Hola a todos. Mi nombre es Jordi Romero y vamos a hablar de MPS, CSAT y churn messages.
Voy a ir un poco más allá de lo que significa en un segundo.
El título original era algo sobre los clientes ir primero y el feedback y todo,
pero lo cambié porque al final del día, estos son las herramientas,
los mecanismos, los clientes o el feedback, pero lo que realmente queremos hablar es la transformación
que un startup llamado Redbooth salió, y cómo implementamos la estrategia,
y luego algo que no estuvimos tan bien como queríamos.
Y luego, por estos mecanismos, pudimos cambiar la estrategia para algo que era mejor.
Cindi already did an introduction. I'm Jordi.
We just started Factorial a few months ago.
It's a software company that offers a platform for HR managers in small and medium companies around the world,
and we also offer the possibility to hire employee benefits from the platform.
So basically, small companies can now manage all of their HR staff in this platform
and also hire things like health insurance, gym memberships, ticket restaurant,
and stuff like that through the SaaS platform.
Before that, I was at Redbooth, and that's actually some of the stories we're going to be talking about.
Redbooth is a project management platform in the cloud.
It's a company that got started in Barcelona around 2010.
So it got started here, then we moved the company to California,
legally moved it there, raised some money there.
It raised up to Series B in financing, and we went through a lot of evolutions,
and we're going to be telling some details later on.
And then I started there as a CTO, Chief Technical Officer,
and then I became the VP Business Development and Platform,
which meant I was in charge of partnership, strategy, and integrations with other companies.
So you can find me on LinkedIn, Twitter, or you can go check factorial out at factorial.co.
So, one disclaimer, this presentation I prepared for today,
but also for an event that happened this weekend, this past weekend in Berlin,
which was a SaaS Founders event organized by a venture capitalist called 0.9 Capital.
So at this conference, there was a gathering of founders of software as a service companies,
and among them there was a lot of people from Zendesk.
Zendesk is one big success story of SaaS, of our service,
and the CEO and some of the top executives of this company were there,
and they were sharing a lot of insights of how they came from the inception
to the successful public company they are now.
And I just wrote down some quotes that they said, this was last Friday morning.
So, Mikkels van was highlighting the success of Zendesk,
because basically they were able to instrument and mechanize the relationship
and the communication between customers and the companies that use Zendesk, the platform,
and capture that right now, for instance, ratings are ubiquitous,
which means that all businesses and all services are using some sort of rating
or qualified benchmark to see if you're doing good, if your aging is doing good,
like you take an Uber and you rate the driver, you go to Airbnb, you rate the host,
you rate the location, you rate the cleanness.
Everything we do right now has immediate quantitative and qualitative feedback,
and that through rating, you can kind of reinforce or detect how well you're doing
in terms of loyalty, which means not just delivering a service,
but alling the customers, making sure they're not just satisfied, but amazed with your service
and they will come back and they will promote your brand and product,
which is very important.
The SVP of marketing, the senior vice president of marketing at Zendesk
was also highlighting some metrics about this,
so I just thought it kind of helped set the stage.
So, what I wish we kind of talk about today,
and please interrupt me anytime, ask questions,
share your experiences, whatever, say your name and what you're doing
so we can all understand where the comments are coming from.
So, we want to talk about how to increase a company's value.
I'm going to talk about this case, what we did in Red Booth,
and in order to get there, we're going to look at some intermediate objectives,
such as the revenue per account,
like how much money we're getting out of each customer,
the attrition, which is the loss of customers over time,
or increasing engagement,
so how much customers are using our product or service.
Some of the outcomes of listening to customers to positively impact these metrics
would be what can we do differently in our product,
what should be our next developments in our product,
how can we change the way we relate to our customers,
like what's the customer experience
needs to be for an increased value,
or just service overall,
like in case our product also has a service attached to it,
like what changes can we do there,
and we'll take a look at that in a second.
So, I made up this framework,
don't pay a lot of attention to it,
it's just an excuse to kind of follow a flow
and go step after step,
it's five steps to actionable customer success.
So, let's pause here for a second,
I think there's two important concepts here,
one is actionable,
and the other one is customer success.
So, customer success, by that,
like it's a very hot term right now,
a lot of people use it,
it can mean a function in a company,
it can mean a department,
it can mean a strategy to relate to customers,
in this case it's kind of going to mean
the whole group of concepts,
it's how our company interacts with customers
to guarantee their success after using our service.
So, our goal is not to make money,
I mean our final goal is to make money,
usually as a company,
but in order to get there,
and not just make money but grow
and get these existing customer base
to help us grow even further,
and make these customers successful
at whatever they do.
So, for instance,
if our platform is project management,
we want to make sure
our customers' projects are on time,
they're either increasing performance
or cutting costs
or delivering a better service.
