0% found this document useful (0 votes)
590 views9 pages

Introduction To SaaS Metrics (SaaShimi Interview Transcript)

Transcription of my interview with Aznaur Midov where Aznaur leads us ably through a structured and thoughtful discussion of SaaS metrics. Listen to the episode at www.saashimi.cloud.

Uploaded by

ramblingman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
590 views9 pages

Introduction To SaaS Metrics (SaaShimi Interview Transcript)

Transcription of my interview with Aznaur Midov where Aznaur leads us ably through a structured and thoughtful discussion of SaaS metrics. Listen to the episode at www.saashimi.cloud.

Uploaded by

ramblingman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

SaaS Metrics

WITH DAVE KELLOGG

independent board director, DK: To me there is one primary


sitting on a board of three SaaS metric - It’s ARR or Annual
companies. Previously, I was a Recurring Revenue. Some people
CEO of Host Analytics and used to say ACV, Annual Contract
MarkLogic. Prior to that I was SVP Value, but the problem is that “C”
at Salesforce in charge of is for contract and could include
customer service applications; things like services or one-time
and prior to that a CMO at setup fees & services. I think most
Business Objects. people have moved away from
ACV to ARR. ARR has a cousin
Dave Kellogg is a technology AM: Tell me about Kellblog called MRR (Monthly Recurring
executive, investor, advisor, and Revenue), which depending on
blogger. Previously, he held CEO DK: I’ve been blogging for over 10 what kind of business you’re in,
positions at Hostanalytics and years. I started blogging at you may want to look at. I think
MarkLogic, and prior to that MarkLogic because our customers MRR is more for small businesses
SVP/GM of Service Cloud at were dealing with the disruption or for a consumer space. If you’re
Salesforce and CMO at Business of social media and I wanted to focused on enterprise SaaS like I
Objects. experience it myself. That was am, I think ARR really is the
actually the impetus for Kellblog. primary SaaS metric. It’s the first
Dave has worked in varied Then over time, it gravitated to thing you should think about.
capacities with companies subjects I was interested in. In
including Breeze, GainSight, fact, the number one reason I AM: How do you calculate it?
MongoDB, and Tableau. He write a post is it’s something I’m
currently sit on the boards of interested in and I want to make DK: ARR is basically a value of
Alation, Nuxeo, and Profisee. Dave sure I really understand it. And your subscription base. If I sign
previously sat on the boards of the best way to really understand you up for $100 a year for a
agtech leader Granular and big something is to write about it. As subscription, you’re worth 100
data leader Aster Data (both got we’ll talk about later this is how I units of ARR. If I sign up five more
acquired). got some of the SaaS metrics people, the same thing, they’re
because people are very quick to 500 units, so a total of 600 units.
Listen to the episode at talk about them over a coffee There are some not as
www.SaaShimi.cloud without doing the math behind straightforward cases; like half-
them. When you start writing a year contracts. If I signed you up
Aznaur Midov (AM): Dave, could post on what churn is and for a 6-month contract for 50
you please say a few words about defining how to calculate it, you units, it gets a little tricky because
yourself? start getting into little details that it’s technically 100 units of ARR,
are very important. but it’s not really 100 unit-
Dave Kellogg: (DK) My name is contract. If you’re doing a lot of
Dave Kellogg, I am currently an AM: So what are the primary SaaS those, it’s actually a case for
independent consultant and an metrics? calculating MRR in my opinion.

1
But if your contracts are actually time, OK, maybe we can put it DK: I think they’re a good thing.
annual, which for most enterprise into ARR terms. But if you want to The fact of the matter is you need
SaaS companies they are, and be an honest person - and to view churn as probabilistic. I
sometimes they’re more than certainly that’s who investors always think of churn a little bit
annual (two or three years), then I want to invest in - if your cadence like rolling the dice. Let’s just say
think ARR is perfectly fine and a of the business is monthly, you instead of going 1 through 12 as
valid metric. could talk about MRR; if your with a dice, it’s just a percentage
cadence is annual, you should talk from 0 to 100. Even better if it is 1
AM: If we exclude services from about ARR. But don’t have a or 0 - renew or not. And if I roll
ACV, how close do we get to monthly cadence and then just go and it is 1 – they renew, and if 0 -
ARR? multiply everything by 12. then I lose that customer’s ARR.

