind of behavioral pattern, so we just measure it. And so it’s still profitable which is why _______. That team is still growing, dollar in produces more than dollar out which is a good thing for us to do. But it’s not as good as we thought it was because there are some intrinsic issues . Like, there are some customers that just don’t fit the profile. Like, no matter what we do, there is a reason they are unhappy. Some of it’s us and we can fix those things, but that’s not the majority of the reason. So we have about a third-sh of them that we can save, but it doesn’t pan out. There will be others unhappy again in 6 months or something like that. We’re trying to figure out now from a demographic perspective, “What’s the proof? What did those customers look like?” Even though they had a low CHI, they are not worth trying to save, essentially, because we’re just not the right fit is what it comes down to.
Question from Audience: Inaudible
It would be. And should be. The question is, “Why talk about average revenue per user and not average profit per lifetime or profit lifetime vs profit value?” And we do. So the answer is, you should definitely look, if you can, at lifetime profit. Because then you can factor in cost of goods sold. It’s like, “Ok, well here’s what goes into it”. So that’s the ideal number that you should be measuring. We use proxies for that. And revenue is good enough for us and we’re not profitable yet. We will be here. And the other thing is that with a software company, the cost of goods sold ends up being not that big a piece of the equation overall so it’s a relatively ________.
I’m done. Thank you very much for your time.
---end of session--