If our service is HR, for instance,
we want to make sure
our company's employees are more satisfied,
they are staying longer in the company,
they're recommending their friends
to come work for them, etc.
So, we're going to focus on how to increase that.
Through customer success,
so asking and relating to these customers,
and by action of all,
I mean that all of these steps
we're going to follow here
should have an outcome,
should say, okay, we learned this,
we measured it,
we have a conclusion out of it,
now this is the next step,
you have to implement this new technique,
you have to change your product in a way.
So, first step is asking,
we have to open channels
of communication with customers,
we need to get the information coming,
then this information
might be in different languages,
different formats,
it's unprocessed and unstructured,
so we want to make sure we do that.
Then we want to connect this
to our existing company,
KPIs or metrics,
so we want to know
what's impacting what here,
so if all the customers
are complaining about
a part of our product,
and if we fix that,
what metric,
what important metric
is being affected by that.
Our customers are staying longer with us,
because now our product
addressed an issue,
and understanding is kind of
going to give us
the segue to acting.
So, once we're measuring things
and we know the impact
of different events,
we can propose solutions
and then we can actually develop them.
Okay, so the first one is fairly easy,
let's say we have
a software as a service company
or any software company
for the matter,
we have people coming to us
using our service,
and then we want to make sure
they have the chance
to give us the information
that we need from them.
There's a couple of important things
to take into account here,
is who should we be asking,
and not just opening up
to everybody,
but knowing exactly
who we're asking at the point.
Is this a customer
that's actually using
the product right now?
Is it somebody who's pissed
and just sent a support request
saying blogging doesn't work,
or is somebody
who just canceled their account?
It's very important
to open communication channels
to all of them
and also categorize accordingly.
What do we want to ask them?
So, I mean,
you have some feedback,
that's awesome,
give us the feedback.
Is there something especially
that makes you really love the product?
I want to know what this is
to give you more of this,
and if there's something
you don't love about it,
or you wish you could receive
from our product
and then you will be a promoter,
that's even more important
because then we can extract
this information,
reinforce it,
and convert a passive customer
into a promoter.
And what problems are we trying to,
are you trying to solve
with our solution?
What do we need to do
for you to be successful?
When do we ask this?
I mean,
it's heavily related to the who,
so there is different interactions
where we want to be asking questions,
and here is
with all these frameworks coming, okay?
So the title had MPS,
CSAT, and chair messages.
These are concepts
that, like,
I didn't come up with them,
they're fairly well known,
best practices in the industry.
The first one is customer satisfaction.
Customer satisfaction
is a transactional feedback,
that you get at certain events
in the life of a customer
with your product.
This is an example,
I don't know if you recently went
to an airport,
almost all of them in the world
right now have these devices,
so after you go through security check,
you can, if you want,
press one of these four buttons,
and they use this to sometimes
measure the performance of the actual agents
that are checking you at that shift,
or, you know, they're validating
if a new practice,
like a new scanner or whatever,
is making passengers in this case
happier or less happy.
In our software,
we can have the same,
for instance,
when a customer sends you an email
saying login doesn't work,
or I tried to do this
with your product,
then I couldn't do that.
After this interaction,
even if we resolved it or not,
it was our fault,
maybe it was their fault,
we're gonna ask them,
like, how do we do here?
Like, are you satisfied
with the resolution of this transaction?
I mean, the wording can be a little bit better.
How would you rate
your overall satisfaction
with the service you received?
Okay, that's better.
And then this, typically,
either one to five,
four smiley faces, etc.
And then we measure this
as either an average,
like from one to five,
and then we track 3.7,
or we just count the satisfied
or top satisfied accounts,
and we say, okay,
75% of our interactions are good.
In SaaS, for instance,
this is something typically
SaaS companies,
software service companies
don't have a problem with.
A good C-Sat is 98%,
an okay C-Sat is 90%.
Anything below that
means you probably have a problem
in your processes,
and you should go ahead
and fix them.
If you're in the phone industry,
for instance,
that's a different story,
because they go after cost optimization,
and they have massive scales,
and then they're usually
probably happy with a lower score.
This is something,
it shouldn't be anything
that a software company
thinks about this too much.
Like, it has to be perfect.
It has to be 95% or better.
If not, you're screwing up,
and there's bigger problems,
as we'll see.
This you have to nail,
and it's easy.
Just hire very good
customer support people,
you know,
implant some process,
use the good tools,
and this will come.
Okay, this in my opinion
is the juiciest mechanism
we're gonna be talking about.
MPS stands for
Net Promoter Score,
and it's a tool that
grow a lot in popularity
in the last few years,
and I'm sure you've had
this question many, many times.
It's like,
would you recommend this
service or product
to a friend or colleague?
You get this,
for instance,
a receipt from an Apple store,
like a physical store.