DK: At least the companies I’ve AM: What’s your take on TCV? The fact of the matter is when you
worked at, ACV was ARR plus do multi-year contracts, there are
services. So to me, ACV minus DK: So in my world, there’s ARR fewer rolls. So if we do a 1-year
services does equal ARR. and TCV. ARR is the Annual contract, you get to roll the dice
Although, I think some people Recurring Revenue and TCV is the every year; and if we do a 3-year
missed that distinction and they Total Contract Value. Some contract, you roll the dice once
actually treat ACV and ARR as people will include services in every three years. Obviously, the
synonyms. I think ARR itself is TCV, but I don’t think that’s really more frequently you roll the dice,
more naturally descriptive - the point of the “T”. The “T” isn’t the higher odds you are going to
Annual Recurring Revenue. As the about services versus not services. churn.
services are not recurring, it’s The “T” is about - if I sign you for a
definitionally excluded. So when 3-year commitment on 100 units The other nice thing about long-
somebody says ACV, the first per year, I would call that 100 term deals as we see increasingly,
thing I do is ask: “OK. How do you units of ARR and I will call that there’s a new up and coming
define ACV?” 300 units of TCV. Or, 300 units of a metric called RPO, Remaining
less popular term - subscription Performance Obligation, and they
AM: Some people define ARR as Bookings. are reflected in RPO. Right now,
Annualized Recurring Revenue they’re virtually impossible to see,
and calculate it by multiplying If there are $50k in services there, because there is no deferred
MRR by 12. What’s your take on let’s say it’s three years at 100 revenue unless they prepaid and
that? units per year, plus 50 units of some companies do prepay multi-
services to get you started, they year deals and we can talk about
DK: I don’t like it. The “A” isn’t would call that TCV or 350 units. the pro-con there.
annualized - It’s supposed to be So as usual, in SaaS, you need to
Annual. If you are annualizing say “OK, what do you mean by But if these are non-prepaid, i.e.
shorter duration contracts, to me, TCV? Are you including services or first year only prepaid, so a 3-year
it’s overstating things. If you agree not?” In my world, you shouldn’t, deal where you pay one year at a
to give me 10 units a month on a because my world is about SaaS time in advance, that would be
month-to-month basis, and you companies, enterprise, and 100 units of ARR. In my book it
can cancel at any time - that is 10 investors; and you want investors would be 300 units of TCV, of
units of MRR. I have more trouble to trust you. And if you’re adding which 200 units is RPO. You’d see
calling it 120 units of ARR. I think services that aren’t recurring into it as remaining performance
it’s misrepresenting because there a number that is kind of obligation that’s not short term.
really is no annual nature of our definitionally recurring, it’s not So overall, I think they’re a good
agreement. It’s a month-to-month going to make you look good. So thing.
agreement. I’m a bit of a purist here.
AM: What metrics besides ARR do
Now, if you’ve got 10,000 AM: What’s your take on long- you think are important to track?
customers with five-year history term contracts? Let’s say three or
and everyone’s renewing all the five years versus one year. DK: So, we’ve talked about ARR,

