October 17, 2018
Serverless Meetup with Cloudflare
On September 11th, 2018, Heavybit member Serverless Inc. hosted the Serverless user meetup at our San Francisco Clubhouse. The event was co-...
Patrick: Thanks for having
me, I'm excited to chat. My
personal background is in econometrics and math. So
I started my career, I worked in US intelligence and then I worked
at Google, basically doing a bunch of fun
value modeling, either hunting bad guys and
girls when I was at NSA or hunting
money when I was at Google. And
then I jumped out and joined the startup game, the SaaS
game, if you will, basically seven, eight years ago.
What we're going to be talking about today is the amalgamation of a lot of learnings that we've picked up, not only from the data that we have access to, which I'll tell you how we got access to that data in a bit, but also just in attacking this problem, not only from building a product in this space, but also just understanding and seeing inside a lot of different types of companies.
And to give you a little bit of a thesis
here, we're going to give you a framework as
well as a bunch of benchmarks in terms
of understanding and reevaluating retention. We're
going to talk about why retention is so important as well as breaking
things down.
There's going to be a lot of strategic bits as well as a lot of tactical bits. And I like to mix it up because retention, very similar to something like monetization as well as your overall growth strategy, it's this continual thing that is just always going to be within your business, very similar product development and the development of the rest of your business.
Put a little bit more succinctly, I
do call this out a little bit here and there,
but there's going to be some things that are going
to be some really easy ways to increase that lifetime value and then
there's going to be some things
that'll require some more effort, but will have higher impact.
But my goal is to make sure that you have those strategies and
those tactics to implement for that overall
retention strategy.
Now who the heck am I? I already gave a little bit of my personal background. But really I want to focus in on the data. So yes, there's a lot here that I've learned and I've studied, but really we're going to be focusing on this data-driven or at least data-informed strategies that are really going deep on what makes retention work well and what makes it not work so well. Where does this data come from?
We build a bunch of different tools for
revenue and retention optimization. And
what we mean by that or just to give you a little bit of color to that,
we have a core product called
ProfitWell metrics. So it plugs
into your billing system, Stripe, Zuora, Braintree, Recurly,
whatever you're using for subscription billing management, and
then essentially we give you free access to a
whole host of really deep metrics
and benchmarks from your MRR, your cohorts, your
churn cohorts, all the way to segmentation done by plan.
We enrich all your data with Clearbit and FullContact for free, and there's a lot of reasons we give this away for free and more than happy to answer any questions on freemium in the Q&A. But right now, not only because it's free, but also because we really focused on accuracy, which was a big thing that was missing in the metrics market. We have about 25% of the entire subscription space using this product.
All the way from Johnny and Jane startups
all the way up to about Fortune 50 companies are using
it, and it's been where we can actually study and understand
a lot of the data that you're about to see. And
then the way we make money is we have a couple of different products
that focus on retention or
optimizing pricing. And I'm giving
you this context not to sell you, it's probably like the worst sales
pitch ever, but more just to give
you that context on where this data is coming from or where
I'm coming from.
So that being said, let's jump in here. So the first thing to set the stage, retention is everything. When you're thinking about a subscription or a SaaS business, anything in the recurring revenue space. And the reason is that we really have to go to the heart of what you're trying to do with a subscription or a recurring revenue business.
The reason I want to start from these first
principles is because I don't think
enough of us think about when we're getting in growth
mode and we get that growth brain, why retention is so important.
And the reason is that the subscription or recurring revenue
model, it's the first commerce model in the history of markets
where the relationship with the customer is
baked directly into how you make money.
So rather than having to have coupons to convince you or your customer to come back to the store every single week and not go to the store over on the other side of town, or being the corner store where you had a monopoly in a particular market, it's one of these things where I have that relationship with that customer and there's some implications to that, but ultimately I don't have to win that person back every single month, I just have to keep providing them value and inherently, if I'm providing them value, they're going to start spending more with us as well as they're going to stick around longer, which is really, really powerful.
Now the issue there is that enough of us
aren't really spending enough time
on that relationship and we're not even thinking about growth
in a recurring revenue business, in my opinion, properly. And
of course not everyone, but just in the average or in the median
here. And what I mean by that is this is why a lot of people have
started talking about flywheels, loops, these types of things versus the
traditional funnel, because your
goal as a recurring revenue business is to acquire that customer,
you got to monetize them and you got to retain them, and then this
loop just goes over and over again.
Now, in that context of not really spending enough time on the relationship, we did a bunch of research and we looked at about 1,300 recurring revenue companies and we actually categorized every single expense within their business. All the way from custodial supplies and office space to sales teams head count, software prices, all of these different things and basically their P&L (profit & loss.)
And what we found out is that when you
actually categorize the amount of time
and budget that goes to these three different growth levers,
it tells you a very stark story. So
of these 1,300 folks, 57% of budget,
and this was a median, was basically spent on acquiring customers,
so sales, marketing, and these different pieces. Basically
nothing was spent on monetization, and these were companies that
weren't super, super early, these were companies that were actually on a
nice scaling path.
Then there's a little bit more time spent on retention, but not much. And this was mainly the product teams, but even a lot of the product teams, you could see line items that were basically going towards more acquisition focused things. And it's not to say that you're not going to spend half, if not more of your budget on acquiring customers. That tends to be the case, even in more dev-orientated products.