In the bottom of every email,
there is a link to answer this.
If they go below 98%,
they might fire somebody
from the store that month.
Like, their Apple
is crazy about MPS,
for instance,
and the Apple retail stores,
they measure almost exclusively
by this metric.
They're almost religious about it.
Maybe a little bit too much,
but it's very important.
We're gonna see
that this score is very,
very negative,
meaning that you can answer
from 1 to 10,
and the customer
doesn't see the colors.
The customer just sees
the numbers
or the runs to click in.
This is a 10.
This is 9, 8, 7, 6, and below.
9 or 10 is good.
It means this customer
is a promoter.
So, it really loves
your product
and would recommend
this product to their friends
or colleagues
or professional network.
This is, for me,
the super painful site.
This is a passive,
which means that a 7
or a 8,
which I thought
was a pretty good grade,
means nothing.
It means
you're not doing
anything at all
for this customer.
They don't hate you,
but it's totally useless
what you're doing
for these customers.
You're not satisfying them.
And anything below,
it's negative.
It's gonna subtract
from your score.
So, typically,
what we see is,
okay, some people
are really pissed at us.
We're not solving any problems.
We're not making them successful.
Some people love us,
and typically,
before you start
taking care of this,
a bulk of your customers
are gonna be here
in the middle,
which sucks,
because the formula
is gonna say,
count the percentage
of customers
that you have here.
So, let's say
20% of your customers
are voting...
No, I don't want
to change slide yet.
Let's say 20%
of your customers
are super satisfied,
and 30% of them
are here.
Can somebody do
the crazy math
and tell me what's
your MPS score?
Exacto.
And it sucks.
Having a minus 10
in MPS,
actually,
you're gonna see later,
this can happen,
but it's not good.
But it's so easy,
because there's
so many of these red dots here.
So, actually,
it's kind of tough
to move it up.
So, it's a very strict score.
It's very useful.
Now, how do you actually
implement a mechanism
like this
if you're somehow
a digital
or software business?
There are some tools
that do specifically this.
Woodrick and Delighted,
they are software-as-a-service
platforms
that only do this.
Basically,
they have like a widget
that you install
in your application
and you configure
some rules.
So, after some time
or some interactions,
the customer's gonna see
like a pop-up
that says,
hey,
from 0 to 10,
would you recommend
this service
to friends or colleagues?
There's Zendesk,
actually,
that is a much
broader platform.
They do customer support,
voice,
phone support,
they do metrics,
they do insights,
they do a lot of stuff,
and they also
do things
for bigger companies
called Gainsight,
which is more
specialized
in customer success.
And there's always
the option of
building your own stuff.
And I'm gonna tell
more about this
because we actually
built our own stuff.
So,
some things to take
into account
if you're in the market
for one of these solutions is,
well, the main goal here
is you wanna score,
right?
Are you doing minus 10
or you're doing 40?
You wanna know that?
And you wanna collect
the comments
because as we're gonna see,
the quantitative aspect of it
is like a guideline
you can follow,
but the actual comments
are the input
that you're gonna use
to change the score.
Without that,
you're lost.
So,
you wanna be able
to obviously measure
the percent of responses
by type,
like buckets
of detractors
and promoters,
and passives
we don't care about
the score,
and the response rate.
Response rate means
how many people
get impacted
by this question,
how many of them ignore it
or answer it.
So, for instance,
at RedBooth,
we used to have
Zendesk, actually.
So,
once in a while,
we send an email
to all of our active customers.
Not when they're using the product,
not when they're having a problem
at the neutral point,
like just Monday morning,
they would all get
an email asking,
hey, you're using RedBooth
for a while now.
Could you rate from 1 to 10?
What's your,
would you recommend,
blah, blah, blah?
And we saw
a response rate
of like 5 to 10 percent
then we moved this
into an in-app banner,
like when you actually launched
RedBooth,
you had there,
you could dismiss it,
but you had there
the banner,
and we multiplied
the response rate by four.
So,
like the tool you choose
and where you show it
makes a big difference.
Things you can think about
is actually the placement,
what I just said,
like an email versus in-app,
versus mobile,
versus whatever it is,
or physical real life,
like the airport example,
and the frequency.
The data ownership
and thinking thing
we're going to talk more
about later.
So,
we were using Zendesk,
actually,
we had some results,
we got some score,
we got worried
about the score
because it was actually
minus 10,
and we knew
we had to do something
about it,
and then we wanted
more data,
more frequent data,
and more segmented.
So we wanted
to be able
to mine this data
by type of customer
when we acquired
this customer,
what product they're using,
basically crossing
with all the information
we had.
So we decided
to break this in-house.
We built
a small open-source
product,
actually,
that you can install
in your web application
and shows in-app
after 30 days
for existing customers,
and then
from that point on
it repeats
every 90 days.
