Xavier: She said, swap,
swap, something like that.
It was a bit funny and then
actually the brand, they were Okay.
I, I'm not sure I want to keep that.
they ending, they ended up keeping
it it's like, just because of that
boom, 350 million view, you know,
Jacob: Yeah.
Xavier: AI could predict that.
No AI could actually do that.
They would have pronounced
the name perfectly,
First you need to have content
that looks authentic, right?
Um, second, it needs to be qualitative,
um, and to have a good script because
if you don't have a good script, usually
creator will be lazy and it won't work.
But still with that, even if, with
the good creators, with, uh, authentic
content, with good scripts, um.
The, you, you still never know
which one is gonna be successful.
Me, I still have no clue when I see an
ad if it's going to be successful or not.
I mean, no one can predict that.
Jacob: Hey Xavier.
Uh, I'm really excited to
have you on the podcast today.
You know, UGC is such a hot topic
in, uh, the mobile app, subscription
world, and user acquisition in general.
So I'm really excited to kind of learn
from you and, and get your insights.
Xavier: Thank you for having me.
Very happy to be here.
Jacob: Cool.
So I guess before we dive into insights
and, and sharing your knowledge.
Uh, and I know Rom dom isn't exclusively
focused on mobile apps, but it looks like
across your career you've had quite a
few stops that did involve mobile apps,
from Orange all the way to Match group.
I, I guess if you wanna give everyone
a quick, uh, uh, backstory on how
you ended up founding rom dom.
Xavier: Yeah, of course.
So I'm Xavier, um, I'm the
CEO co-founder of Random.
But before Randa, I actually
lived in the intersection of
music, tech and consumer apps.
I started my career at Universal Music.
And then, um, launched an app
that has been acquired by the
Match Group dating app, obviously.
Um, so we stayed at Match with,
with me and my co-founder for five
years working with, uh, AI teams.
And then we left Match a few years ago and
we launched, uh, start, um, an app studio.
um, RDA was a pvo.
Uh, let me tell you the
story of rda actually.
We launched a few apps and one of
the app was, was an app for, uh, a
car search app for us millennials.
uh, we managed to scale this app
in the US thanks to one TikTok.
A alright, so actually spend months,
uh, trying to get tiktoks doing
some short promo video of my ads
using it as a, as a creative part
of my social media ad campaigns.
Took me a while to get, at the
end of the day, one creatives that
divided my cost per style by five.
And then actually we, we,
um, we were profitable.
We became actually one of the top, uh, car
search app in the US for various reasons.
We had to kill this app, but the
idea remained alright, so I managed
to get this top performing ad.
Struggling to get those in three
months, like maybe 50 different
creatives in three months.
What if I could build a tool
where I could actually provide
these 50 creatives in three days
and this is how random was born.
so we build a tool that help any
consumer app to get creator ads for their
social media, ad campaigns, creators,
authentic creators, and now AI creators.
Maybe we'll talk about that.
Super seamlessly, uh, so that
they can actually scale on
social media, on TikTok and meta.
Jacob: Yeah.
That's, that's amazing.
And, and I think a lot of great ideas
are often come out of solving your own
problem, solving your own frustrations.
Right.
And, and that's really the best
way to understand that it is a
problem that you are challenged.
Uh, uh, you were challenged with it
and, and try to solve that for yourself.
It makes so much sense and clearly, uh,
um, a need for many other people as well.
Uh, uh, you know, ba based
on the success so far.
Um, so how about we
start with some basics?
Um, I'm not a paid at ads expert,
so maybe this is common knowledge.
But I thought, uh, you know, I was
reading some of the materials you,
you were sharing, and I thought the
different formats of UGC content
was quite interesting, where you had
classic UGC trends, carousels, uh, and
it makes sense that you'd wanna test
a, a variety of styles and formats.
Maybe you could explain these
different types of UGC, uh, for me.
Xavier: Yeah, of course.
So basically what we call UGC is like a
content we call, call it also creator ads.
So it's a content that looks
natural and that you will use
on TikTok ads on meta ads.
YouTube ads whenever, wherever.
And that doesn't really look like an ad
because now the more it looks like an
ad, the more likely people will skip it.
So it needs to look natural.
So this is what we call UGC.
And you have different kinds of UGC
and we provide that different kinds
of UGC, what we call classic UGC.
This is actually what most people think
about, like for example, a creator
that talks directly to camera show the.
The app, maybe a before and after
focusing on one key benefit or feature.
Uh, a quick, uh, like a quick
walkthrough, a very simple story
that illustrates a pain points.
This is a classic QGC style, right?
you also have, for example, trends.
Um, if you go to TikTok and on
meta now, uh, trends like is
um, is a sound, a mam a joke?
the idea.
Is to wrap as a map, your value
proposition inside that format, right?
So it's usually between five
to 10 seconds, 15 second max.
And like the beauty of it is that
the platform, uh, like TikTok, they
already know this format very well.
It, they know that it performs well.
So actually it can
actually make your app win.
Okay?
Make your ad win.
Uh, and we see actually a lot of top
performing trends for subscription
utility app, for example, The
product is not very sexy, but the
format is, so it can be super useful
to use trends you have carousel.
So Caral is a very, um,
uh, unknown in a way.
Features, especially for
TikTok, even for Meta.
It's like, for example, for
dating ads we see is where we
see the most top performing.
Um, ads are from Carousel,
so it's a slideshow carousel.
Um, it's like.
great to tell a mini story
in a very interactive way.
So we'll see like maybe five different
pictures with, um, with, um, with
a creator illustrating a story.
And then the user can, can, um, can slide
into each of the, uh, pictures And, and
the good thing is like when creators do
a carousel, they will talk about their
transformation, about their stories.
If, if you guys as app developers
wants to do a carousel, you'll
talk about your features.
And people, they don't really
care about your features.
When creators do that, they
talk about their transformation,
but their experience, and this
is what people care about.
So yeah, this is a.
Different kind of formats you
can actually use and we can
help you doing that at random.
Uh, providing you advice on which
format you could privilege to search,
to get traction on social media
Jacob: Yeah.
And the trends is interesting where you're
kind of, uh, riding a wave, right, of
existing things that are already popular.