Over time, it's just inevitability. But
the thing is that you have to think about not
necessarily the fact that you're not going to spend that money,
but you have to realize that in the market that
we're in, acquisition has now
become table stakes. Put another
way, you need to be really, really good at spending that 57% of
your budget on acquiring customers just to survive, just to
get a seat at the table.
Now, there are exceptions, of course, if
you're trying to build just like a bootstrap
company that you're not trying to go for the moon, if you
will. But it is one of those things that if you're trying to be a fast
growing hypergrowth company, it does require
not only being good at acquisition, but being
good at the other growth levers.
Just to give you a little context here, we built a model where we isolated these three main areas; acquisition, monetization and retention. We wanted to find out that all things being equal, if we were able to improve or increase each of those levers by the same relative amount, what would be the impact on revenue? What we found is that if you increase your acquisition volume by about 1%, so your net new leads, your conversion volume, et cetera, you're going to see about a 2-3% boost in your overall revenue.
Now, if you focus in and basically improve
your revenue per customer, i.e.
your monetization or your actual overall retention rate,
how long those folks stick around, by the same 1%
level, you'll actually start to see about a 4 to 8x improvement
or a 4 to 8x impact on your overall revenue. Again,
this isn't to say you're not going to spend most of your money over
here, it's just to say that if we
really evaluate what's going on in
our own businesses and how much time we're spending on not only making
sure that we're actually monetizing those
users, but also keeping those users
around, those customers around for a long time, it
is pretty telling on where we're putting our priorities.
As the market continues to get denser
and denser, it's going to be more important for us, not only to be
good at acquisition, but also good at this
other aspects of balanced growth.
So that background aside, let's go
deep here into retention. What I want
to start off with is just making sure we're all on the same page
with how to look at retention. The
reason for this is I think that our industry, we have a
pretty big misconception of what retention involves.
The reason for this, especially when we're looking at products that are, let's say dev-heavy, if you will, which is obviously a lot of folks in this group, we have a very almost Pollyannish look of, hey, if there's value and you're going after the right customer, obviously those customers will sticker around. It's not to say that this is wrong, it's not wrong at all. It's not nuance enough, it doesn't go deep enough. It's like saying, well, if we build it, they'll come, right? And if it's valuable enough, people will just use it. And we all know that the devil's actually in the details here.
And so when you think about retention, to
give you a model here, we have the
spectrum and on one end, we have these advocates. These
are people who are bringing you more users because they love your
product so much. These are the
people who you can't even imagine will ever cancel and they're also
probably buying more of what you're
selling. Then on the other hand, you
have these people who have already canceled, they've already churned,
they're probably bad mouthing you because they had such a bad
experience. Then in the middle here,
you have people who are obviously on the precipice of canceling
a product.
Now it's interesting, most of us, when we think about retention and we talk about retention, we're talking about these ends, these things that are heavily influenced by product. This is the strategic retention bits. Getting those people who are the right types of customers, getting them to the point of value and moving them along the ends of these particular spectrums to become advocates or at least make sure you're not getting those really, really improper customers.
But what we found at least in looking at the
data, and we're going to go deeper into this, is that
in the middle here, there's basically this tactical retention
zone. And it's not the largest bucket of retention,
but there's this tactical zone where you can nudge
people from almost canceling to sticking
around so that they can hopefully see the value more, or taking
that person who had a payment failure and before they actually churn
out, saving them to stay beyond that point of
cancellation.
When we actually look at the data and we look at these two different types of retention, typically 60 to 75% of retention is going to come down to that product. It's going to be that death by a thousand paper cuts that you as a product leader are going to be marching against those features, those right customers. Then the other part, like 25 to about 40% is all this tactical bits. Now, the reason I bring this out and draw this distinction is because most of us, because we're so focused on the strategic bits, we don't realize that there's this pretty large chunk of retention that can be solved with much more tactical pieces that are somewhat easy to implement and can get more complicated depending on your size, but it is one of those things where you do need to balance both of these.
So we're going to be talking about, in the
context of these two types of retention, a bunch of
different pieces that you can implement or think about to actually boost
retention overall. So how do we boost
it? Well first, how do we measure
it, right? So the primary
measurements are typically net revenue retention. This
is a combination of those folks who are canceling offset
by those folks who are upgrading or spending more money with you.
The other way we measure is typically just raw
cancellation, so that churn that
typically comes across. I've
already alluded to it, but you got three big areas of retention.
You have your active cancellations. These
are people who are actively hitting a button saying, I don't want your
product anymore. You have expansion
revenue, which is probably one of the most underutilized pieces,
especially for product led growth companies.
And then ultimately you have delinquencies, which is not the sexiest topic in the world, but for products that cost typically less than $500 a month, at least at the entry point, it's one of the largest single buckets of lost customers, and we'll talk about those in a little bit.
But starting off here, talk a little bit
about active cancellations. There's
a couple of things to think about here and most
of these are going to be very product orientated. Time
to value, the target segment, this is, again, that death by a thousand
paper cuts. We'll talk about this and give you
a little bit of a framework to think through this in
a second. But in addition to that, there's some tactical things
that you should implement as soon as humanly possible. Things
like term optimization, as well as wind back and triaging.