You can configure that,
of course.
That's kind of like
what we saw,
is enough information
without bothering
too much the customers.
One,
of our goals
was plugging it
into our own BI,
like our own business
intelligence tool,
so we could have
a unified dashboard
with MRR,
like brand new growth
metrics,
and MPS,
and correlate everything,
and integrated
with CRM,
like Salesforce data,
Zendesk data,
whatever,
HubSpot,
GainSight.
So as I already said,
we increased
the response rate
for X,
and we went from
having once a quarter
via email
to having it
in-app
following this formula.
So that's what MPS,
this is a timeline
of MPS scores,
or MPS service,
after time.
These are more or less
real data.
I removed some events
to make it easy
to follow,
but basically,
what happened here
is almost the case
I said before.
We had a 25% promoter
base,
40% passive phase,
which is pretty common.
Most people say,
yeah,
you're doing a pretty good job,
seven, eight,
that's okay,
doesn't matter,
and 35% detractor,
so six or less,
which means
you really have
an issue
and you have to figure
something out.
So that's a minus 10,
just here between
seven,
five,
and 15.
So we're going to look
later at how we
changed things,
like how we
basically
detected some challenges
that these detractors
were having
and addressed them
right away,
like get the hottest topics
and go fix them.
But what was really
delighting our promoters
and we did a little bit
more of that.
So what is the time frame
from the first?
These are months.
These are months.
Each bar is a month.
A month.
Yeah.
Yeah, yeah.
So MPS
is something you can
change quickly
if you have hot issues.
Like you're really
screwing up somewhere.
For instance,
everybody is telling you
to do something
and you don't do it
because you believe
and they really hate you
for that
because it's like
you should just do that
and then you do it
and then you're like
boom, boom, boom.
That's it
and that's actually
what happened.
So this Red Booth
project management software
was satisfying
some customers
like they could manage
their tasks and so on.
But there was a common issue
actually.
They wanted one kind
of report
that's called the Gantt report
and everybody was
asking for it,
like for years.
We even had it
and we removed it.
And at some point
we did this.
The score was bad
and they said,
okay, let's do something
about it.
The team built
prototype of the Gantt chart
was exposed
to some of the people
that told us
they cared about it.
They were immediately
very satisfied.
Then it got released
to everybody
and a bunch of people
went from passive
to like, yeah,
I can manage my tasks
to awesome.
I can finally run
the report they always wanted.
So that's a way
to get a quick bump.
As you can see
there are some other issues
that are long term
and they take
a whole company
to change
to be able to address them.
So that's how
the picture can change
from an extent
to a 35, I guess.
Any questions?
Cool.
Okay,
so now kind
of moving away
from NPS
for a moment.
Oh, sorry.
I have a question.
Can the negative
stop being influenced
by, I don't know,
people being dissolved
by your banner,
by your email?
Can this be influenced
by being detected?
Should we move it around?
So the question is
if the tractors
can be influenced
by banners
or other
marketing campaigns
or other interactions,
yes, totally.
Like these measures
the overall service
that you're offering
this customer.
If you're annoying
this customer
all the time
with retargeting,
for instance,
and you should not
do it,
in our experience
it's much deeper stuff
that gets in here.
Like,
people tend to care
like if we're trying
to make them productive
with a tool
that's what they care
the most about
when they answer.
But if you have
terrible customer support
people
and it's time they call
they like disrespectful
or, you know,
not helpful
and so on
it's going to impact
this heavily.
That's when it's good
to have
different metrics
like CSAT.
You can detect that early,
get that thing
out of the picture
using MPS.
The particular case
of retargeting
I don't know
how you can segment this
out.
If you could cross it
in your data set
for instance, say
people who are most
heavily attacked
by retargeting
have a lower MPS
than those who are not
that's going to take
some data mining
and somebody in the team
who's really checking
all the data
against different sources
but I guess you could
detect it.
But it's a tricky one.
Cool.
So, quickly into churn
messages.
You lose for
lost business
so you acquire
a subscription
and at some point
you lose this revenue,
you lose this usage,
you lose this customer.
That's a churn customer
and when this customer
leaves
it's very important
that you take as much
information as you can
from this interaction
because
they're passionate,
they are
definitely pissed,
definitely disappointed
by you
so they really
had expectations
and you spend some money
you're proud to
spend time adopting
training people,
convincing, spending
social capital
business
and they didn't succeed
so that's pretty bad
but at the end
of the day
if you send somebody
from let's say
your product team,
marketing team,
customer support
or success teams
to go talk to this person
and say hey
first of all
I'm not going to try
to sell you
anything again.