And so those are probably, um, pretty,
probably pretty short-lived where
they're not gonna have a, a long lifespan
in terms of the, the, the content.
Xavier: Yeah, I mean, any
UGC now has a very short
Jacob: Okay.
Xavier: um, um, this is a good thing
for us as a business, but, uh, the,
the, the thing is actually usually.
When you have a winning ad in TikTok
on meta, it, it doesn't, it, it
doesn't last for more than a month,
maybe two months if you're very lucky.
So the, the, the lifespan is
between two or three weeks.
When you have a top performing ad,
of course trends tend to be even
Jacob: Yeah.
Xavier: But, but, uh, but you
need actually to refuel and
to get new UGC every month.
If you want to keep winning and
keep getting, uh, low CPI or high
S, you don't have any choice.
Jacob: And so you mentioned on
TikTok that they're actually seeing.
The trends and maybe help
boosting your, your post because
it's tapping into that trend?
Or is that just that, uh, naturally
people are searching for that trend?
And so, uh, does this TikTok have some
inherent thing where they're identifying
trends and showing those to users?
Or is it more just general
algorithmic, you know, suggestions?
Xavier: It's both actually.
If you have a trend that is posted
organically, then the algorithm from
TikTok or even Meta will boost it
naturally because they know that
this format is actually popular.
If you use it as an ad, it might be
also, uh, more efficient because people
are used to see that kind of trend.
So they'll be curious.
Maybe they spend more time,
they're not, they're not gonna
leave after one mil is gone.
They, they might stick a little
bit more and then you have higher
chance to, to people clicking on it.
Downloading your app so it can
drive to better performances.
Jacob: Do you think that, do you
typically see that one style ends
up working best for a client?
Do they, people usually have a mix.
Does it depend on maybe the creator
too, where like the creator and
the, uh, UGC style and the brand or
product all need to have alignment?
How do you think about like the different
styles and how they, how they mix?
Xavier: Right.
I it's always a, a matter
of diversity and volume.
Okay.
Volume always win.
So you should, you should, you
should never stick to one specific
angle, one specific kind of creator.
The idea is some, I, I mean, some
briefs works better on TikTok than meta.
Obviously, industries are
naturally more trend friendly.
If you go to social media or dating
it, it, it might be good, some
utility, some are more demo driven.
Some brands need more storytelling,
some other, more social
proof, et cetera, et cetera.
But the bottom line is
you need to test a lot.
You need to test different formats.
You need to test different creators.
You need to test different angles.
And usually what we see the pattern of,
uh, winning clients is clients that accept
the idea of volume and experimentation
than, for example, searching just one
perfect format or one perfect persona.
it's, it's always better,
for example, to have.
50 different creators that doesn't
really match exactly the type of
persona you're looking for, rather
than five personas that match
exactly what you're looking for.
So yeah, experiment,
experiment, experiment.
Of course, be aware of what would
be successful for your industry for,
for your, for the platform you are
actually, uh, um, advertising on.
And we can help with that.
But experimentation is key.
Jacob: Do you, um.
Probably look for the, the client and,
and to provide some guidance on what
they see work, give you some information
so you can better align kind of the
briefs you're providing, uh, uh, to the
creators based on what they see working.
Or do you, I guess, how, how
much of a helping hand, uh,
uh, do you give to the clients?
Or is it that clients, you know,
they, they need to come with more,
uh, understanding of, of what kind
of the value of their product is?
Xavier: Yeah, so basically when a client
comes to us, what we do is we have
what we call the creative strategist
that will actually go deep into their
product, see what's our value proposition,
what makes them different from
competitor, and then the idea is first.
Basic check what competitor did all
the top performing for competitors.
So we, we can get all the
different, uh, concepts from
competitor, from competitors.
also know ourselves, uh, what are
the current trends that are winning?
What are the current
concepts that are winning?
Because we are doing that every day.
We are, we have like 10,000,
uh, creatives per months.
So know what are the current trends.
Um.
Then you, not you, you.
So we use AI obviously, to source all the
winning ads from competitor, et cetera.
But at the end of the day, you need the
human brain from a creative strategy
that come up with five perfect briefs.
of course on your platform, if you wanna
advertise on TikTok, it's gonna be very D.
It's gonna be slightly different
than if you advertise on meta.
If you advertise on both, actually
you would need some concept that match
both platform, but based on that.
You need to come up with maybe five
different briefs and having different
video and um, for each of these briefs.
Jacob: Got it.
Got it.
Xavier: have some trends.
You can have some classic
UGC, some, um, uh, yeah.
Jacob: Yeah.
And so it sounds like generally
the UGC creative process, um.
This follows like a similar format to
other ad creative processes where you
start with a variety of options, hone
in on on which ones are performing, find
your kind of champions, and then try
to iterate and, and double down on, on
kind of your, your creative champions.
Is that right?
Xavier: A hundred percent
is exactly the way it works.
You start very broad.
So you will, you will test many
different concepts with many different
creators, with many different formats.
And then, um.
You will have enough volume, and then
you will get sign signal you know, uh,
maybe some briefs that won't work at all.
So we'll forget about it, and then
you'll see some briefs that will be
more successful, and then you will
iterate on this brief and then, and
then double down on some briefs and dumb
on the, on some kind of creators, some
kind of tones, et cetera, et cetera.
So yeah, start broad, get seniors
and then, uh, and, and then, um, dig,
dig into some, uh, winning formats.
Jacob: How many creators does
someone need to start with to kind
of narrow down and kind of have a
better chance of finding a winner?
Xavier: Um, so for a serious test
in a country, so if you wanna start
a test in the US for example, um,
what we usually recommend is starting
with around 20 creators, right?
So you can start between, yeah,
between 20 and 40 creators.
come up with four or five different
concepts and then you, it's because
I will give you ation to find early
winners and see some first patterns.
Jacob: Got it.
And so you, you have, it's based on
the geo where, based on like kind
of the country that you say, okay,
well if focus on the US or focus on
Western Europe, so you need different
creators for different regions.
Xavier: Yeah, so usually our clients are
usually, uh, starting with the US market
Jacob: Yeah.
Xavier: they expand to, for Europe,
for example, or Latin America.