We're going to talk about these in some quick hits. Everything's going to be shared with everyone and I'm happy to go deeper in the Q&A. But in terms of term optimization, the thing you got to think about is these longer term contracts, lower cancellations. Obviously that's correlated, but the basic idea is that when someone is on a 12 month contract, a six month contract or a quarterly contract, or even a two year contract, depending on your type of business, they have less purchasing decisions, even implied purchasing decisions than if they're on a monthly contract and they're getting a receipt every month and thinking about your product.
To give you some data here, what you're
looking at here is on the X axis is
basically of these 3000 businesses,
those who had a percentage or what percentage of their contracts
were annuals compared to monthlys. And
then here, you're looking at gross revenue churn. And
so, again, highly correlative, obviously these companies that are doing
100% annuals, they'll typically have
also a higher end sales process.
But to give you a more direct data
point, annuals typically have 30% better retention
overall than monthlys. Quarterlies
typically have about 20%, and so it's
one of those things depending on your type of business, you got to be
sure to actually ask for these
upgrades beyond just the signup.
The signup is typically one of those things where the person hasn't even seen the value yet, but it's the only time we're asking for that annual or quarterly upgrade. And it's okay to ask there, but you got to make sure they then do it later, basically, in their life cycle. And typically, you want to do this programmatically and study your engagement data a little bit as well as your planned data. But to give you like an easy route, typically you want to go after folks from two to nine months in their life cycle, if they're on monthlys, to try to get them on longer term plans.
If they're already on a 10-month or a
12-month trajectory, they're typically going to
stick around at least in looking at
a bunch of different data. And
going out and reaching out to them, in-app notifications and email,
it can be super straightforward. And
then once they basically hit Let's go, making sure that
they just have a really quick experience. The
experience here really matters. If
you're going to hack it a little bit, just doing email replies works
well.
One little note, and I can send the data on this if anyone's curious. Anyone who has looked at pricing data, there's always this battle between percentage discounts or whole number discounts. So two months free versus the equivalent percentage off. Whole numbers when it comes to upgrading, particularly with this type of promotion, works significantly better than using percentages. So put another way, use $100 off, use 2 months off, use that type of language over using percentages.
Now, second big thing here is
tactically having these triage of these cancellations
because churn's a fact of life, and there are people who, obviously,
are going to leave your product because they're just not seeing value
and they're just not the right type
of customer. But there's plenty of
people who are on that precipice who maybe didn't have enough time
to actually use the product and see the value, or maybe they're
just not ready for it either.
So when you have someone who wants to leave,
the biggest thing is to add some
friction. Now to be super clear
here, do not hijack your customer. Again,
it's a relationship, don't hijack them, don't make them call you, don't
make them do a bunch of these types of
things. There are probably some
circumstances where that makes sense, but mostly they don't
make sense.
The thing is, is adding like a one-question,
maybe a two-question, survey is super, super
reasonable, but it also gives you
an opportunity to not only learn so that you can
catalog and hopefully predict some future cancellations or some
future churn, but it also gives you the ability
to do some triage on why that
person left.
So if I ask a question like, "Hey, of the
following reasons, what was the biggest reason
you're leaving and what's the smallest reason you're leaving?" And they
select for the most important reason
they're leaving something like, "Oh, I just
didn't have time to use the product." Well, what I want to do is I want
to give them a salvage offer.
I want to say, "Okay, cool. Well,
why don't we give you the next month for free," in some cases,
"Or for half off next month?"
It's pretty
simple, you can do this with some simple modals, "Hey, here's a free
month." Have them continue. You
got to be careful because people can abuse these, but that's pretty
easy to handle. But it keeps that user around
and it also it helps them basically
start to see the value in the product because sometimes
that relationship needs a little bit more time to fall
into that symbiotic relationship going forward.
Now, some other things that are very similar but actually can protect in certain cases are like maintenance plans. So this is basically like, "Hey, we'll store all of your data so that when you come back, it's all here and ready to go and you don't have to restart." This works with a lot of lock-in type products. These products that have something to be locked in. Normally it's less than what you were getting, but ultimately what it does is it allows you to have that signal and have some MRR coming in. And then as soon as they come back and start using the product, you can instantly get them back up to that original price point. Pause plans were very, very commonly used.
I recommend these less and less in
dev-orientated products because typically
there's actual like either background usage, meaning
it's running some sort of infrastructure or some sort of reporting,
or there's actual active usage on it. And
in that case, a maintenance plan makes a lot more sense. But
in some cases doing a pause plan, meaning it's a bit
better than a full key installation. And
as soon as they come back, they're a little different than just a brand
new customer and you can get them
back on to their paid plan.
Then there's also freemium tiers, highly used
typically in dev or engineering
oriented products. The other thing I
always recommend, especially for a group like this, is it's okay to play
chicken. I think a lot of folks, they
want to play chicken more so than
they should, but there's other folks who are very insecure and are just
like, "I don't want to lose that
money." Right? For a lot of
infrastructural products or products that are in workflows,
it's okay to play chicken because you want that customer to realize
what it's like without you, right? And
you can do that a little bit more easily in certain cases, but it's okay
to like stand that line.
