Thanks, everyone for having me, James in particular for inviting me. It's been a real pleasure to see Heavybit grow into the huge success that it is today. I'm thrilled to be here.
As Tom mentioned, I like to invest,or we as a firm like to invest in software-as-a-service and platform-as-a-service companies. The great part about those businessesis that they tend to be very technically focusedand they tend to be started by engineers a lot of the time.I'm also an engineer, but what I knowabout an engineer's mind is they like to make decisions,or generally engineers like to make decisions using data.As I started to work with a lot of these businesses quite a few questions come up throughoutthe progress and the evolution of these as-a-service businesses, as to how you developthe business side of the business in additionto the engineering side, which seems to bethe expertise of a lot of the founders.
My goal in this talk is really to kind ofshare with you a lot of the answers that I've developedfor founders that I work with as to how youbuild the business side of the businessand what benchmarks exist in the market that you can useas you build your developer-facing services or platform-as-a-service businessesor software-as-a-service businesses.That's where we're going to get started.If there are any questions throughout the talk, feel free to interrupt.This is meant to be totally conversational.
First things first, every business is a snowflake, every business is unique.These benchmarks aren't rules.They're not recipes for success.They're just data points. And I'm happyto walk through the biases, and the sample sizes,and the details of all those analyses; but I just want to get that out front.
Revenue growth is really kind of the ultimate goal ofmost businesses that we invest in as venture investors.
A very simple way of thinking about revenue growthis you grow revenue by increasing your customers,by increasing the average amount you chargethose customers, and by maximizing the differencebetween the growth rate of the businessand the number of customers who churn. What I've done is I'm going to structurethis presentation around this equation, because we're all engineers, and we'll talk about each component.
This is a chart. On the Y axis you've gotannual revenue, revenue growth, and on the X axisyou have the log10 of enterprise value in dollars. These are for all the publicly tradedas-a-service businesses in the market today.And you can see that there's actually a pretty strong correlation.There are about 36 publicly traded as-a-service businessesand there's a really strong correlation, actually the R2 is about 0.56, between revenue growth and the value of the business.So what does that tell you?
It means in order to be a valuable businessyou have to grow really fast.
The most valuable businesses growaround 200%, in the last year, and that includes businesses like Workday,Viva, which just went public,FireEye and a bunch of others. This is why revenue growth is so importantand that's why I start with it in the equation.
If you look at those companies on a cohort basisby number of years since founding,this is the revenue growth actuallysmoothed over the life of those businesses.They started around 0 and then within 5 yearsthey're roughly at around $75 million on average. And so what we want, I think, as foundersand as investors is to find companies thathave these kinds of characteristicsand to** inform** those businesses and foundersso that they can create these kinds of businessesthat grow really fast so thateveryone in the company benefits.
Let's first talk about customers.A lot of the businesses that are in this roomand in Heavybit are freemium businesses,and at the bottom of the freemium pyramidyou have freemium marketing. And that freemium marketingcan be any number of different things.For Zendesk, it was really community marketing.For Expensify, it's app store. There are lots and lots of different â€” for some businesses like Intercom, it's all done through content marketing.Lots of individual users sign upand then there's a monetization vector.
Ideally at some point what ends up happeningis the business builds an inside sales teamin order to capture the larger customers. This is the customer acquisition modelwe'll be talking about.
One of the** key** questions that comes aroundfrom technical founders is: What are the relativesizes of the sales team and the engineering teamsover the company's life? How much should I be investing in building the productversus actually taking the product to marketand selling it and putting itin the hands of customers?Of the 36 publicly traded companiesthat I showed you at the beginning,this is the ratio between their spend in sales and marketing,and research and development.In order words, customer acquisition versus product.You can see that over the life of the business it's a 2:1 ratio on average for these companies.
For every dollar they put into engineering,they tend to invest two dollars into sales and marketing.That trend continues basicallyover the life of the business.
When we think about what thosesales teams ought to be doing,one of the key questions is: How do Imeasure the success of a salesperson? When I started in the venture businessI didn't know how much a salesperson was paid,what the productivity of that salesperson ought to be,what the differences between an insideand an outside sales rep are.