So the way it works usually is you
start with the us, you get some top
performing in the us and then if
you want to expand to, uh, Germany
or France, you can actually use,
uh, top performing from the us.
It's in, it works, it's counterintuitive,
but it can work and you, you
keep the same, um, you keep the
same, uh, winning for format
and then you use local creators.
Or you can use AI having, for
example, a top performing, uh,
creatives with the US creators and
then just duplicate it with, uh,
Jacob: Interesting
Xavier: Using AI or to
have a brand new persona?
You can do
Jacob: is that, um.
How, what is the percentage chance
that the creative that performs in the
US will also perform in other geos?
Is that pretty high or is it, is it a mix?
Xavier: actually.
Usually what we see is when you
have a top performing in the US it's
very likely that it's going to be a
top performing, uh, anywhere else,
Jacob: Okay.
Xavier: likely.
Very, very likely.
So this is why we always recommend our
clients to start, I mean, unless they're
not targeting the US market, but it's
very rare, um, to, to focus, focus on the
US market first, get some winners, and
then once you have that you can, you can,
um, you can expand to other countries.
Jacob: And, and so really you need the
new creative because you're putting
in a different language where it's,
it's, you need the French person for
the French market to feel natural.
Xavier: Yes and
Jacob: Okay.
Xavier: Um, why?
Yes and no, because actually
that's actually counterintuitive.
we, I mean, if you have a top
performing creatives in, um, in
English, in US English, it can actually.
B, a top performing for
non-US, non-English speaking
countries, French and Germany.
I mean, it
Jacob: Yeah.
Xavier: it doesn't work as
much as in the US but it works.
So the, so you can use actually
the original ad for better result.
What we recommend is getting also
this same creatives with local
credit, with French credit if
you're targeting the French market.
But my, uh, my advice would be do both
Jacob: Yeah.
Xavier: test.
English speaking ads, but also use
local creators or use AI to get, um,
to get, um, creative in local language.
Jacob: Do you think of the, let's say,
uh, English, English UGC in the French
market and French UGC in the French market
as like different placements where you
have, you know, audience network and you
have Facebook feed, where potentially the
different creators reach different people,
and so you have like different reach?
Or is it just that sometimes
it works, sometimes it doesn't.
Xavier: Yeah, no, it's a good question.
Especially with meta.
They, they, you know, they
launched, uh, under meta recently,
and, um, the diversity of key.
So, for example, if you launch an a
campaign, um, so let's say you, you
launch in 20 different creatives.
If you get 20 different creators by,
let's say, two or three creators,
or maybe you are only one creator,
meta or even TikTok, they will
consider that it's almost the same ad.
Okay?
So they will not reach,
reach out to new audiences.
it's.
super important to have
a diversity of creators.
The more diverse your creators,
uh, base is the more likely
you reach out to new customers.
So, so it's very, very important to
have many, many different creators and
this is how we can help because we.
Usually provide one creative for one
creator, and we have like 50 K active
creator including 20 5K in the us.
So you can get many different,
uh, many different faces, uh, to
promote your app super seamlessly.
Also, with ai, now you can define
any personnel you would like.
That makes it even easier to, um,
to, um, to multiply the incarnation
that will, uh, talk about your app.
Jacob: Got it.
Got it.
That makes sense.
That makes sense.
Uh, so, so we start with 20 to 40, how
many do you usually end up with, uh,
you know, based on that performance?
What do you, what do you usually see?
Xavier: So just first a number.
You, you need to know
that it's completely okay.
Uh, like, and, and an ad that
doesn't perform is pretty normal.
So usually the success rate is
around maybe 5%, something like that.
So, so.
Basically, as I told you earlier, the more
you scale, the more creative you need.
Plus you will need the ad
fatigue, so you will end up with
hundreds and hundreds of ads.
I mean, our number one clients now,
uh, today, they, we are delivering
more than 1000 creatives per month
Jacob: Wow.
Xavier: you an idea.
Jacob: Yeah.
What, what, like, you don't
have to give the exact, but like
what spend level do you need to
support that level of creatives?
Xavier: Uh, yeah, on their side, are, um,
they are, they are, um, seven figures per
Jacob: Right, right.
Xavier: in terms of media spend.
Um, so they're, they're big spender
on, uh, both TikTok and meta ads.
but, uh, but even with, uh, five or
six figures, you need, uh, minimum
20 to 50 creatives per month.
Jacob: Yeah.
Yeah.
Just because of the
creative fatigue is so fast.
Uh, yeah.
Yeah.
And not all of it performs.
Xavier: Yeah, basically a game of numbers,
Jacob: Yeah.
Xavier: so you have rules to make sure
that your ads is gonna be successful.
First you need to have content
that looks authentic, right?
Um, second, it needs to be qualitative,
um, and to have a good script because
if you don't have a good script, usually
creator will be lazy and it won't work.
But still with that, even if, with
the good creators, with, uh, authentic
content, with good scripts, um.
The, you, you still never know
which one is gonna be successful.
Me, I still have no clue when I see an
ad if it's going to be successful or not.
I mean, no one can predict that.
It's very similar to the,
music industry actually.
Uh, it's very, it's very impo.
It's impossible to predict which one
is gonna be a hit, single, but maybe
5% of them or even less, will take all
the money, capture all the budgets.
And the other one will only be here
to actually provide signals, you know?
Uh,
Jacob: Yeah.
Xavier: so yeah.
Fascinating industry.
Jacob: Yeah, well in the music
industry there's some, some ways
to kind of engineer hits where
I can't, uh, bribe Facebook to
play my, uh, to play my ad more.
But yeah, I understand what you're saying.
I, I understand what you're saying.
Um, so, so, so on that note, in
terms of like styles and formats.
Like once you find a style
that once you find a style that
performs, do you, um, do you tell
creators, okay, this is what works.
Go create more of this?
Because there's also some that you,
you know, it has to be authentic or, or
so do you have to still give creators
kind of the freedom to try new things?
How do you think about that?
Uh, you know, that balance between,
you know, authenticity, but you also
know this style works, so you want
them to create something like that.
Xavier: Very good question.
So first very important when you provide a
brief, usually not, not being too strict.