The other thing is, win-back campaigns. So after someone's canceled, it's a great marketing channel for you to go back and get folks. Now what's interesting here is that typically percentages work better in this scenario because they're looking for that big, healthy discount. So it defies some of the other notions that we've learned and typically around upgrades in these types of things. I mean, because it's a different profile psychologically of a user.
Now the last piece on active cancellations,
this is a bit more strategic, a little bit more nebulous.
Really understanding your customers.
And the reason I'd like to talk about this is because we've
all heard this advice about, hey, we should do our research, we should
talk to our customers, we should do
our customer development. And then
there's someone at the company, maybe yourself included, who likes
to quote Henry Ford, who probably actually didn't say this quote,
or Steve Jobs of like, "The customer, of
course, they're going to say what
they want." Right?
The difference is that the biggest misconception I should say with customer research is that you don't have to listen to your customer. In addition to that, you shouldn't be asking them exactly what they want, you should be asking everything around that particular problem or that value that you're providing. And that allows you to understand like how much of a hill you have to climb, or it helps you understand like the shape of your products.
And ultimately, you're going to make the
decision filtering this information market data, your vision,
et cetera. And so it's really, really important,
especially in this day and age as well, to understand based on some
data I'm about to show you because we don't
have a good conception of where our
customers actually have value in
products and what we're actually building. And
I'm going to support this, but also give you a little bit of a framework
to think through value of features
and these types of things in one foul
swoop here.
When we think about a product, and
it doesn't matter what type of product it is, yeah,
it doesn't matter if it's function at all could be a cup of coffee, it
could be an enterprise piece of software,
you typically have two axes of value.
You have the relative value of the features or the functionality that you're offering, and then you also have a willingness to pay. And so your job, at least in my opinion, when you're looking at product and you're looking at features and functionality is you first want to understand from that segment or that customer's perspective what is the relative value of the features.
So if I'm talking
about a cup of coffee, I want to understand out
of taste, country of origin, temperature, these types of things, what
is the most valued feature in that
list. What's the least valued feature in that list. And
if I ask a group, or my target customer group, or a sample of a
group, those questions, I might find out things
like taste is the most important
feature in that group. It doesn't
mean everyone cares about tastes as their number one, but in that group,
it comes out as number one. And
then things like country of origin are on the low end.
There are people who care about country of origin as their number one,
there's just not a lot of them,
right?
I then want to basically overlay willingness
to pay information, where I want to
ask them about that cup of coffee. What
point is this way too expensive? What
point is this a good deal? And then
basically you in that composite data, what I want to do is I basically
want to get a baseline. Okay, the
entire group, this is where their willingness to pay is. And
then those people who cared about a particular feature, this is their
willingness to pay relative to others.
So I might find out those who care about taste, their willingness to pay is about 15% more. Those who care about country of origin as their number one, they're willing to pay about 20% more. And where this becomes powerful even if you don't collect the data, even if you just do a thought exercise is this gives us a bit of a framework for measuring relative value of features or functionality.
So if I find an aspect of my product that is
high value where I presume it's high value
relative to the other feature and the people who
care about it as their number one are willing to pay more, that's a
differentiable feature. Something
low value relative to the other features, but the people who care about
it as their number one are willing
to pay more is an add-on. Something
high value, but low willingness to pay is a core feature, and then
commoditized features, which we all tend to
have to build at some point.
What's interesting is we went out to a bunch of product leaders and we made sure they understood this grid and then we asked them, "Out of your last end features, where do you think they fall?" And this is basically what they said for just under 5,000 features. We then went out to about one million or so of actual customers of these products, it was a composite study, and using our Price Intelligently software, we basically found out this is where actual value was. And this isn't to say that you're not confident or you don't know what you're doing, it just means that like the market is getting stretched more and more.
It's super, super crucial to use a model
like this either qualitatively or
more quantitatively to actually
validate what you think is valuable or
at least the level to which it's valuable because we're long beyond
the market where we know exactly
what our next 18 months of roadmap looks like, like
in 2005, 2006, it's a very, very different market today.
And so there's a lot more we could
go into there. But I'm going to
save that for the Q and A and move on here to expansion revenue.
Expansion revenue, I already mentioned this, it's one of the most underutilized pieces of retention mainly because we're so focused on acquisition and we're so focused on just the concept of keeping people around, we forget to ask them to pay us more money, right? Through cross-sells, up-sells add-ons and then the value metric, which is one of my favorite topics to talk about. And the basic idea here is that, again, you have that existing customer, it's not only easy to keep them, easier to keep them then to get a new customer, it's also easier to get them to pay you more.
And we should be taking advantage of
these advocates. To give you a little bit of
benchmark here, the best recurring
revenue businesses out there, they have 20% or more of
their new revenue each month coming from their existing customer base.
But most of us we have about less than 10% or
it's very, very anemic because we
don't really focus on this part. A
couple of pieces to go deep into here, cross-sells
and up-sells. These are typically
other products that you can sell someone or getting them to upgrade
to a more premium version of your product. There's
a lot written on upgrades and we can get into
that in the Q&A.
The one thing I want to point out a lot of future growth is coming from multi-product companies. In the first wave, there were a lot of companies that waited till 100 million+ to be multi-product. And we're seeing the second wave, a lot of people waiting. They're not waiting that long and getting into multi-product very, very quickly because the basic idea is you get really good at building and building becomes not the easiest thing in the world, but easier than distribution.