You know like
you have to be very honest
like one thing
we noticed
we screwed up
is when you lost
a customer
because your product
didn't
solve any problem
for them
and this person
sold it
and that's horrible
because like
if you didn't solve
a problem well enough
to pay in this case
5 bucks a month per user
like it's not about
20% discount
it's about understanding
what they were trying
to do that
you failed at
fulfilling
so yeah
pretty straight forward stuff
this was very useful
as long as you
collect all the data
segment it
and cross it
with the others
as we're going to see later
so we open communication
channels
we're listening to customers
we have many ways
for us to listen
to customers
it might come
even in many languages
it comes
like it's free form
a lot of times
like there's an MPS score
which is a number
but the most
important information
is the actual message
they attach to it
we need to process this
so things that we actually
did that were really useful
and once we started doing
this like
actually MPS meant
something to us
so they were
oh I understand that
oh hey that's good
no it's actually
people are complaining
about performance
they're complaining
about our mobile apps
like for instance
nobody like
mobile app ratings
tell you stuff
but when people
evaluate your entire service
and the thing
they are most
most pissed about
is the mobile app
and the performance
and the UI
that means
you should seriously take
that into account
and we're going to see
that's a real example
and another thing
that we started doing
and I saw having
really good effects
in the
collaboration
between teams
was sharing this information
so usually
some
well not usually
sometimes you have
product people
looking at this stuff
sometimes you have
customer support
the success people
looking at that
and then
they do their metrics
they have their flashboards
maybe they report to their boss
and they say
look we're going up
we're going down whatever
and nobody really hears
about it
or cares about it
one thing that we saw
made a big difference
is when
the people responsible
for this
collected all this information
kind of
drew some charts
and shared some information
collected all the feedback
into categories
and then shared it
at an all hands
and the whole company
would see guys
we're doing okay
we're doing great
we're not doing so well
people love us
because
our mobile applications
but they hate us
because of our performance
like
everybody understands this
everybody has a shared mission
and objective
and like
you can have
like
for instance an engineer
that wouldn't have this input before
come up with a solution
say
I was thinking of this refactor
I didn't propose it
because I thought it was not relevant
but if our biggest problem is performance
maybe it's a good idea
to work in this refactor
so actually
this helped a lot
this is an example
is actually a real example
of the components
as we call them
which is group
all of the feedback
that we get
into different
either product components
or root causes of contact
so product components
are parts of your product
or service
that the people
might be either happy
or unhappy about
y root causes of contact
are
why did they
come back to us
in the first place
if it's MPS
it's neutral
if it's customer support
why are they writing us
we had a lot of
this is a
this is components
so we don't have root causes here
but we had a huge problem
with billing inbound requests
so
a lot of our customer support time
was handling somebody who said
I wanted to add a new user to my account
but the credit card didn't take it
so I had to do this and that
or I wanted to
change credit card
and it didn't work
and it's not just inefficient
because you have people
actually solving stuff
that should be automated
these people are dissatisfied
because if they have to do something
then it doesn't work
then reach out to us
have to figure out etc
it's a problem
so it's good to categorize root causes of contact
and components is critical
because you know what's working
and what's not
in this case
this was
the tractors components
so
usability was a big problem
usability
is a terrible component
because it's too generic
it doesn't really give us an insight
on what to fix
but that was the product
product management team's job
to go after these people
and say
hey, you're not so happy
with the usability
what are you trying to achieve
like
what Red would not do so well for you
and then after some interviews
we got this information
and then one after one
the team went like
engineering, marketing, product management
they went one after one
and nailed them
and you see the NPS score
going up and up
ok, so we basically categorized everything
and segmented a little bit
the causes for contact
or satisfaction
or dissatisfaction
the next thing you want to do
is connect these measurements
and these sources of information
with your business metrics
like
the beginning of the presentation
we said our goal here
is to increase company value
ok, that's
in general
every executive's job
in a company is to increase the company's value
how do we do that?