Uh, if you don't know which, uh, if you
don't know the winning, uh, formats,
uh, when you do a brief, not being too
strict and not being too loose, because
if you're too loose, the clutter will be
lazy it won't work if you're too strict.
Uh, you don't leave room for what we
call happy accidents and creativity
and usually top performing come from
a happy accident and creativity.
Now, let's say you have a
winning format, alright?
You have, you have a new winning concept.
You absolutely, need do what
we call replicate campaign.
And you have actually two ways to do it.
You can.
You can, for example, give like
pretty broad concept with app
specific guidelines and then you let
K interpret them in their own style.
Um, or you can give very specific concept
base, uh, specific concepts and ask K to
replicate a winning angle loss structure.
Alright.
some of our best performing
campaigns are replicate campaigns.
Jacob: Hmm.
Xavier: You know, so they take a wind
structure and they ask creators to
recreate in their own words, but they
need to recreate in their own world.
So they shouldn't actually,
uh, uh, replicate word by word.
The backbone would stay similar, same
hook, same narratives, maybe same
call to action, but the faces energy
or the newest nuances will change.
And usually this is how you get
top performing with real creators.
Now, if you wanna.
If, if you wanna have the exact
same structure, the exact same
word, you also can, but my advice
would be use AI to do that.
Jacob: Hmm.
Xavier: if you ask a creator to
stick to a very, very specific, uh,
script, usually it's not natural.
Usually it won't work.
Jacob: Yeah.
Yeah.
So it sounds like early on.
You need to give maybe more, uh, uh,
broad guidance and less structure
to figure out the winning ad
concepts kind of guide, help this,
uh, the creative process happen.
Have some happy accidents happen
that to learn, okay, oh wow, that
that creator did something unique
there and that really worked.
They worded something a certain way.
And so early on it's about kind of more
exploration of figuring out what concept.
And then once you figured out the concept.
Give a little more structure, but
still like it has to feel authentic.
Yeah.
Xavier: exactly.
Absolutely.
Hundred percent.
Jacob: Yeah.
And, and, and so you can talk about
different styles a little bit.
Uh, do you, what, what do you see
the differences or nuances between
different ad platforms and, and
how kind of things styles change
or performance differs by style?
Xavier: Yeah, so first the main
playgrounds for us are TikTok and
Jacob: Yeah.
Xavier: of course,
working with YouTube Snap.
But really the, the two, two main
playgrounds are TikTok and Meta.
And, and even if we are talking
about what we call the TikTok
ization of social media.
So if you look at, for example,
materials, it's very similar to TikTok.
So it, it, it tends to be sort of
similar, but, um, but they behave,
they still behave a bit differently.
For example, on TikTok, things
over perform very often.
First, very short video, like I
I told you earlier about trends,
but around ten second no optimal.
CTA very organic, very low, just
maybe a strong relatable hook.
Uh, good watch rate.
Very important, a trending sun, because
sun is very important on TikTok, can get
a huge reach while looking super simple.
Also, I told you earlier on TikTok,
many advertiser ignore it, but carousel
Jacob: Yeah.
Xavier: can perform extremely well,
and especially when the caption
and slides tells the personal soul.
So this is TikTok on meta.
I mean, you still have carousel and you
start having some top performing with
trends, but the winning pattern is usually
clearer, more structured narratives.
Uh, product demo, the sweet spot
would be a bit longer, like, uh,
maybe between 15 and 30 seconds.
Maybe where the creators work around.
Some use case show the product,
reinforce the value in a very direct way.
so yeah, I would say that's just,
that's, those are the, the main
differences between TikTok and meta.
But you can have still some top
performing ads on TikTok that performs
on meta and vice versa, obviously.
Jacob: Yeah.
Xavier: of course, all
platforms the same will apply.
The recipe for success is
volume, volume, volume.
You need to test a lot of creators,
a lot of concepts, a lot of video.
And then the data will, will tell you
which one, uh, is gonna be successful
and which one you should replicate.
Uh.
Jacob: So on, on volume.
Um, volume is hard if you
don't have enough budget.
Do you see that there's a.
A starting point, like on a minimum
spend, maybe not just with, not, maybe
not specific to rom dom, but for,
for, for an advertiser to see success
with success with UGC generally.
Um, do, do you say, Hey, like this is
probably not the first thing you test.
This is a, you know, the second or
third once you've, once you've reached
a little scale, or how do you think
about when it's the right time?
Xavier: Yeah, so.
Let's say you are a subscription app.
You're very big using, uh, I dunno,
Google for example, you want to, you
want to go on social media, right?
Okay.
I, I, I wanna scale on social media.
So first my advice would be focus on one,
Jacob: Mm-hmm.
Xavier: only TikTok or only
Jacob: Okay.
Xavier: Don't try to do both.
When you want start, try to
focus on only one platform.
and if you wanna focus
on meta, for example.
My advice would be, as I told
you, start with a minimum of
20 creatives from 20 creators.
Just focus obviously on one country.
example, the US market, and the
minimum spend for one month would
be, uh, I say between 20 and 50 K
Jacob: Yeah.
Xavier: for the first months.
Below that you would not have
enough data, enough senior to see
if you actually can scale using
this, um, using this, uh, network.
Jacob: Yeah, I mean it's probably similar
to just other types of advertising too,
where I think the, the minimum spends you
need to learn what works on, on meta ha ha
has gone up over the years, uh, because of
ad costs, but also because of signal loss.
And so that's probably aligns
pretty closely with just like.
How you should think about
budgets in general for meta?
Is that
Xavier: Hundred percent.
A
Jacob: Yeah.
Okay.
Okay.
Okay, cool.
That makes sense.
That makes sense.
Um, so I, I, I guess I'm, I'm curious
how, how we can connect, uh, add
learnings, add performance to the
actual product or, or the funnel.
So maybe if you, if you look at.
Like, let's say some best performing for
specifically subscription apps for your
best performing subscription app clients.
Um, what are the common elements
maybe in kind of the funnel either
in, uh, you know, web to app funnel
or in-app onboarding funnel that,
you know, make creator ads work?
Are there differences that, uh, happen in
the product if were for the subscription
apps that you see that UGC works and
performs versus those who doesn't?
Xavier: Yeah, I mean.