2
which is by far the most that will help you value the school, they taught – and this is
important SaaS metric and by the company. Then the very next very important - the Payback
way if and when you sell your question is how much does it cost metrics are risk metrics. How long
company, it will be valued based to pour a dollar in? There are a is your money on the table? So
on ARR. You may have a services couple of derivatives of this ratio; CAC Ratio is how much you spend
business there, but people will it’s CAC or CAC Ratio. If you spent on sales and marketing, and CAC
tend to look through that services $1 million dollars on sales and Payback Ratio is how many years
business as if it doesn’t exist marketing and that got you 10 of subscription gross margin does
unless it’s either super profitable customers, then you’d say we it cost to get that back.
or super unprofitable in which have a CAC of $100.
case people may adjust for it. But So let’s just stick with our
if it’s running around breakeven, Personally, I prefer doing example. We’ve spent 200 units to
its size is irrelevant to your everything in ratios. I don’t know get 100 units of ARR. Therefore,
valuation. People will just look at how many customers you got, I our CAC Ratio was 2.0. Let’s say
ARR and value your business as a want to know how much ARR you our subscription gross margin was
multiple ARR and how fast it’s got. So if you spent $1 million in 75%. The question is how long
growing. marketing to get $500K in ARR, does it take me to get back my $2
then your CAC Ratio would be 2.0, at $0.75 a year? So it’s going to
If you have to pick two metrics, which by the way is a fairly high take me one year for the first
one should be ARR and the other one. $0.75, two years of $1.50, and I’m
one is ARR growth. With those thinking somewhere like a 2.6, so
two metrics alone you can pretty That’s the difference with CAC somewhere in the middle of year
much figure out what a SaaS Ratio and just CAC - I don’t know three I could pay back. That’s a
company is worth. When you if $100K is a big number or a CAC Payback Period and that’s
move on from them, I think small number, because I don’t where I was saying earlier that 12
bookings are somewhat know how much ARR they’re months is an excellent CAC
important. So I know how big your giving you. But as soon as I can do Payback Period, 18 months is
subscription base is, but how long things in a ratio-tied ARR, I can getting on a long side, 24 is
are the contracts you’re doing? If say that spending $2 on sales and getting longer. In many ways, you
you’re doing all 1-year contracts, marketing to get a $1 worth of can think of a CAC Payback Period
there’s probably a little more risk ARR is a fair amount of money. is kind of CAC adjusted by how
than if you’re doing a 5-year long it takes to recoup it on the
contracts. So, how stable is that One of the best companies I’ve gross margin basis. So in general,
renewing base. ever seen have CAC Ratio of 0.4, your CAC Payback Period in
which is really, really good. I months will always be higher than
The next really important metric would say 1.0 is excellent; 1.0 to 1.5 your CAC ratio times 12 unless
to me is Customer Acquisition is getting on the high side, and 1.5 your subscription gross margins
Cost (“CAC”). I think of a SaaS to 2.0 is going to start to make are a 100%.
company like a “Leaky Bucket”. So investors scratch their heads. A
your leaky bucket full of what? It’s friend of mine who’s a VC once One of the issue is VCs tend to like
full of ARR and every quarter sales said, “I don’t even like to meet what I call compound metrics
dumps more ARR in, and every with companies that have a CAC that take two different things and
quarter the customer success Ratio above 2.0”, which is a slight blur them together so they can
team is trying to put their fingers exaggeration, but people are see all at once - how long it takes
on the holes in the bucket, to going to have a lot of questions if to get your CAC back. Operators
keep water from leaking out. you’re spending more than $2 for like me tend to dislike compound
$1 of ARR. metrics because we want to go fix
When you think of a SaaS the problem. If you tell me your
company this way, when you get The reason for that is people look CAC Payback Period is too high, I
the first two metrics - which is at another ratio related to CAC don’t know if that’s because your
what’s the ARR level in the called CAC Payback Period. You CAC is too high or your
bucket, how fast is that going off - know, when I went to business subscription margin is too low.