well, we can take these intermediate objectives
and say we have to increase
our
average revenue per user
like we have these customers
we want to make more money out of them
or we need to reduce the attrition
we don't want to lose customers
so how do you connect
these measurements
these sources of information
these root causes of contact
with this
basically
dump all the data
somewhere where you can query it
so data warehouse
is like a big database
where you can connect everything
and then
graph it out
come up with conclusions etc
process it
so it's all in the same format
you can actually
cross
and plug this in
into some dashboards
charts etc
so that would be what BI stands for
the rules here
something we learned
the hard way actually
define company standard segments
like
who knows here
what an SMB is
can somebody say
what an SMB is
small-medium business
ok define SMB
by number of employees
how many
I don't know
up to 500 it's million
ok
before 50
ok
somebody has a different number in mind
depends on the country
depends
exactly
so it's a mess
it's a huge mess
so the first thing like
we have a problem with SMB
no that's not true
SMBs are doing very well
well it depends
what's an SMB
so like
somebody decides
and says
this is an SMB to us
it doesn't matter
if everybody has different opinions
we're gonna go with this assumption
from now on
don't just connect everything
real time
and send reports to everybody
to look at
unless they have information
that's a common mistake
we did as well
and always track your own progress
that's something we've done
and I've seen a lot of people do
which is
have a problem
create a dashboard
look at the dashboard
come up with some conclusion
and then go home
and then the next time
you have a new problem
you create a new dashboard
you look at the dashboard
you come up with some conclusions
and then you go home
but when you actually put them together
it's like oh
that was totally wrong
like each time you want an answer
like if you want your BI
to tell you something
it's gonna tell you
whatever you want it to tell you
so what you have to do
is just define some standards
and let them evolve
so that's very important
otherwise we're gonna lie to ourselves
this is just a couple of screenshots
of two BI tools
I've used
at Redwood
we use this one
I've been playing with this one now
this is chartio.com
and this is a mode analytics
I actually recommend Chartio
so the way this works
is basically
you connect data sources
so databases basically
that fit data into this platform
and then it has a very easy
drag and drop interface
to create whatever graphs
or instrumentation
you want here
it's smart
so you can do a lot
drag and drop
you can share definitions
like if you say
this is an SMB
so you can kind of
create a canonical definition
and then expand it
or you can go hardcore
and edit the SQL yourself
mode is kind of more
do it yourself
but it's also cool and it's cheap
ok so the
second last step here
is we have information
we have customer feedback
that's coming through the process
we're processing it
we're segmenting it
putting it inside of buckets
connecting its evolution
with business metrics
now we need to come up
with some insights
some ideas
right
so for me this is the most important
the most important
it's like
here's the happy idea moment
it's like
this is what we need to do
you saw a pattern
you found a pattern
you validated it
the whole team looked at it
there's some proposal
you share this information
with the team
they come back
with some conclusion
and then you brainstorm
some solution
you make sure
the solution is
like compatible
with all the teams
and you kind of define
ok we're going to do this change now
this change is going to have
an impact in this segment
and that's going to fix XYZ
this sounds very abstract
so that's where I want to go deep
is I want to tell you
a horror story
this is
well you're not going to say anything
until I explain you
so this is
the result
of our
first NPS
after we realized
we had a problem as a company
we
at Redbooth
offered this project management
software in the cloud
so people came in
they pay for it
they use it
and they go
so this is called churn
churn is a big problem
in subscription businesses
it's especially big
in collaboration businesses
there's many reasons for that
we need to go deep
and it was a problem for us
because we were spending a lot of money
to bring customers in
we were spending a lot of money
to develop a product
that they used
and then they left
before we could get enough money
out of them
so the whole machine
was like a leaky bucket
you put water in it
and then you walk three steps
and you're out of water
and you put water in it
and then you're always without water
and that sucks
that happens in general
but it happened a little bit more
to us
than it happened to other people
so we knew we had the problem
we had to address it
we ran some MPS scores
we saw a minus 10
we have a problem
it's good
there's a low starting point
so we can go all the way up from there
and this is the process
that actually the product management team did
processing this information
so one of the first segmentations
the team did is
we're going to segment it
by the role of the person
that answers this MPS survey
so in red booth
you can be the first person
that comes in
and creates an account
this means you really care
about that
because you're looking
for project management solutions
you find two or three
you sign up
you set it up
you invite some people
and you're a customer
okay, that's the creator
then you might say
okay in the company
it's not just me managing things
it's two or three more people
so I'm going to invite them
and I'm going to make their admins
and then there is a bunch
of other people
who actually do the work
that gets tracked in tasks
and those are the participants
so this is the whole team
using red booth
when we segmented
the MPS score by role
we saw that the person
that thought this was a good idea
in the first place
they were happy about it
of course they chose it
so they're super biased
if they said this was terrible
it means they're admitting
their own mistake
and people don't like to do that
so they said
this is a pretty good tool
I chose it of course
it's a good tool
their peers said
it's okay
I have some reports
like people do the work
and now I can see it
which is better than asking
this is done yet
this is done yet
no I have my reports
project management is cool
but it's not like