And
therefore you already have this base to distribute to. To
give you some benchmarks here, multi-product companies typically have
30 to 50% higher growth rates than those who
are just single product. You're
looking at some data here for companies between 10
and 100 million, so a bit further along, and these are different
pairs essentially of companies depending on their price
point. The left side of the pair are single product, right side are
multi-product.
Now to be super upfront, if you're a bit smaller, if you're in the one to 10 million range, you should probably stay single product. Typically, we see growth rates of single product from 1 to 10 much, much higher than multi-product. And then when people get over 10, as soon as they have like their infrastructure with either sales or like product-led growth down pat, that's when things start to accelerate. So it's like, that's probably the time to wait. Now, one thing you don't have to wait for, and probably one of the easiest things to implement today, regardless of your size, is a better add-on strategy.
This in particular is one of the most
underutilized aspects of pricing,
retention, et cetera. I'm going to
give you some numbers here. What we
typically see is that those customers who have at least one add-on
typically have about 20 to 50% higher lifetime value
than those who don't. And this
comes from two places. One, the
most obvious, is they're obviously paying you more for an add-on,
but in addition to that, their retention or their churn
rates tend to be much, much lower, which is great,
and normally this comes from they're more entrenched
or entrenched with your product or with the ecosystem that you're
building.
Now, what should you use? You can use this framework that we talked about, things like priority support, things like analytics and certain places, these things that not everyone really wants or finds valuable, but the people who do find it valuable are willing to pay more. Another way to look at this is if you have a feature or a piece of your product that of the people in that tier of that people in the group, less than 40% are using that feature, that's typically a good indication that it's a good candidate basically for an add-on.
So that can start the journey and thinking
through that. The last big piece
here, if you're going to spend any
time strategically, particularly when it comes to
products like a lot of yours, it's on your value
metric. And your value metric is how you charge. So
left side here, this is Intercom. They
charge based on the number of end people that go through
the live chat widget, whereas Drift a competitor of theirs,
basically charges on a per-user basis.
I wanted
to show these two different competitors because Drift
primarily targets sales folks, as well as a little
bit of marketing, whereas Intercom goes after product
support, sales and marketing. And
so it's a little bit more of a need for a genericized value metric.
Now, the reason this is so powerful is because if you remember back to
your economics class in high school
or college, what you'll remember is that your professor or
teacher basically said, "Oh, we have this demand curve.
Here's the price point on the demand curve," and they shaded in the part underneath that point and said, "Oh, this is your revenue," right? Well, with a value metric, you presumably have infinite points on that demand curve, and therefore your revenue will increase. You're not going to have actually infinite, but this essentially makes sure you're not charging Disney who comes in for your product the same amount of money as you're charging Johnny or Jane Startup.
To give you some actual data, cancellations
are typically much, much lower. You
will see more downgrades, but people will downgrade to,
"Oh, I only need two seats," but then they'll upgrade. "Oh,
I need five seats." Right? You'll
have that more so than actual straight-up cancellations. So
you're looking at here about 6,000 companies. These
are pure value-metric companies on the right, and these are feature-differentiated,
old-school product pricing.
The churn is basically half for those who are value-metric and then expansion revenue is basically baked into growth because it's not like you have to convince someone to upgrade or buy another product. It's just, "Oh, you're using more or you're getting more value from this," because it doesn't always have to be usage, "therefore we're just going to upgrade you to this plan." And so you typically see that expansion is about double for value-metric-based pricing.
To bring this home, if there's one thing from
a monetization and retention
perspective, the value metric is the one thing that,
if you get everything else wrong, but your value metric is on
the right path, you'll tend to be okay. And
so that's why I recommend, don't worry too much in the early days
when it comes to monetization about a
particular price point. Don't worry
too much about, add-ons in the early days.
First get your value metric right, and then start to worry about those other pieces, because the value metric will save you from yourself, or at least from the lack of knowledge you have as you're pushing things forward. Now the final piece, and I'm not going to be able to go too deep into this just for time, but I have learned more about credit cards than I have ever wanted to in my life, and so I can definitely talk for hours about payments, but I will give you a good primer here.
Delinquencies are funny. There's 130+
different reasons why a credit card
fails. There is no single one of those reasons
that breaks probably 12-15% of
delinquencies, meaning there's just a lot of reasons
and they're all over the place, and depending on your company, they'll
be all over the place. But there's
some things from an infrastructure perspective that you can
do, and then of course, tactical things you can do before
a point of failure and after a point of failure.
Just to give you a little bit of context here,
it's the largest single bucket of
lost customers. We typically see,
for products that are credit-card based
and not all products are credit-card based, but both in B2B and
in consumer products, it typically accounts for about 20-40% of
lost customers. So you have a hundred
customers that you lost, 20 to 40 of them are going to be because
of payment failures.
It defies logic sometimes, because you're just like, "Well, this doesn't make sense." Right? Well, the problem is that the underlying technology for credit card processing, while computerization has obviously been updated, there's some things around risk that have come a long way.
A lot of the underlying processing technology
hasn't really been updated since
the 1960s, 1970s. And it's a little
bit of a generalization, but the
thing is they're mechanical devices, subject to failure.