The, uh, I mean, the first thing
is if you have an app that already
makes money on TikTok and meta,
Jacob: Yeah,
Xavier: you c usually scales.
It's as simple
Jacob: with organic content.
Xavier: Yeah.
Jacob: Okay.
Xavier: You see, uh, yeah,
usually c usually scales, right?
So, so the, the, the thing
is, it's, it's pretty simple.
The, the winner keeps usually
one story from ad to paywall.
Okay.
And then they can deliver the
promise, quick, win, and fast.
It, it's, I mean, it,
it must be super simple.
Um.
So, so for example, what winners do,
usually they have one narrative end
to end for, for example, if the ad
promise is like sleep better in seven
minutes, it should be like the same
thing in the landing, in the store
page, in the onboarding, in the ad.
And you need to repeat the promise
and show the same hero feature
immediately at any touch points.
Usually what we see.
Um, where UGC underperforms
usually, is when the product
needs longer technical demos.
Jacob: Uh.
Xavier: you think, uh, like heavy
utilities like PDF, heavy Utilities
app or some games also, some of them
you cannot crack it using social
media and they often scale better.
For example, on Google,
on uploading with deeper.
Jacob: Gotcha.
And, and I imagine there's some level of,
um, emotional connection in the product.
If you have an emotional connection
in the product, that probably performs
better because, you know, this UGC
content is coming from a person and, and
inherently there's some emotional hook
that kind of gets people, people in.
Um, and I, I guess
that's, that's generally.
And I'm thinking about it aligns with
what works on, on social media in
general, like you're saying, where,
you know, you're, you're, if you think
about the behavior, someone is, uh,
scrolling, they're not, they don't have
high intent of trying to do something.
And so you need to have a product
that's, uh, can be valuable.
Um.
You know, for, for most people where
it can be interesting and they can
substitute it where it's entertainment
almost, there's some substitute for
scrolling Facebook, for looking at
YouTube, for looking at TikTok, where
it can fill that time verse like you're
saying, a product that needs higher
intent, where someone is searching for
a solution, they want some, a solution
to solve that, that uh, uh, problem.
Um, yeah.
Does that sound right?
Xavier: Yeah, completely right?
Jacob: Yeah.
Xavier: I, I completely
agree with you on that.
Jacob: Yeah.
A a and so do, uh, do you see
teams learn from the UGC content?
Like based on what works where you, you
know, you were saying that, you know, you
test a bunch of creative, you might have
some creative breakthrough, some happy
accident where you see some hook works.
Do, do you ever see teams take
learnings from those and then
bring them into the product?
Xavier: Yes, sometimes it's difficult
because the UA team are very
different from the product team,
but we can see some success stories.
So, we did this on a
campaign, for example.
It was, uh, they, they like,
they, they like have a generic
unlocked premium kind of stuff.
Um, uh.
And, and then they change.
This swaps, this generic and log
premium for the exact creator
promise we had in the ad.
Jacob: Hmm.
Xavier: Um, they just kept one rating
line and the clean seven day trial,
nothing fancy, but actually trial and
service moved thanks to this change
from, uh, from, from was like a super top
performing a that took all the budget.
Jacob: Yeah.
Xavier: you have some examples
where, where actually products
and tech team inspiration from a
winning ads and it actually works.
It's completely consistent
from add to payments.
Jacob: I, I, uh, I've, I've seen
that, I've seen it go both ways
where, um, not with UGC, but other
types of ads where, um, I think I was
working for a meditation app and they
saw that, uh, anxiety focused ads.
Uh, performed best and we tried adding
more anxiety copy to the onboarding flow
and it just, people didn't care where
there's something about like, I guess
sometimes it works, sometimes it doesn't.
There's something about f.
Once people are in the product,
they kind of forget everything else.
You see an ad and it's kind of just like,
it's like a subject line in the email.
Once you open the email, you don't
remember what the subject line was.
Now you're focused on this, and so
like there's some kind of disconnect
where as long as you get someone into
the product, doesn't really matter
how you get them there, but maybe
if you can take those learnings.
It's, it's a test.
It's not a guarantee though.
Xavier: Absolutely.
Absolutely.
You, you can have ads that are actually
very good for CPI, so people will
actually be, uh, attracted to click on
the ad, eventually download the app.
But you don't have the right promise.
Jacob: Hmm.
Xavier: have, maybe the features
that you have pushed in the, um,
in the ad is not the right one.
And then you have a very,
very low, actually, um.
Uh, install to trial are
Jacob: Right.
Xavier: installed to a sub rate, so
it's also very important, even if you
forget about the subject, it's important
that the subjects, uh, and the ads
match with what you wanna sell what you
wanna actually, um, because if you are,
for example, selling only the three
features in the ad, then you might be
very disappointed by the performance
of your ad campaign in terms of a os.
Jacob: Yeah.
Xavier: also very important to make sure.
That you measure actually the,
um, that the UA team measure the,
um, the a OS and not only the CPI?
Jacob: It is tricky to do that though now
with if a bunch of different creatives
and you're trying a bunch of different
concepts, uh, to, to really hone in on,
you know, the ones that drive trials.
I guess you have to structure your ad
campaigns correctly and, and make sure.
Xavier: so yeah.
Usually the, the way most of our
client, um, uh, our clients do it
is, um, they have what they call.
ad campaigns where they
Jacob: Hmm.
Xavier: measure CPI and then this is
where they actually get quick winner.
Okay?
So for example, you put all your
creatives in it, uh, and then
you will have some quick winner.
You will, um, and all these quick
winner, they go into your BAU
or your, uh, scale campaigns.
And on this scale campaigns, you
already have winner from the test.
And the, and the original winners.
And then you only measure here, um, for
example, uh, uh, subscription rate or a
os, and this is where all the money goes.
So you have winners because
it's easier to measure quickly
the CPI, all the CPIs winner.
Then you put in this, uh, ad campaigns,
and then you scale, you scale, you
scale, and you, um, and you can
grow, uh, using that technique.
I mean, most of all our clients
are using that technique.
Jacob: Yeah, because that's what I was
gonna ask about, about the feedback
loop, because you have to, if you
have to wait till day 30 retention
or, you know, trial conversion rate,
the feedback loop is much slower.