rocking my world
and then there's this poor guy
that suddenly gets
tons of emails
and tasks
and things
saying do this, do that
do this other thing
and everything's out of control
because we spend a lot of time
building stuff for these people
and nobody really cared
about this poor guy here
and it's the essence
if work doesn't get done faster
and if people go against the tool
it's not gonna go in favor
of customer success
so this is the problem
how did we address this problem
well
we detected there was a lack
of functionalities
there was good reporting
and there was good tools
for the managers
to kind of see things
from the top
but when you receive
a lot of inbound requests
you wanna see
what am I supposed to do next
this is tricky
this is one of the examples
that didn't get fixed
in two weeks
this was feeding input
into our product roadmap
and the team is still
working on them today
and this was
a year, almost a year ago
when we run this
so a year after
we still executing this
because this means
changing your product
and changing your product
is not something
you do overnight
especially when you're like
6 year old
you have a team of 60
and a lot of customers
you're gonna just change
everything overnight
you need to evolve
so that's one thing we saw
then things started getting
really ugly
because this we kind of
okay
it's just product
we will fix it
but here it gets ugly
because
this is very hard to understand
at first
side
this is one NPS survey
one
it was done
at the end of 2015
one day
we send an email
to 4,000 companies
I don't know
how many employees
to everybody using red booth
and we said
would you recommend red booth
to a friend or colleague
from one to ten
and we got the score
so we're segmenting them here
by when they paid
so we had some customers
here
that have been with us
for two and a half years
or more
or three, four, five
we have some customers
that have been here
for one and a half years
once that just came
earlier that year
and once that we just acquired
so there is an obvious
evolution here
and we knew
what happened
between these years
and I'm gonna tell you
because then you start
realizing a lot of things
right
in 2013
and before
we used to be
a freemium product
so a product
that offered
a decent service
for very small companies
for free forever
and we had some growth
like people came in
used it
many people didn't pay
for it
and that was okay
because they recommended
it to other people
and they liked it a lot
and some paid
because they were bigger
companies
and at some point
you had to pay
and that was okay
we saw that this was
not very efficient
monetization wise
so we decided to make
a change to
paid trial
so we had the 30 day
trial period
that's free
and then you have to pay
it doesn't matter
if you're one person
five or 500
after 30
days
there is no free product
so we killed
all the virality
of a freemium product
there
but in exchange
we hope to get
some more money
because people couldn't
use it for free
so they have to pay
even if they were small
and then later on
we saw that it was
very hard to build
a growing business
with very small
accounts
one at a time
because it costs a lot
to acquire them
to serve them
and they're only paying
like what $40 a month
so we decided
to go
a fish bigger fish
basically
so during 2014
we said
we're going to move
from the SME
to the meat market
we're going to
stop
fulfilling the needs
of the smaller companies
we got
they don't give us enough money
we're going to go
after bigger companies
are we going to
these bigger companies
they don't just come
on their own
so we need to change
from a freemium to trial
and they don't just
figure things out themselves
they need a sales rep
they need a person
that sells this product
to them
so we did
the whole
strategy change
new teams
new funding rounds
new people in the team
changing roles
in the company
etc etc
and we didn't see this
before we were flying blind
and at some point
we saw this
and it's like
okay what's happening here
the people that came
on their one promise
on average
they're 30
that's pretty good
the people that came
when we were
we stopped caring
about the small companies
and we started
we started caring
only about this
meat market
like a few hundred
y so on
didn't like us as much
and when we really
started only selling
through the sales team
doing a training
like acting enterprise
then they actually
didn't like us at all
so that's very bad
when our average
was almost minus 10
that was very bad
so at some point
this year
we started realizing this
and there were a couple
of experiments
where we said okay
something's not going well
here
let's start
rolling back
let's start undoing
some of the things we did
particularly only
so we're going to take
a segment of customers
any company with
200 employees or less
is going to go
totally autonomous
nobody's going to talk to them
it's going to be self-service
they're going to have the features
that we used to have
and we removed
because enterprise
don't want these features
and so on
and we saw that this
actual segment
was doing pretty well
and then the last
part of the story
is when we went ahead
and said okay
let's do one thing
let's query the database
and separate everybody
that somebody talked to
and everybody
so there's no touch
means people that came in
on their own
they figured out the tool
they figured out the business
the price
everything
and they set it up
and sales touch
means the people
that they were blocked
from self-serving
and they had to talk
to a human
and this human
had the monthly quota
that was probably
not well set up
and this human
wanted to convince them
that this was a good tool
for them
so what we saw is
these people
okay we were not growing
as fast as we wanted
here
but they were happy
what they found
and they chose to use
was going pretty well
these people here
they were forced
to use something
and especially
what we detected
is they were over promised
a lot of things
so the sales people
heard the message
that we're going to go enterprise
we're going to be building
all of these features
our product is not for
small companies anymore
we don't have it yet
but it's coming
so a lot of people came in
under this promise
and this promise
didn't really get delivered
because the transition
didn't go well
and what we did here
was both the go to market
and the strategy
didn't go as planned
at this point
we stopped everything
there was a drastic change
in the business here
it included
one VP getting fired