3
So I think you’ll find different Well, then it’s very hard to argue Net Churn. So when you’re talking
folks like to look at these things that they are not in sales and about churn, are we talking about
different. Another one would Net marketing. In that case, they have gross or net, meaning what’s
Expansion or Net Churn. I always to be, because at most SaaS going in the numerator? The
like to look at gross and upsell companies 30% of your new ARR leakage in total or the leakage
because, again, which problem do might be coming from existing before upsell? And by the way, if
I have - I am I not selling enough customers depending on your size you have more upsell than
to my base or am I leaking too and scale. leakage, then you have “negative
much out of my base? And if you churn.” So if I have 10 units of
give it to me netted out, I can’t I think you need to stop and think, leakage in 20 units upsell, I’ll have
tell what the issue is. So we’ll talk “Wait a minute, there’s no formula negative 10 units of churn.
about that too. but what do our customer success
people do because the intent of The denominator, unfortunately,
AM: Regarding CAC calculation - the ratio is to measure how much is also complicated in churn
how do you look at the Customer it costs to put $1 of new ARR in because if you look at it at the
Success versus Sales and the bucket”. If they’re helping to first order, you’d say, “Well, how
Marketing? put new ARR in the bucket, then high is the level in the bucket?”
they should probably be included You might say, “Well, I’ve got $10
DK: Yeah. All these metrics seem in the cost. If they’re not, then million of ARR in my bucket”. If
real simple, right? CAC ratio was they shouldn’t. the company only does 1-year
just sales and marketing cost deals, then all $10 million of ARR
divided by a new ARR, right? And AM: You briefly mentioned churn is definitionally up for renewal
then the question is, well, wait a earlier. Could you please define it every year.
minute: “What’s the sales and for us and explain the
marketing?” And the answer was – significance of the metric? So I can just say, “OK, in a year,
“well, just sales and marketing”. you leaked out $1 million. You had
“OK, that’s pretty easy, but what DK: Yeah, let’s speak about churn. $10 million in your bucket; all of it
about customer success?” We’ll do two things: First, we’ll do was up for renewal, so your churn
the basics and Second, we’re rate was 10%”. The issue is in the
This one’s tricky because let’s just going to talk about Net Dollar world of multiyear deals, maybe
say you have a model where all Expansion rate, which I think is only half of it is up for renewal.
your customer success team does increasingly replacing churn as a Maybe $5 million is up for renewal
is renewals. They never do upsell. core SaaS metric. But first, let’s this year and $5 million in the
They simply call and check “How talk about the basics of churn. next year because you’re doing 2-
are you doing?” “Are you happy?” year deals. And in that world, you
“Are you planning to renew?”. Churn is how much leaks out of then really want to do what I call
When the renewal comes, they your bucket. The issue with churn ATR churn, which is measuring
take it. And to make Reductio Ad is it’s a fraction, so you have a churn against the Available To
Absurdum, they don’t even ask for numerator and denominator. The Renew base, which to me is the
price increase. They just renew numerator is in theory the leakage more interesting statistic. If I want
the contract as is. It’s hard to and the denominator in theory is to know how long your contract
argue that any of that should go the bucket level. It all sounds durations are, I can ask you about
into the CAC Ratio, because really simple, but it gets that, right? If you do lots and lots
they’re not acquiring new ARR. complicated fast. Why? Because in of long deals, that’s great, I’ll
the numerator, are we counting know that. But when I ask churn,
On the flip side, let’s just say your gross leakage or net leakage? typically, I think the honest
customer success team does all person answers on the ATR basis.
customer base sales. Therefore, So 20 units leaked out of my
sales only brings new customers bucket, but I upsold 10. So do I If we had this many people up for
and your customer success team have 20 units of churn or 10 units renewal, how many did we renew?
does all upsell, i.e. all new sales to of churn? People would call the In the prior example, where we
the existing customer base. 20 a Gross Churn and the 10 a had a $1 million in churn with the