And so what ends up happening is that even
if Stripe or someone is 99.9% good,
that 0.1%, given just the straight-up volume
that happens, ends up accounting for
a certain portion of these users where you actually need to reach
out to them to get updated payment information.
Even the ACH fails, which is always fascinating. It fails at a lower rate than credit cards, but it still fails as a high enough rate where you need to at least set some things up on the backend here. The main reason is we're just really bad at recovering them. Just to give you some numbers, and I can send you deeper reports on this, if a hundred people have their payment failed, on average we're only recovering about 30 of them.
If we look at some of the best folks out
there, they're typically recovering
about 60 to 80. So you're going to
have some people, this is their excuse to leave you, but
then there's plenty of people who just don't even realize it. And
there's that delta that most of us should cross. I
do see some people are really good at 95% recovery, and
I certainly see people like much lower than this at 5% recovery.
It's one of those things to just think about, and the way you think about this is you got to treat these folks as a marketing channel. Most people that we look at, because we sell a product that solves this, they send very bill-collector emails. Not quite, "Where's my money? I'm going to break your legs." But pretty close. And again, remember that relationship, right? And so what's great is that this is purely mechanical cancellations, and so there's a lot of things you can do to basically solve for this.
Before the point of failure, the most
obvious thing is exploration tracking.
Just to give some tidbits here, if you're not going to put in
the effort to study the data, do not send
emails for expiration updating. Emails
typically increase active cancellations by about 10-20%. So
use in-app notifications here. And
I wrote a lot on this so I can send more details, but typically
like actual in-app notifications that will come into the
user's field of vision, rather than just a persistent bar
inside the app, works much, much better as well.
If you
want to get more advanced and actually study the payment data,
this is where you'll start to add things like,
"Oh, we noticed these debit cards,
they always fail every X months. We're
going to get ahead of them as well." But don't worry about that too much
if you're not going to spend the
time. Just expiration tracking is a good starting point. And
then after the point of failure, there's a bunch of things you can do.
The bulk of recovery is going to come from email, in-app and SMS. But what I will say is, the instinct is, "Oh, let's create really beautiful emails." Beautiful emails actually don't perform that well. The plain text, "Hey, John, Patrick here from ProfitWell. It looks like your credit card failed. Could you update it for me?" Those emails work so much better. There's a little bit of a reciprocity that happens where people will either respond, "No, I don't want the product anymore," or they'll fix it because they don't want to let the person that's sending the email down.
Whereas very heavily branded emails are really
easy to ignore. And then in-app is
a really good place here. And SMS,
if you have it, got to be careful. I
always recommend just one SMS message. They
typically have 90-plus open rates. So
if you see an SMS message and you
ignore it, it's a pretty good signal that you shouldn't
send another one. And then the one
technical thing I would do is make it so your user doesn't have to
sign in to update their payment information.
It's a good little friction point. We
see the biggest drop-off if they have to sign in.
There's some good technology you can use, we have this in our product, but you could build it yourself, I'm sure with some work, to figure out how to get that updated payment info without that friction. And then the one last thing I'll say is, please make sure you lock out your customers who aren't paying you. I know that comes off a little condescending, but the reason I say that is there's typically one out of 15-20 companies where whoever was in the billing was setting up flows to maybe treat people who are delinquent different than people who actively cancel, which is the right move.
They close the loop for the people who
actively cancel, locking them out of the app, but
there's a certain portion, it's not a majority by any means, it's one
out of 15-20, which is not a lot,
but of those people, they don't close the loop for
the delinquencies. They keep them in this purgatory. And
so I've been on calls where we've been talking about this, and
I always say this now because there was one
company, and this is a fun one to
talk about, where they had the equivalent of about
$10 million in annual revenue using the product and
they weren't locking them out.
basically they started locking people out and I think they didn't recover everyone, but they recovered, I think, six to seven million, which was the greatest growth hack ever, which is not really a growth hack, but it was a really, really good day for them. And I can go deeper. There's a bunch of other things around like smart retries using Stripe's billing systems versus your own, these types of things that I can get into. But yeah, I think hopefully it's enough. Lot of information, lot of stuff going down here.
Just to recap before we get to Q&A.
The biggest thing that I would mention here, it's a game
of incremental optimization. That's
the biggest thing to think about. When
you're looking at this spectrum here,
you got a lot of things that you can focus on.
I think in the early days, understanding your user, your
segments, that's going to pay off a lot. Getting
your value metric right and then setting up some of the basics of some
of the tactics here are super,
super useful. As you get larger or
as you get obviously more revenue or there's more at risk,
starting to use more of these tactics and just double down
on some of those strategic bets is super, super important.
We talked
about these three big areas. You
got your active cancellations. And
this is how I would start is just by tracking these three areas.
I see a lot of businesses, they don't even know what their active
cancellations look like versus their expansion revenue, et cetera, let
alone what their net retention looks like.
Like all things, once you start tracking, you
tend to improve it or at least realize
what you're giving up, which I think is the first step in anything.
The big thing here is, nothing here is rocket science. Nothing here is like, "Oh, you need an advanced degree for or you need to become an expert." A lot of it's just the effort in breaking down the problem. And so if you ever have any questions, feel free to hit me up. We've written a ton on this. There's a lot of data we publish on this. And of course let's get to some questions.