But if, I guess that makes sense, you're,
you know, uh, uh, measuring on top level
performance metrics, click rate install.
Okay.
This has strong signal, it
could be a winner and then put
into your business as usual.
And then because it's getting most of
the money, you can clearly tell if you
see, uh, you know, trial conversion rate
tank, you go, okay, well that's not good.
That's that, that one,
that one's not performing.
Yeah.
Okay.
Okay.
That makes sense.
That makes sense.
Yeah.
Yeah.
Okay.
Okay.
Um, cool.
Uh, so we, we, we've touched
upon it a little bit, uh, uh,
previously, but we can't escape it.
We've gotta talk about ai, uh, you know,
everybody, uh, is talking about it.
And so, you know, you, you've
mentioned, uh, replicating, uh,
top performing creatives with ai.
Um, I, I also, um, I
think I saw that you, you.
Your training models based on the
AI videos, uh, to see which perform
best and worse, and using that to
predict, um, which brief or, or
style, uh, and which creator are
likely to win for specific clients.
I guess, can you touch upon that first
about how that, uh, how that process
works of, um, those, the learning
signals and then trying to tailor, uh,
uh, styles and formats to, to creators?
Xavier: Yeah, sure.
So, as I told you earlier, if
you wanna be successful, it's
super important to get volume,
Jacob: Yeah.
Xavier: of creative.
But of course, you, you need, you need.
content at scale.
Okay?
It's a q and q quality and quantity.
So you can, it's getting a
lot of volume of, uh, crappy
content is completely useless.
And so to be able to scale,
this quality, you need ai.
Why?
So the way it works actually for us
is whenever you launch a campaign,
for example, like, uh, a nap that will
help you sleeping better, for example,
the first thing is we will analyze.
that are likely to succeed, and we will
match those briefs with thousands and
thousands of creators in the us for
example, thousands of creators, and we
will calculate a score on the brief.
The creators.
So we look for example, at the
creators or all the past campaigns
they did, the performance, they,
the, the, the, the performance they
had on their past campaigns, their
social metrics, their bio, their
hashtags, all their video, et cetera.
And then based on that, they will
be able to, um, to actually rate.
Each of these creators and only push
the campaign to K likely to succeed.
And only the AI can do that because
no human would be able to analyze
thousands and thousands of K each
of them having hundreds and hundreds
of, uh, past collaborations.
Alright, so the way we see it is having
an, an AI algorithm, uh, AI algorithm
that would match any campaigns,
any brief with the top creators.
And then what we also do If you, if
you want to have like maybe 20 or
maybe a hundred creatives per month.
The AI also very good to analyze
each video before it's sent to the
clients and make sure that the video
is actually aligned with the brief
so no one wastes their time with
a video that didn't respect it.
The save zone that didn't
show the, uh, app in action.
They didn't put the right screen cast.
The AI is very good actually at analyzing
everything, so AI for us is just.
A way to scale, to scale quality, uh, and
then increase considerably the chances of
success of our client for real creators.
Okay?
Jacob: Yeah, yeah,
Xavier: Uh, this is how we
use AI for real creators,
Jacob: This.
Xavier: are also using AI and
maybe we want to, you want to
talk about that, to generate
content, and that's another story.
Jacob: Yeah.
Yeah.
I, I do wanna ask about that.
I, I, I have one follow up question, kind
of the briefs and kind of what you built.
Are you using, um, what are you
using, uh, existing tools out there?
Uh, have you built your
own in-house tools?
Uh, to, for that AI layer?
Xavier: So you're talking about the
Jacob: The, the briefs, the,
the matching, the video reviews?
Yeah.
Xavier: Yeah, so, so, yeah, sorry.
So about that it's all internal.
So we have trained a model actually, where
we actually, um, each briefs, we match it
with creators and then we see the output.
I mean, has this creator been
successful for this campaign?
Yes or no?
And then it help us training
and improving the model.
And has, we have now thousands and
thousands of campaign, Almost hundreds
of thousands of existing collaborations.
Past collaborations.
We are starting to have enough data
to train a model and making sure that
we can clearly increase the chance of
success, um, of each collaboration.
Jacob: Do you, do you see, can you,
can you measure that the, the chance of
success, the success rate is going up?
Xavier: Yeah,
Jacob: Yeah.
Xavier: Uh, we can see for each brief,
we can see the, the, um, the success
rate for each brief, for each creator,
for each video, and this actually help
us to train and improve the model.
Jacob: Yeah.
Got it.
Got it.
Um, that makes sense.
I guess one, one thing I, I was curious
about is do you, um, do you incentivize
the creators based on performance?
Xavier: No, I mean, yes and no.
So basically they all have a flat fee,
Jacob: Okay.
Xavier: uh, because at the end of the day.
never know which one is
gonna be successful, okay?
You can increase your chances of
success by having the right creators,
by having the right brief, uh, by having
the videos that match with the brief.
But you never know if
it's gonna be successful.
All you can do is increasing
your chance of success.
Okay?
So there is still, um, a part of
luck, uh, and, and then a creator can
be successful or not successful and
not being fully responsible of it.
So.
They have a flat fee, if they
perform, what we're gonna do is for
the next collaboration, they will
have a flat fee that will increase.
Jacob: Got it.
Got it.
So there's some level of performance
bonus because you want them
to continue to work with you.
You want them to stick around.
Xavier: yeah,
Jacob: Okay.
Yeah.
Yeah.
Okay.
Interesting, interesting.
Okay, so let's, let's talk
about the ai, AI creators.
Um, do you, what, what do you
see the future of this is?
Is this AI generated video gonna
replace real people, creators?
Xavier: No, I think it's gonna be,
uh, it's never gonna replace, um.
I think it's gonna be a
Jacob: I, yeah.
Xavier: Um, actually we have started
working on AI generated video.
We have a huge demand
on AI generated video.
Um, I think they are incredibly useful,
uh, and we, and we see actually great
performance in very specific use case.
for example, if you need to
reach audiences that are very
hard to source in real life.
So for
Jacob: Hmm.
Xavier: looking for very high income
persona, which are usually very
difficult to, uh, target for, um, UGC.