4
$10 million of ARR, we calculated companies that do that. The first Or how about even the year 67
on an ARR basis that it would be thing sophisticated people do payment? The math doesn’t know
10% churn and if it was all annual while reading S-1 is look at, “OK, that. The math is just to assign
contracts, it would be 10% churn. are they survivor-biasing the Net that very large lifetime value and
But say we’re doing 50/50 and the Dollar Expansion rate or not”? An a lot of it is in the very distant
ATR base was actually only $5 honest company would say “All future. That’s one problem.
million…Well, that is 20% churn. the customers who were
customers a year ago if there were The other problem is with
So churn itself is inherently 100 units then, what’s the total negative churn, because the
confusing, because you just go number of units now?”. That’s the calculation becomes
“Oh, Gosh, it’s just a fraction correct way to do the expansion mathematically meaningless, as
leakage divided by base and then, rate. The survivor-biased way is to you have an infinite customer
well, how do you calculate run backwards and say: “Of the value. So what some people
leakage?” And then, “Oh well, how people who are customers now, (bankers) used to do is just cap it.
do you calculate the base?” And let’s go back in time”. They’d simply bring the actual
all of a sudden you went up to number down to a 5 or 10% churn
four different churn rates and it It’s a trick that mutual funds used rate, somewhere between those
makes things complicated. That is to do a long time ago. They’d start two. So we’ll have an artificially
why you’re seeing public SaaS 10 funds and they’d kill off all the cap somewhere you know 10 or
companies talk about Net Dollar ones that didn’t perform and then 20-year customer lifetime and
Expansion rate. What they’re talk about the great then we’ll go do the LTV. And then
doing is they’re just grabbing a performances. It’s the same idea. say: “Well if you have 100 units in
cohort and saying of all the Excluding the customers who 10 years, then you’re a $1 million
customers who were customers stopped doing business with us, of lifetime value”.
one year ago, if they were worth our expansion rate is great which
100 units of ARR a year ago, what is Whoa! It’s a pretty big thing to I just think the whole notion of
are they worth now? It might be exclude. having to quietly introduce a cap
120; exceptional ones are 140 or revealed an underlying weakness
150, an OK-ones are 105 to 110, AM: Churn is also used in with the metric. Again, this is why
that’s really kind of unexciting. calculating Customer Lifetime I think Net Dollar Expansion rate
Value. Could you please discuss is not only replacing churn, but
I think you’re going to start to see that metric? it’s also replacing LTV. Now, the
more and more companies other thing that people do, and I
focusing on Net Dollar Expansion DK: This is one of the problems. actually like this metric, I wish
rate because it’s just one number. So LTV or Life Time Value was and LTV worked better, but the notion
As you remember, with churn, we has been a very popular SaaS of LTV / CAC. I really like it
created four numbers in a very metric. The idea is simple enough personally because ultimately,
short period of time and they’re on the face of it, which is - gosh, if what you pay for something
all interesting numbers. But I I have a 10% churn rate and you’re should be a function of what it’s
think it’s a little bit harder to 100-unit customer then your worth.
game Net Dollar Expansion rate. lifetime value is going to be a $1
There’s only one way to game Net million because you should last Like I told you I had a CAC of 2
Dollar Expansion rate, which is to for a 10-year lifetime. with a lifetime of 20 years, “Well,
calculate it in the survivor biased gosh, I’d pay $2 to get $1 a year for
way which is calculated There are a couple of problems 20 years”. If I was really sure I was
backwards to say: “Oh, let’s look at with that. The smaller the number going to get it, I’d do that all day
all the people who are customers gets the bigger the lifetime value long, right? So is 2 really that high
right now. Let’s look at those gets and the more of that lifetime of a CAC? And then when you look
people and what their ARR was a value is in the very distant future. at it from a payback period, wait a
year ago”. So with a 1% churn your lifetime is minute, you’re getting paid back
100 years and am I really going to in gross margin not in revenue, so
And there are a handful of get that year 97 payment? the 2 is really three years to get