An inside sales rep, as we talked about before,is someone who they're called inside salesbecause they're not making outbound calls.There's business coming in, there are users signing up,there's Google starting to use your product,or there's Heroku that's starting to use your productand so you call them. You've got a bunch ofdata about that customer and your goal isto close them and convert them and make thempay you more money or pay you money at all. An outside sales rep is avery different kind of salesperson.An outside sales rep sells by being out in the field,by building relationships, by playing golf,going out to steak dinners, all that kind of stuff.
The** sales cycles** for these 2 kindsof sales reps are very different. The typical inside sales rep,and we'll walk through the numbers in a second,but the typical inside sales rep sells about10 customers a quarter, while the typicaloutside sales rep sells about 4 a year, 4 or 5 a year. Let's walk through these numbersand we can kind of differentiate the difference between an inside and salesperson.
The first line here is OTE.That's a sales term that stands for "On Target Earnings." Every salesperson when they sign up for a jobthey're given a quota, which is the amount of moneythey need to sell of the productin a given year or some given period.For an inside salesperson, that's roughly around $300,000 to $600,000 depending on the product.For an outside salesperson,that's $1 million to $2 million; if they hit that quota, then their OTEor on target earnings is their total compensation.Typically for an inside sales rep,depending on the sector, it's around 65K â€” half of that is salary, half of that is bonus. Then for an outside sales rep, that's split differently.Instead of being split 50/50, it's split 25/75 â€” 25% salary, 75% bonus.
You can see the second line is the quota.The third is the average size ofthe sale that they typically make. An inside sales rep often sells somethingbetween 5K to 25K a year in revenue.An outside salesperson is selling somewherebetween $250,000 a yearto multimillion dollar contractsdepending on the seniority of that salesperson.Just kind of walking through the math,that means that the inside salespersonhas to sell 10 customers a quarter.The outside salesperson, I said it's 4 to 5 a year.
Typically what we see with inside sales teamsis their close rate. The close rateis out of 5 customers they call,1 of them will convert to be paid.That's a 20% close rate.For field sales reps, because the field sales rep is much more individualizedand personalized, is around 30% to 50%.And if you back that out, that means you probably need 200 leads for an inside sales repin order for them to hit their quota; you need about 17 for an outside sales rep. The ratio between the revenue they generateto the cost of the sale is about 5:3.
What I'm trying to give you here isa sense of how a sales team operates,how expensive it is, and what the returnon investment ought to be typicallyfor these kinds of businesses.
Again, for those 36 publicly traded companies,they invest a lot of time into acquiring customersand the question is: How much shouldI be spending to acquire these customers? On a cohort basis and averaged outover the entire group, this is the averagepayback period, or also known as the sales efficiencyor magic number, for these as-a-service businesses.The sales efficiency is: How much did I spend last quarterto generate this quarter's incremental revenue? Or, how much did I spend last year in sales and marketingto acquire this year's incremental revenue? If you spend $1 this year,how many dollars do you get next year out of revenue?
The higher the multiple, the more efficientyour sales process is, and actually the inverseof that multiple is the payback period.
How long does it take for yourcustomer to be paying in order for youto recoup the costs of acquiring that customer,also known as the payback period?
The typical publicly traded companytakes about 15 months to pay backthe acquisition of their customers. For startups, best in class tend to bebetween 3 to 6 months, so all the dollarsinvested in sales and marketing can be recouped in 3 to 6 months.
At scale we kind of see businesseswith a sales efficiency ratio between 0.8 and 1.2.The higher that number, the more you want to investin sales because what that tells you iseach incremental dollar in sales generatesa disproportionate amount of revenue back to the business and you can grow really, really fast.And this is an arbitrary or a very, verydeterministic metric of understandinghow performant the sales team is.
Moving on from customers,we can talk about average contract value.Average contract value is justa really fancy word of saying pricing.This is kind of the way to think about pricingfor as-a-service businesses.There are typically four pricing modelsI've seen in the market.The first and the most common is freemium. And a freemium, we all know what this is.Typically what ends up happening is a user uses a product.The value to the end user increaseswith time as you put in more data.