If you're looking for specific life stage,
think of a pregnant woman, for example.
It's very, very difficult to to, to
get, or if you want to produce very
some complex scene very quickly,
like street interviews, that kind of
stuff, or very cinematic concepts,
then AI can be incredibly useful.
But what AI can do for now is replace what
we call, what we call a accident of real
Jacob: Right.
Xavier: So when a real person
actually tests your app,
connects it to their own life.
Come up with a line out with
that nobody in the marketing
team will have ever written.
is where actually the magic come from.
So, for example, I have this,
our number one, uh, ad ever.
350 million views in the us.
Uh, it was sponsored, but still
it was a huge top performing ad.
And actually the creator, they
like pronounced the brand name.
He called the app called Swipe
Wipe with a funny accent.
She said, swap, swap, something like that.
It was a bit funny and, and and, and
then actually the brand, they were Okay.
I, I'm not sure I want to keep that.
they ending, they ended up keeping
it it's like, just because of that
boom, 350 million view, you know,
Jacob: Yeah.
Xavier: AI could predict that.
No AI could actually do that.
They would have pronounced the name
perfectly, the brand name perfectly.
yeah, this is a perfect illustration that
need AI sometimes for very specific use
case, but you still need real creators.
Jacob: And so maybe there, it's,
it's, it sounds like it's valuable
for maybe scaling once you've found
a creative concept that works because
you know what works and then you can,
you can crank out variations and, and,
and new, new iterations, duration.
Xavier: Yeah, a hundred percent.
let's say we have a winning, um.
Winning hat for a garage like a plumber.
And then you wanna extend to many
different jobs, like, uh, like for
a lawyer, any, we have this case
because we are working with, um,
AI concierge assistant, an app,
and they need many different jobs.
And they found like a winning hat
for one specific job with a real,
and then we want to expand to.
All kind of jobs.
And then AI could be absolutely
incredible doing that.
Um, so finding the winning concept with
human and then scale and expand with AI
can be super useful or maybe vice versa.
What you can do is testing
some first concept with ai.
Jacob: Hmm.
Xavier: You see some first
traction with some initial concepts
because you want to test it.
And then once you have a, a winning
concept, you can ask for a real human
uh, do that concept and maybe find
some top performing, uh, from there.
Jacob: Is AI video priced, uh, lower?
Xavier: Yeah, it's probably slower.
So on our side, the AI video
starts at, $30, $35, the the ai.
But on, on our side, it's not only,
only a prompted video because we have AI
operators that will produce, uh, videos
that is completely, uh, fully edited.
Jacob: Got it.
Got it.
So there's still, there's still manual
intervention because you need to add
text or add overlays or, or whatever,
whatever it is to actually make it, uh,
the right kind of concept for an ad.
Xavier: Yeah, especially now because
you have many models around, so you
obviously have, so two from OpenAI.
You obviously have VO 3.1
from Google, but you
have, uh, many others.
You have ENTs from TikTok and actually.
of the, uh, models are very
useful for some specific scripts.
Jacob: Hmm.
Xavier: don't have like one model
that is good for all the different,
um, uh, all the different use cases.
for example, if you need
replica, you might need actually,
uh, um, uh, one, two, 0.2.
You know, it's, it's another model.
So you have different models for each
video, you would need a lot of different
prompts, uh, video generation for one
prompt, because one prompt can lead
to something that is, uh, that is not
Jacob: Yeah.
Xavier: Each prompt, by the way.
Cost money because, for
example, 10 cent per second.
So 32nd multiply by 20, uh, twice.
So you, the, the, the price are
increasing, um, the price increasing fast.
And once you've got uh, video from
points, video generations from
points you a hundred percent of
the time you need post editing.
You need to add subtitles, you
need to, to do some clean, um.
Some clean shot from one to another.
So, so the bottom line is you need
someone that actually is a specialist
to do that kind of video if you don't
wanna waste your time and your money.
Jacob: Yeah.
Yeah.
Xavier: And what we have, what we call
AI operator, that has automatically,
uh, attributed to your brief whenever
you launch a brief and in the
hours they will provide with, uh,
some, um, some ads that, and videos
that perfectly, uh, feature brief.
Jacob: This is probably, uh,
the beginning of like what we
see as a new job title, right?
Yeah.
Xavier: Actually you have a
name, we call them AI operator.
Jacob: Yeah.
Yeah.
So, so I, I would imagine the more this
will expand to other companies, it's kind
of like, you know, there's, early on it
was like prompt engineer 'cause people
were talking about similar Yeah, yeah.
Similar, uh,
Xavier: it's not only
prompting generics more complex
Jacob: yes.
Xavier: sometimes we
need, uh, lip syncing.
For the audio.
You will need, uh, post
editing, et cetera.
So we call them, yeah, video AI operator.
Jacob: Yeah.
Yeah, that, that makes sense.
Um, and I think, you know.
One interesting thing, kind of, uh,
one through line is that when, when
we're talking about all this is that,
you know, fundamentally, you know,
marketing is, is still very similar
at the end of the day where, you know,
humans haven't really changed, right?
We're still, you know, uh, for,
from, for hundreds or thousands of
years, our motivations, desires,
psychology are still the same.
Uh, and so I'm curious, you know, we
talk about, you know, the emotional
hooks and, and you know, uh, uh.
Getting people interested and these
same concepts hold true and that there's
something special about human creators
actually having that happy accident.
You know, I, from, from your perspective,
you know, what, what are these
foundations of marketing that that
still hold true that people need to
remember when working with AI creatives?
Xavier: I mean, it's, I
I think it's very simple.
I mean, people have for, for years,
thousands of years people have a
problem or a desire they want to
understand quickly how something,
a product, a service, will change
something and solve their problem.
So if you look at the ad, they don't
care if the ad is very polished,
they just care if it feel real,
relevant, and easy to understand.
Jacob: Yeah.
Xavier: they have two things that
matter actually, the clarity.
It's, you need to show very
clearly how your app solve one
specific problem, not 10, just one.
The clearer the problem is, the easier
is for the viewer to project themselves.
number two is authenticity.
Jacob: Mm-hmm.
Xavier: Uh, the ad really should look,
uh, and feel like the content you see
between two ads, you know, symptom,
same type of person, same environment.