5
paid back. So if you’re trying to it but if nobody is surveying realistically going to move to? You
build a SaaS company, it’s going people and publishing data on it, I don’t have that many alternatives.
to take a lot of capital because don’t know if it’s good or bad, Happy customers don’t renew
you’re going to keep buying right? Almost every metric we’ve because they just got acquired or
customers and keep taking three talked about today you know I they just got a new CFO and he
years to get your money back can tell you like LTV / CAC, 3.0 is /she wants to bring in a different
from having bought one. the minimum and 5.0 is good; planning package. So, what you’ll
CAC Ratio - 1.0 is very good, below find is this what I call as loose
This is why VCs think 2 or 3 is a 1.0 is great, above 2.0 raises correlation between NPS and
large number. But back to the eyebrows and the subscription renewal.
main point, the LTV / CAC ratio is gross margin 80% is best practice,
supposed to be the lifetime value 75% is getting low, 50% way too So when I do these surveys, I
divided by the cost. Basically, if low, 90% super high, right? actually measure both NPS and
you were a $1 million in lifetime we also explicitly ask for an intent
value, you’re a 100-unit customer, For all these metrics, we know to renew. It’s a different question.
and I have a 10% churn rate. I had benchmarks. The issue with a new So there’s a – well, would you
a CAC of 2 (let’s stick with that metric is you don’t have recommend us on a scale of 1 to
example). Your LTV / CAC would benchmarks, you don’t have data. 10, the NPS question. Then right
be a $1 million divided by 200 So one of the best things about after that we’d say, do you intend
which would be 5. So we would NPS is we have lots of data. But to renew your software, yes or no?
have a lifetime value of 5 times what’s a good one and what’s a Because the two are not tightly
what we paid for you. bad one? coupled. And if you ask both
questions, I think it’s great
I like that metric because it’s kind The other good thing about NPS is because then, you want to follow
of got both sides to scale; what’s an outward-in metric and I like up on the low NPS customers. But
it worth versus what I paid. I think those. I think many times SaaS you want to follow up even faster
it’s very solid and a very good companies spend so much time on the people who say they don’t
metric. But as we’ve talked about slicing and dicing customer data. have the intent to renew.
lifetime value, unfortunately, is a We are only seeing that segment
bit of a squishy metric. with the market we’re selling to The other thing about NPS I think
and they might be more to it. So people miss, it’s a little bit tricky
AM: I’ve seen some companies we’re completely blind to people because it’s kind of a composite
tracking Net Promoter Score outside our sample if you will. metric in the sense that you mail
(NPS). How critical is that metric customer survey to 30 end users
in your book? Another good thing is it keeps at your customer. Take as an
your eyes on your customer, example financial planning
DK: The NPS is a very popular because that survey typically has software which I once sold - you
SaaS metric, and I think more than just one question. could sample the end users, the
companies should track it. I think You’re going to ask other people who own the budgets, and
the best practice in tracking NPS questions and get a lot of data out the regular managers inside the
is to survey one quarter of your of that. The bad thing about NPS company. The fact of the matter is
base each quarter. So you’re is it’s not as correlated to renewal if you calculated an NPS score,
serving each customer once a year as one hopes. Well, one, the first you may do it end user score
but you’re getting data four times story you’d go, “gosh! high NPS versus buyer score - this is a way
a year because you’re cutting the customers always renewing, low to cut that data. If we are careful
base in four. NPS customers don’t.” And in about the weighting of the data,
some ways, you’d wish life was by the way because you don’t
There are few things that are good that simple. It’s not. For example, want to weigh the end users 15
about NPS. First, other people unhappy customers renew times more heavily than the
measure it too, so you can get because of switching costs; if finance people, because you have
benchmarks. We can create a you’re unhappy with Salesforce only two of finance folks and 30 of
great new metric and calculate for many years, what do you end users. The most important