Evernote is a great example.At the beginning, you're notreally willing to pay for it.As you fill in more and more notebooks,the ability to search across that databecomes more valuable and so youend up being paid for it. Typically those products are pretty simple.Most of the time the end user actuallyends up buying it, but sometimes it actuallygoes up to a purchaser. So in the case of Expensify,which is a freemium expense reporting business,individual employees within a companyadopt Expensify and then convince an accountantto buy a site license. That would be a casewhere the end user actually doesn't buy,but it's the controller, or the CFO,or the VP of Finance who does.
And typically for most freemium businesses,the average seat price is somewherebetween $5 to $75.On the $5 to $10 range,those tend to be kind of more "consumery" products,and on the $75 range you have products like CRM tools or kind of real workflow tools.
The next three pricing structures or pricing models, they're all kind of marketing tactics in order toentice customers in order to convert;and the next three are kind of geared at averagecontract values in the $2,000 to $25,000, $100,000 a year range.
The freemium works really well whenyou have a huge funnel of incoming users,a huge number of leads, and your goal isto convert somewhere between 1% to 4%of those leads into paying customers,and the market size is large enough togenerate a big business given that funnel structure.
When you're targeting a smaller set of users,say fewer than 1 million, your goal thenis to use an inside sales teamin order to convert those customers.And the marketing tactics that SaaS companiesor platform-as-a-service companies use, they're threefold.
The first is a limited free trial â€”Salesforce uses this extensively.They first started with a 30-day free trial,and then they tested a 7-day free trialand a 14-day free trial.They found no difference betweenthe 14-day conversion rate and the 30-dayconversion rate. Typically what ends uphappening is smaller free trials for these kindsof products converts just as much becausethe users want to end up either committingto the product immediately or they churn anyway.
Limited free trial works really wellwhen a user is trying to understand a productand will really only understand itwhen they experience it. It's another product like a freemium productwhose user value increases with time.The product complexity for limited free trialstends to be more complex, like CRM tools.Often the end user buys, but sometimes not,and the average seat value is kind of low or medium.As you move up towards the ACVs,you get these different structures. One is an upfront paymentand the other is a money back guarantee.
An upfront payment is: the only way that a customer canuse your product is if they pay you upfront for a year. Those tend to work really wellwhen a user can perceive the value right away, they think it's an absolute must-have product, and sothe product complexity tends to be simple.The end user often buys, so the decision making processis fast and tight, and the average seat price is very highand justifies the cost of an inside salesperson.
Then the last one is a money back guarantee. It's basically a free trial that's subsidizedby a customer paying for 3 months or 6 months,and it works for those kind of categories and products.
This is a conceptual framework fortypically how we tend to think about pricing structures.There's lots of variants, different sales teams might move froma limited free trial to money back guarantee; but these are all the different vectors that we see.
Q: On that last slide, what do you think the value is of Low/Medium/High?
A: I think a low seat value is probably $5 to $25 a month, and then a higherseat value is $50 to $75.Those don't overlap, but that'skind of how I think about it.On an annualized basis, if you were to lookat it on a per customer basis, the freemium productstend to generate somewhere between $500 to $2,000 per user/company per year,and then the mid-range products tend to generatesomewhere between $25,000 to $150,000 a customer per year. Then high would be north of that; so high is anything a quarter of a million and above.
Q: Can you give me an example of the upfront payment where it's an obvious must-have and it's a simple product?
A: Yeah, it tends to be things like LinkedIn would be a really good example. The recruiter product for LinkedIn â€” you cannot try that product as a recruiterbecause you would extract so much value from yourinitial usage that you wouldn't want to pay for it anymore. Often they tend to be social network productscoupled with SaaS products.
Any other questions on this slide?Okay.
Let's talk about growth rate. How do you deduce growth? There are three ways to grow thatwe see in these as-a-service businesses.The first is to grow through sales teams.Let's walk through this table. There are 2 tables here; the first is a hypothetical business.The first line, the customer's line, demonstratesa growth from 100 customers to50,000 paying customers.The average contract value, the amountthe typical customer pays per year, is held flat at 20,000.
You can see how the revenue scales,you can see how revenue churns,and then you can see over time the typical SaaS business tends to churn between 1.5% to 3% of its revenue per month.You annualize that, it's roughly 20%.So they're losing 1/5 of their business every year.That's the leaky bucket.