and like
eight to ten people
moving on
we stopped having
the sales force
like the people
that were selling the product
over promising
and pushing this
down to people's throats
and went back to
a no touch model
changed the metrics
pretty well
soon
and then we started
implementing some of these
changes like progressively
so when we go back
that's my laptop
when we go back
to the other thing
right
so this is long
this is low
like
we started the experiment
okay
we see some really good results
we're gonna keep going
we go more drastically
now everybody's
no touch
there's nobody trying to
oversell them
trying to push
for a bigger quota
trying to make them
like what we would do
is we would go
from somebody
who is
at a
at a six
or a seven
in MPS
so the kind of
not really loving us
they're using our product
for ten people
and we would say
hey
invite twenty more people
and go to our more
expensive product
and they said
fuck you basically
because they're not even
happy in the first place
and we're
pushing more sales
down to their throats
so
that was a big learning there
and basically
the conclusion here
our example is a little bit extreme
it's not just
tweaking your product
we actually have to change
the company strategy
I think it's interesting
to share
but usually what happens
is you go back
to all your customers
it's like okay
you said
you don't like the user experience
or the interface
tell me more about it
what was your final goal
you know
what were you trying to accomplish
how does your team
try to do this
without redwood etc etc
so we improved
some of these experiences
and interactions
with the customers
we obviously
fit things into the product world map
right now at redwood
all of the inputs
to the product world map
come from this process
it used to be kind of like
happy idea
like
I think we should do this
what do you guys think
yeah I think it's a good idea
so let's do it
not anymore
so now it comes from this process
one interesting part is
when a team develops something
it goes into a better feature
which means
it's not automatically
released to everybody
we selectively
deploy it to some people
who do we deploy it to
people who told us
they care about it
so
they wanted to get
something out of this
they're going to be
the most qualified feedback
out of this
they help us tune it
then it goes out to everybody
and everybody's a little bit happier
and eventually
there's some processes
like customer interaction
if your sysad is very bad
then you want to tweak that
you want to make sure you answer fast
you don't delegate
and then delegate
and delegate again
and then come back
oh sorry the technical team
is looking at it
like improve this process
a little bit
it's not work in science
but you have to do it very well
you need good people there
do some retrospective analysis
like
you know
what are we doing wrong here
what did we do wrong
during all this process
let's look back
let's look at the evolution
let's be honest about it
and in our case
it led to a big acknowledgement
our strategy is not working well
we have to change
our old strategy
which means for a start up
with venture capital
I don't know if you saw in the first slide
it said something like 22 million
dollars in venture funding
it means going back to your board
of directors
and said remember when we told you
we were going to do this
and it was going to be awesome
well it's not going very well
we have to undo some of these things
and we need more money
so that's a tough conversation to have
and you need to be very sure about it
so here's what I already showed you
happy ending
sort of like
a lot of things improved
you know
still a lot of work to do
but we did fix a lot of stuff
and I'm going to quote
actually my colleague here
John Soninshine
he is the one that led
many of these
changes in the company
like he saw something was wrong
with churn
it used to be they had the product
and as I had the product
the CEO would say
your churn sucks
go fix the product
and he's like
okay I'm going to change
something in the product
and churn still sucks
and then we said
no that's because marketing
is bringing in the wrong customers
so let's bring customer
let's bring marketing for a while
and then marketing said
look I'm just doing the same as always
and suddenly they didn't like it anymore
and then we say
okay that's because sales
are just pushing stuff
down their throats
because their company is wrong
and eventually it's like
okay let's use some science here
let's listen to our customers
let's come out
with all of these conclusions
and he quoted this
in a medium post
that kind of talks
about this transition
we cut our net MRR churn
by 50%
that means the percentage
of revenue we lose
every month
cut divided by 2
which is pretty good
so we went from being bad
in the industry
to being above
the industry standard
so now we're actually
like there's always room
for improvement
but it's not our biggest problem
right now
and important thing
this happened to have
a kind of happy coincidence
or happy side effect
which is our lifetime value
to cac ratio
so how much we get
out of our customer
versus how much we spend
to get this customer
improved 5x
why is that?
because customers now
like our product more
so they stay longer
so they pay us more money
because it's a monthly fee
and since
we happened to get rid
of a 10 people sales team
that was a lot of money
we were spending
to acquire customers
and now this became
automation systems
in online marketing
and drip campaigns
and customer success process
and so on
we spent a lot less
to get a bunch more
so it was like 2x
2.5x
and combined is 5x
which is pretty good
so because your companies
are now bigger companies
so they pay more money
no, that was the fallacy
that's what we thought
they ended up being the same companies
they're just more satisfied now
that's what we thought
what happened
but it's a longer story here
about why bigger
or smaller companies
use some collaboration products
or not
the truth is this product
Red Booth
is very good
for a size of customer
for a big kind of customer
this level of transparency
and freedom
doesn't work well
and we thought
we would leave this change
but we didn't
and then we kind of
reverted some of these assumptions
based
and backed with data
and then
went to what we knew
how to do
but actually doing it very well
or better
so yeah
it had a happy ending
so that's everything I have
I hope you have some questions
and yeah
¡hasta la próxima!