6
question to me is who’s writing DK: Billings is another metric you know if it’s a calculation problem
the check, who’s going to actually hear about from SaaS companies; or if there’s actually something
do the renewal because if all the typically, a public SaaS companies. unhealthy with the business.
users are only moderately happy, Because in a private company, I
well, end users don’t always like can just ask the question I care No private company investor
being forced to use systems. about, which is how much new asked about Billings because they
There’s a lot of things that may ARR did you sign up last quarter. don’t have to - they can just ask
not affect your renewal at all. That’s actually a question people about what was the new ARR. So
care about. Billings is a way to try it’s an interesting metric, but not
So it is important to ask right and guess that using public one I encourage. Once in a while I
questions. How many people did company metrics. Billings is do meet a private company that
you survey inside the company? calculated - and some financial uses Billing and I kind of scratch
How many different roles do they analysts call this calculated my head and say, “Why are you
have and what is the influence of Billings to be explicit that they’re doing that?” - no one really cares
those roles, direct or indirect, on calculating it - Revenue in a about Billings.
the renewal decision? quarter plus change and deferred
revenue in a quarter, and I did a AM: Do investors and operators
AM: What’s a good NPS score - a blog post all about Billings use the same metrics?
rule of thumb?
Ceteris Paribus [all things being DK: It’s a great question. If there’s
DK: The reason this was equal] for any company - Billings one big difference, it’s that
particularly hard is I actually think works. But for example, if a investors like compound metrics,
it’s very closely guarded. Like end customer changes billing while operators like individual or
users in general are less happy durations, say it is to do a one-year granular metrics and I’ll give you
than the people who buy the deals and there’s a downturn and some examples of those. One
software for them, right? I would all the customers want to do six example is just CAC Payback
say a good buyer NPS [which is months payment instead of one Period. We’ll start with that one.
the one that matters] is anything year payments, Billings will look
North of 30. I think North of 50 is horrific. In that scenario, ARR will Well, CAC Payback Period rolls
great. Remember, NPS is a metric be fine because all those people together two different things. It
that can go negative. It’s not 0 to have said, “Hey, I want to keep rolls together your CAC Ratio, how
100, it’s -100 to +100. If every renewing, I just want to pay you six much it costs to acquire a
single one of your customers months intervals”. But deferred customer with your subscription
hates you, you’d to have a revenue will drop radically, Gross Margin, which is really a
negative 100 NPS score. because they’re only paying 6 measure of how efficiently you
months, not 12. run your SaaS service. Now, why
There is a guy named Bruce would you want to combine those
Temkin who publishes some This actually happened to me two metrics? To get a sense for
survey data on this and I think when I was at Salesforce where we how long it takes to get the
he’s consistent with what I said. had a quarter during the money back that you spent on
He’d give you Oracle’s aggregate downturn and there were a bunch acquiring customers.
NPS. He won’t tell you by product of six months deals; and therefore,
line or anything else but is on his deferred revenue wasn’t what So, it’s a reasonable metric. The
blog he’ll talk about that. I’ve people thought. Financial analysts problem with compound metrics
heard of 80, by the way, and is the thought thatthere were no new in my mind is if you’re an
highest score I’ve ever heard. sales and it wasn’t the case. It was operator, you don’t know what to
Basically, it means that 8 out of 10 a change in payment terms. So this fix. So if you tell me you have a
of your customers are raving fans. is why Billings is not a great metric bad CAC Payback Period, I don’t
because it works when all other know if it’s because you run your
AM: I have also seen companies things being equal. As soon as SaaS service inefficiently and we
tracking Billings. What’s your something is not equal, Billings need to worry about subscription
take on that? can get thrown off and you don't Gross Margins or if it’s because