If you really want to grow,one of the key parts is mitigating that churn risk.
You can see as it grows â€” at a $20 million clip, the businessis losing $4 million every year. In order to just hold its revenue steady it needs that $4 million in incremental business every year.It's a real weight on the business.And there are three strategies to accomplishmitigating this churn. The first is to replace the customerswho've churned with new customers. Here's all the math, then that net is it's actually really expensive to do it that way.
That's because acquiring new customersis the most expensive way to grow.
In order to recoup $40 million in revenueat the 10,000 customer mark, you need to invest about $50 million in this hypothetical scenario,if your payback period is 14 months, 15 months.
The next more efficient way, but still less efficient wayis to upsell your existing customers.You have customers that are leaving, what you want is to grow your existing customersand have them pay more and more over time. This is a model where what you have isyou have a team that's a customer success teamand they use their time in order toget customers to buy more products.And this is a **more efficient **model.We can walk through the math later.
And then the most efficient model isif customers grow by themselves.You guys will be most familiarwith this with Amazon Web Services.This is utility-based pricing where customersjust kind of grow by virtue of their own growthand the incremental investment that you needto invest in order to have thosecustomers grow is basically minimal. It's just product development work. Most of the businesses that we work with,they use a combination of these three.
But, the more that you can grow organically, the more efficiently you can grow, the more capital efficient business you can build and the better valuation you can command.
The last component of the equation is churn. As I mentioned before, the typicalas-a-service business churns between 1%to 3% of customers per month.This is a chart I use with a lot of our companies.It's a cohort chart and it's a Revenue at Risk chart. Every month 1% to 3% of our revenue is at risk,and what we do is we look for the months in whicha lot of revenue is at risk becauseit's not evenly distributed.The sales process actually is very seasonal.
This chart is showing on the X axisthe months since a customer became paidand on the Y axis the revenue that that customer generates. And in the board meetings what we do is we lookat the ones who are coming up on their, say 12-month renewal cycle or their 24-month renewal cycleon their contract and then we double click on the onesthat are really, really big and we makesure that we keep those customersbecause they disproportionately impact churn.
And then the next question is: Well, if our goal isto mitigate revenue at risk, how much should a business spend mitigating revenue at risk?
Let's take a look at the unit economicsfor a business in their first year.If you sign up a business, let's say we've got a hypothetical businessthat's got a customer who's going to spendabout 24 months on a product.The first 15 months of that is the revenuethat we use to pay back the coststo acquire the customer, and then the15 to 24-month mark is what's known asthe contribution margin,that's all the profit. If they were to churn, then the revenue coming from this customer would go away.
But if we were to invest a bunch of time â€” here we've got the same chart and what we're trying to do is we're trying toelongate the life of the customerfrom 24 months to 36 months.We've got the same thing; we've got the first 15 months allocated towards acquiring the customer, the next few monthsis contribution margin, some period of timewhere we actually invest in customer supportand service in order to keep that customer. And then the last part is asaved contribution margin, additionalprofit as a result of having made the investmentand keeping that customer.
And if you do the math, a really greatsimple rule of thumb is you want to spendabout three months worth of a customer's revenuein saving customers.That's what this all kind of boils down to. That's kind of the equation in a nutshell. I just wanted to give you a high level overviewof the way that I look at these businesses.
If you're able to maximize your revenue growthin this way, then you can benefitfrom what's happening in the market today.
This is a chart of the valuations,basically of valuation to revenue multiples ofas-a-service businesses over the last,call it 10 years, 9 years. The 2 different lines are: the red onesare all businesses and the blue ones are high fliers.High fliers tend to be newer IPOsor better known brand names. You cansee that these companies are fetchingmultiples of 20x revenue which is really unheard of in the market. Part of it is there's a lot of enthusiasmaround these subscription businessesand investor really darlings at this point.
The great part about this for entrepreneurs is the valuations in the public markets are oftenvery, very quickly reflected in the private markets,which is why we've seen, and I've seen firsthand,pretty dramatic acceleration ofas-a-service business valuations of late.
I know that was a lot, but that's my presentation.Thank you very much for the time.