If it screams ad people will,
uh, people will scroll basically.
So, so yeah, just basically
in a very authentic way.
Explain a product can solve the problem.
As easy as that.
Jacob: Yeah, I think.
We often forget.
That's what, you know,
that's what your business is.
You're solving problems for people.
Uh, you know, don't forget that's, that's
the whole, the whole core of what what
you're doing is, is solve a problem.
And, and usually just
one problem is enough.
Uh, don't, don't overcomplicate things.
Focus, focus, focus.
Yeah.
Yeah.
It, it, it makes sense.
Um, I'm curious, you know, I like
asking, you know, this type of question,
but like, what's one counterintuitive
thing you've learned over the last
few years helping companies run?
You know, UGC creative, something
that's uh, uh, you know,
you surprised you, I guess.
Xavier: Yeah, so I told you earlier,
but uh, the, the thing that actually
surprised me a lot is you have a lot of US
English ads that can perform outside the
Jacob: Yeah.
Xavier: even in US speaking in non us.
I would never have bet on that,
but it's, I mean, it's a case.
Um, so that's, that's the first one.
another content cont intuitive is, I,
I'd say, is that actually your intuition,
uh, intuition about what will work is.
Often wrong.
Jacob: Mm-hmm.
Mm.
Xavier: a you are a hundred
percent sure will, uh, win
will most likely underperform.
And the one that feels a bit weird
can end up changing your business,
you know, and changing your life.
You know, so I told you about the
swipe right, you know, the swipe link
that the client wanted to reject the
video, uh, because, uh, the name of
the brand is pronounced correctly.
It's 350 million view.
Alright.
And it's completely changed,
obviously, the trajectory of the app.
Jacob: Yeah,
Xavier: So, so yeah.
Jacob: I think that's, you know, on
that point, that's something people
should remember that literally.
One top performing ad can
change your whole business.
Where, where it's, it's
not, uh, it's not linear.
There's not slow progress,
but like, you know,
Xavier: it's,
Jacob: orders of magnitude difference.
Can, can, yeah.
Xavier: Zero to one, zero or one.
Uh, this is why it's very
similar to the music industry.
Jacob: Yeah, yeah.
Yeah.
Xavier: thing, you know.
It's, uh, it's uh, whereas it takes
all the money, all the attention,
uh, or, uh, it doesn't exist.
Jacob: Yeah.
Yeah.
It's, it's, it's so fascinating
where, you know, I, I've worked most
of my career on kind of the, the
post acquisition monetization side.
Uh, you know, once people actually
install the app, how do you
get them to pay and opt funnel
optimization and stuff like that.
But just the, the, um.
The, the, the world of like ad champions
and how performance differs so much.
And it's, I, I think, you know, of course
it depends on the algorithmic e elements,
but, but it's, it's human emotion, right?
Can you, can you tap into something that
resonates, that connects with people?
And it is, it's so interesting.
Yeah.
Yeah.
Um, yeah.
Okay.
So, so we, we've talked a little
bit about rom dom throughout, but
do you wanna give a quick plug for,
for rom dom and, and maybe, yeah.
Xavier: Yeah, we talk about
Runda, but very simple.
We have, we help subscription
apps and consumer apps, scale
creator ads on TikTok and Meta.
Jacob: Yeah.
Xavier: we work with
leaders in the category.
We work with Tinder, we work with
Photo Room, we work with Flow.
So most of the leader in the category
and our sweet spot is apps that
already invest in page social.
And that one scale, uh, without
turning their internal team in
that into a content factory query.
So we are kind of a creative wingman
and, um, they usually understand the
value of creative volume our clients,
but they lack time and bandwidth.
Jacob: Hmm.
Xavier: So what we're gonna
do is handling everything from
brief to creator AI production.
We are gonna deliver ready to run assets.
With two obsessions num number
one, obsession providing you
with this top performing ad.
This is what we do provide
top performing ad obsession.
Number two, we, we are
not gonna waste your time.
Jacob: Yeah, and, and so there's lots
of companies doing similar things.
How are you guys different?
Xavier: So we are not an agency in a sense
are we are like most of our company are,
um, for company employees are engineers.
everything is streamlined.
So we have basically a tool that will.
Automatize and streamline everything.
So as I told you, um, whenever you
launch a brief, um, the creators are
automatically, um, are, are gonna
automatically being ran and automatically
being, uh, pushed, uh, um, a notification
so they can automatically apply.
So we will select them.
The AI will make sure that the video is
automatically aligned with the brief.
So thanks to all the tech we have
put and all the AI can deliver.
Spend time, 15 time more than any
competitor with a better quality.
Jacob: Got it.
Got it.
Well, um, yeah, I, I, I,
I've heard great things.
I, I know some people that work with
you all and, and, and, you know,
heard, heard very positive things.
So, um, this was awesome.
I, I really appreciate you
sharing all this knowledge.
Uh, any, any last things
you wanna promote, plug any
closing notes to, to share?
Xavier: No, just if you can, you can
reach out to me on LinkedIn, on mail.
I dunno if you have the contact.
And if you say that you
come from the botsy podcast,
I'll do a personalized demo.
Jacob: Okay.
Xavier: I'll, I, I'll take
time to make sure that I, uh,
prepare the right brief for you.
So just bring me an email, at Ramdat am
and then I'll make sure to treat you well.
Jacob: won't be an AI
Xavier, it'll be a reels.
It'll be
Xavier: this.
Jacob: okay.
I love it.
I love it.
Okay.
Well, we'll, we'll, we'll link, um, we'll
link to your LinkedIn in the show notes.
We'll link to, to rom dom in the
show notes so people can go find it.
And I, I think, um, yeah, you
may have one or two case studies.
We, we can also share
there what we'll see.
Uh, but this was awesome.
I, I really appreciate you joining.
Xavier: Thank you Jacob.
Jacob: All right.
Bye.
Thanks.
Xavier: Bye-bye.
Thanks for listening.
Hope you enjoyed.
Please go to price power podcast.com
to see all the episodes.
Go to Spotify and YouTube and
give us a subscribe and follow
so you don't miss any episodes.
Alright, talk to you next time.
All.