7
you pay too much for a dollar of has 150 units, you’re asking for 100 I guess the other one would be
ARR. units of subscription and 50 units cash burn. VCs are in the business
of services. of selling money, so they will tend
Well, LTV / CAC is another to like high cash consumptive
compound metric. If your LTV / A company would actually be companies because it lets them
CAC is low, I don’t know if that’s better off if you could play a little invest. If you’re breaking even on
because your CAC is high and sleight of hand and turn that into day one and you don’t need to
you’re spending too much to 120 units of subscription and 30 raise $100 million round, I can’t
acquire a customer or your units of services that would wipe put $100 million capital to work.
lifetime value is too low, i.e. you’re out your 40% gross margin on the When VCs invest $10 million in a
churning too much out of your services, but now we have 120 company, they’re not thinking
ARR bucket. So, they’re useful units of ARR instead of 100, and that’s the only $10 million they’ll
metrics for investors because they are renewing every year, and put in. They’ve earmarked another
they’re cool and they roll together suddenly your company is 20% $40 or $50 million to say, “Hey, I’m
multiple things. Investors like more valuable. going to be able put $40 or $50
them because they’re kind of million of capital to work on this
compact and dense. To a certain Most SaaS companies, I’m not opportunity. I’m going to do it
extent if you have a bad one, they sure if they’re even consciously over multiple rounds”. Maybe not
just don’t want to invest. But your aware of that’s what they’re that much, maybe they’re putting
decision has to be “What do I do doing, but they’re trying to take $10 and maybe they’ve earmarked
about it”, because you’re an the year one money and divide it $20 just to be part realistic.
operator and you want to see the in a way that as much as possible
metric kind of torn apart. of it recurs. The Reductio Ad That makes sense, because if
Absurdum by the way is all of it you’ve got $800 million fund,
So I prefer the more individual, or recurs when you say services are you’re not going to be able to
granular, or kind of one- free. So you say “Hey, I’m going to invest it all $3 million at a time.
dimensional metrics to the have 150 units per year and free You’re not going to have 200
compound metrics. Other than set up” in which case your services investments sitting on 200 board;
that difference, I think investors business is going to run at minus it’s just not going to work. You’re
and operators care about the 100% gross margins but you will going to want to have a “Check
same thing. Put it like this, have made the company 50% Size” - is the buzzword - of $20
operator should care about what more valuable by moving million or $50 million so you can
investors care about and investors everything to ARR. And that’s a do 16 investments across the $800
care about ARR. Let me give you big question investors and million and you and your partner
an example - this would be operators can debate all day long. are going to sit on 16 boards. So, I
Services Gross Margin. One of Personally, my opinion on in this, think that’s another one –
these age-old questions of the the rule of services in a SaaS investors and operators look
SaaS company is: Should you have company, we should always start differently at cash flow.
a services business and should it with the mission and the mission
be profitable? is to make happy customers and AM: What’s your take on the Rule
not lose money. Even better yet, of 40?
If you and I started a services to maximize ARR and not lose
business, we could probably money. And if you follow that DK: Rule of 40 is another
generate 35% to 40% gross mission statement, I think you’ll compound metric in attempt to
margins pretty easily. In fact, in end up in the right place that balance growth versus
perpetual license software investors like. As a result, you’re profitability. And the Rule of 40 is
companies back in the day, those roughly breaking even on your traditionally calculated as
were the kind of margins you services business, it’s not costing
drove. Once in a while I meet a you anything, and it's actually Growth % - Profitability % = 40
SaaS company that drives 40% helping to maximize ARR, which
services gross margins, but it’s maximizes the value the So let’s say you’re growing 100%
like why are you doing that? company. I think that's one metric and your profitability is negative
Because if in year one a customer operators and investors may look 60%, therefore, you have a Rule of
at differently. 40 score of 40.
8
If instead you’re growing at 100% Kellblog (www.kellblog.com). something called the Magic
but losing 200%, your rule of 40 David Skok’s is another one Number, which is kind of the
score would be -100 (100 – 200). I (www.forentrepreneurs.com). He’s inverse of the CAC Ratio. He does
think it’s a good rule in general. I a great metrics blogger. I think some content on this. The other
don’t think it’s a religion and by some of the PE firms, like guy I would look at is Tomasz
the way it’s mathematically OpenView, do a nice job of Tunguz at Redpoint Ventures. He’s
interesting when people do blogging about SaaS metrics. more a data-oriented than
correlations of valuations to Rule There’s an annual study done by metrics-oriented, but still a very
of 40 score. The most interesting KeyBank - if there’s only one thing good blogger for all things
chart you can show me is Rule of you could look at, that’s what I quantitative about SaaS.
40 score on the X axis by would look at because it provides
valuation multiple on the Y axis. all the benchmark data along with AM: Thank you, Dave. It was a
People have done that and in the the definitions of how they pleasure speaking with you.
general, Rule of 40 has maybe calculate it. That study itself is
10% more predictive power than probably the single thing I would
just simply growth because that’s look at every year.
the idea. If I could just do growth
versus multiple, for the Rule of 40 If you’re an operator and you want
to be really meaningful, it should more detailed SaaS metrics, like
have more predictive power than how much do people spend on
simply growth. The good news is marketing demand generation,
it does, but it’s not dramatically versus marketing people, versus
more unless you call 10% marketing programs - none of that
dramatic. The last time I looked at stuff we’ve talked about. We’ve
it is it had incremental predictive talked about not even marketing
power. by itself but marketing aggreged
with sales, we’ve talked about it in
The other thing you have to high level. So, there’s a company
remember about the Rule of 40, called OPEXEngine that sells
which is a problem for private benchmark data on a very granular
SaaS companies is all the level. If you’re for example a
benchmarks are public. So you financial planning professional and
have a huge case of mis-sampling you’re trying to get data for SaaS
or sample bias. You are sampling companies between $10 and $30
a bunch of companies in the million, how much are they
standard that got to be $100 spending on marketing head count
million, got to go public, have versus marketing lead gen, you can
raised on average $300 million, get that from OPEXEngine.
about 15 years old on average …
and then you’re comparing your Well, obviously, Bessemer - we
$8 million SaaS company to that. can’t not speak of the Bessemer. I
That I think is misleading. I think mean, Bessemer as far as I know
when you get bigger you, the $50 are the first people who really
- $100 million in ARR, then I think popularized the CAC Ratio and
you should start thinking about they actually – I think did it wrong
the Rule of 40. the first time to be fair. But the
good news about them is they
AM: Last question. What have changed over time how they
resources would you recommend calculate things to keep up.
to people who would like to learn
more about SaaS metrics? There’s another guy named Rory
O’Driscoll at Scale Venture
DK: Let me see, one is obviously Partners, he came up with

You might also like