Yeah, it's a really good question.
So I think the last thing you just said about
how much advocacy takes place.
So there's certain products where the buyer and
the user are the same person. There
are certain products where the buyer and the user are not the same
person, but they're close. So this
is like director of engineering and then engineers.
And then there's products where it's very different, right?
Where I'll sign off on our AWS
stuff, but I am not the person to ask about it. Right?
so in that latter case, you
actually don't have to worry as much. You
almost have to treat your user as the buyer, because in the case
of our AWS spend or DevOps tools or these types
of things, I'm just going to trust
our head of engineering and our head of product. Right?
And you might have to care about me a little bit just
in terms of terms or something like that, but probably not as much,
but the other two, normally when there's
a disconnect, you want to care
about the buyer from the actual purchasing perspective
but if that buyer has cancellation
authority as well, you tend to want
to worry about them a little bit more when it comes to retention.
Put another way, basically what you're trying to find is, is that when it comes to retention, you want to worry about who the primary user is and the buying that comes in the world of pricing, but for retention it's who really has that cancellation authority. In the case of, hey, that user has a lot of influence on cancellation or is the person who could cancel, care about them, but they also might be the buyer in that case. And if it's disconnected, worry about whoever has that cancellation authority.
Yeah, it's a really good question.
Well, there's a couple
of ways I could answer this. I'm
going to answer it a couple of ways. For
one, integrations and I actually, I should've put this data in
there, sorry. I forgot to put it in there.
Integrations actually tend to increase
retention quite a bit. You used to
spend or used to actually charge for integrations.
If you wanted a Salesforce integration, you'd charge an additional,
five to $900 a month for it in
addition to the product.
But what
we're finding is, is that integrations
are a better tool to not only get better retention because people are
more entrenched in your product, even if
they're not actively using your product. I
think the increase is somewhere around someone who has
one, to three integrations, I think they typically
have 10 to 15% better net retention
than those who have zero integrations. And
then those who have more than four or more typically it's another
five to maybe 7% bump.
That's a big, big thing to think about is making sure your product can permeate throughout the workflow of your user. And the reason this is important is because we used to have this phase of building product where it was like, oh, everything needs to be a source of truth. This is a source of truth for this, a source for truth for that. And it's such a high lift, especially today when you have the Salesforce, HubSpots of the world, as well depending on the dev product, you have your equivalents.
And so I think the big thing to think about is
I don't have another word for it,
but I should come up with a better word, because
this is very negative, but more of what's your parasite strategy?
What's your strategy for hey, we've created this value
and then we've gotten it in every other app or every
other product that's out there so that the idea of tearing this
out feels bad. Now
in terms of expansion revenue, I think that more integrations
do increase willingness to pay typically
because people appreciate your product more. But
I think there are ways to expand revenue through
partnerships, which I think is what Dana's getting at a little bit here.
It's just tough and I would go deep on some of
the partnership side of things.
Research some of that, because it's really what
I found with partnerships is it's
really tough to find perfect overlap
where you're both incented for the same thing.
And that's the big thing is if you could find that and someone's going to distribute your product or help you distribute your product or is going to basically give you a cut when people integrate and start paying them, it's a phenomenal thing to do because you can double or triple dip there, but it does take some effort and it probably takes more effort than you always anticipate because you see how perfect it is, but they might not see it. And then even if you both do, it's just like all the effort of figuring that out.
It's getting rarer and rarer that you find an addon that is an integration, if that makes sense. It's getting rare. Now it does exist. Some people will charge but it's one of those things where, because features are getting more and more commoditized, because there's more stuff out there and it's because we've gotten good at shipping stuff quicker, ultimately it feels more nickel and dim-y than anything. You got to be careful about charging for those integrations. It is possible, but it depends on your niche and the integration.
Number one is value metric. I
talked about that. I belabored that a little bit, but the reason is, is
because it's the one thing you can
figure out that'll help with your pricing and help with your
retention. It also helps with your growth because you'll be able to get
people in. And in the world of more
dev-orientated products, you
typically see that this is how pricing already
is.
Think of buying AWS. AWS
has a lot of value metrics. I think
they have like 16 or 17, depending on what you're trying to do for.
And when you have a more dev-orientated product
with a dev-orientated customer,
they can handle a little bit more
complexity. You don't have to think
as deeply about it as if we're selling to
marketers who they can really only handle one metric
because they're not used to buying
in that manner. And so yeah, that's
something to think about, the value metric.
And then the customer. You've gotten this advice if you're early on, understand your customer, do your customer research. It's that one thing where only 20% of people that I talk to about it will actually do it. We'll all tweet the articles about talking to your customer and focusing on your customer but very few of us are actually doing it. But what you find is, is that the more you know about that customer, that segment, the better off you end up are not only in building, but also in your go to market. You have to get good at filtering. But filtering is something you gain over time.
And I'm not talking about oh, someone asked for this in a support ticket, therefore we should build it. That's not what I'm talking about. I'm actually talking about understand your market, understand that customer. Those are the two things that pay the most dividends overall. There's a couple different ways to build. I'm a huge fan, especially in B2B, of building in a way that you get 10 people really, really happy or high NPS or really fulfilled customers and you get 25, then you get 75, then you get a 100, before you even do hardcore marketing at all.
This is
how Drift was built, this is how he
does it, Hiten does it this way. There's
a lot of people who build this way because it prevents the noise
from coming in. Yeah, I would also
think about making sure your retention is really good before
you start scaling, because that means you've discovered why someone uses
your product, why they convert on
your product and ultimately why they're happy with your product.
I am not a fan of using competitive based data as the primary data. I think it's secondary or tertiary data works. And what I mean by that is you typically have three inputs if you're not a market product. A market product would be like gold, where it's clear market's going to set the price. Some products we're building aren't those products. The three inputs are value based data, so actual research going out, measuring willingness to pay, choosing a customer.
And there's a bunch of ways you can do that.
Competitive data and cost data. Your
customers don't care about your cost, they care about their costs.
In addition to that, your costs are probably
very small relative to the value
that you're providing. It's why
we're all in this world of SaaS and software, it's because there's
that huge difference. And from a competitive
standpoint, you're assuming your
competitors have done their homework. Apart
from having some sort of signal that there exist, I
almost guarantee you, they haven't done their homework, just guarantee
that they haven't done their
research around their pricing.
So you want to be careful about copying
not the best kid in the class, if that makes
sense. And so what I recommend
doing is doing your value based data, thinking about your
value metric, thinking about your general price point based on research.
And it doesn't have to be complicated.
You could talk to 20 people
qualitatively and just go out into the market
and then validate more later.
But from there, what I then recommend
is basically looking at the market and are you
getting compared every single
conversation to that competitor, without
you bringing them up? If that's the
case, then maybe you do need to weigh that competitive data
more, but more often than not, you find that's not the case, but
there are some markets where that's the case.
Even if you're trying to build the next Jira, there's a world where like, yeah, Jira doesn't really come up because, yes, that's what they're using, but they don't like Jira and that's why they want to come to you or that's why they're getting on the phone. You just have to be careful of how much you look at it. And it's not a point of not doing it, it's more of just how much do you weigh that data?
For one, I'm not
a fan of keeping people on their
legacy plans. This is unpopular with more dev-oriented
founders and builders. And the reason is I'm
not a fan of it is because it makes
your life that much harder. And
what I mean by that is going from one
to 10, you can go from one to 10 without raising
prices on your customers in any way, even your new customers.
It makes it much more difficult, but you can. Going
from 10 to a 100 it's really hard. And
if you haven't built up that ability to
communicate, understand and that muscle
about changing up your monetization, it gets even harder
after that larger point.
It's something that again, you only can see what's in front of you right now, but I'm just giving that advice for hey, this is coming, do the work now to get there. Maybe half people will listen to that, the other half won't. But in terms of the earliest of customers, you can keep them on their legacy pricing forever if you want as a thank you but just be careful as you can be on that. Now, how to handle those early folks in transitioning.
Normally what I
would recommend is, hey, we don't know
what our pricing is for right now, but we just
think a nice token amount of 50 bucks a month or something like that
works. And then you'll be the first
to know and we actually want your input
on our pricing as we get closer to that. Then
you'll definitely get a deal for like the first year, once we
established that pricing. Because
you've subtly, if not overtly set the expectation
that this is going to change so that it's not a surprise. And
oftentimes we're scared to do that because we
think it's politics where we're like, oh, we
can't talk about numbers. We can't
talk about price.
Your customers know things cost money. And yes, maybe they want a deal and maybe they were going to talk to you to get a deal for a while, but it's okay. Now, as you're established and when you're changing prices, a couple common mistakes and I can send a little guide over too on how to change prices, but a couple mistakes are, I see a lot of price increase emails that make it all about the company, not about the customer. Hey, our prices went up, so your prices are going to go up and it's not necessarily overt as that, but that's what happens.
I also see people don't clearly communicate the value that they're providing. When I send one of these emails, I want to be like, hey, we've been able to build a lot for you, looks like you've gotten 10,000 contacts over the past 12 months.
You're reminding them of how much they've used
the product, how much value. Oh,
you asked for feature A and we have feature A and you use feature
B every week. That's awesome. Remind
them of the value, then put it in the context of them and
say, "Hey, in order for us to continue to invest in making this product
better for you, we need to raise
prices."
And some people go as
clear as, "We need to hire more engineers, we need to do this.
This is what we're going to do with this.
And then because you've been so loyal, we're
going to raise prices on everyone
else new today, but because you're
awesome," and you don't have to say it that overtly, but basically
that's what you're saying, "you're
going to keep your price for the next six months."
This is known as a legacy
discount. And basically, "You can keep
it for the next six months and then we'll put you up to the next
price." And then my favorite thing to do is obviously,
"Hey, if you have any questions, let me know. And
then P.S. If this material impacts your business or your life,
if you're a consumer product, let us know and we'll work something
out." And that PS is typically for two types of people.
It's for people like me, who I run a business and I don't want to pay more for things, even if they're valuable but if I see that and you've reminded me how valuable things are, I'll basically not make a fuss. It's for also people who are actually affected. And most of the time people respond or want to negotiate with you and then you can have a really good brand moment where you go, "No, no, no, don't worry about it. We'll talk in 12 months. We'll talk whenever, you just figure out your stuff right now."
Join our mailing list to receive the latest Library updates. After subscribing, tell us your preferences to receive only the email you want.