January 30, 2025 • 53 mins
Discover how Jonathan Moffie, a pioneer in the CTV ad tech universe, is revolutionizing television advertising for small businesses with his company, Streamr. Join us as we unpack his journey from launching some of the first fast channels at Yahoo to leading video ventures at Shutterstock, and how these experiences culminated in the creation of a game-changing service that promises to put your business on TV for a mere $250. Jonathan’s insights into the world of connected TV (CTV) advertising reveal the challenges and triumphs involved in making this space more accessible and less intimidating, especially for small business owners who might find the jargon overwhelming.
Listen to Jonathan as he shares the remarkable story behind crafting a user-friendly ads manager, simplified into just 10 steps, setting it apart from the tedious processes of competitors. Learn how Streamr tackles the common obstacle of businesses lacking video ads by offering a cost-effective solution that makes TV advertising as easy as pie—perfect for your local pizza shop or neighborhood café. This conversation shines a light on the transformative power of innovation in CTV advertising and its potential to level the playing field for businesses of all sizes.
Stay tuned for more in-depth insights on video technology, trends, and practical applications. Subscribe to Voices of Video: Inside the Tech for exclusive, hands-on knowledge from the experts. For more resources, visit Voices of Video.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:07):
Voices of Video.
Voices of Video.
Voices of Video.
Voices of Video.
Speaker 2 (00:18):
Well, good morning everyone and welcome to this new
edition, next edition of Voicesof Video.
Hey, I'm Mark Donegan and I amhere with Jonathan Moffey.
Jonathan, welcome to Voices ofVideo, mark thanks for having me
.
Yeah.
(00:38):
So you know, we're going tojump in and have a great
conversation about CTVadvertising, which you know.
I don't think there's anybodywho's working in video today who
doesn't care about that.
Right Agreed.
Speaker 3 (00:52):
For sure yeah absolutely so.
Speaker 2 (00:58):
I know you built the company Streamr, but before you
know, I let you talk aboutStreamr, what you guys are doing
, and then we have aconversation around um, you know
, I let you talk about streamer,what you guys are doing, and
then we have a conversationaround uh, you know what's
happening in the world of ctvadvertising.
I think you brought a videoright to tee up this whole thing
.
Speaker 3 (01:17):
Let's check it out, our first promo video oh, that's
awesome.
Speaker 2 (01:20):
yeah, it's like one minute long, so let's roll the
video.
Speaker 1 (01:25):
Business owner add on TV for free Well, not free
cheap.
Already interested?
Visit streamerai If you needmore info.
Keep watching $250, two minutesyour ad on TV.
You can show up to our websitewith nothing at all and we'll
create an ad for your businessin less than a minute with
nothing at all and we'll createan ad for your business in less
(01:47):
than a minute.
Let's summarize visitstreamerai, upload or create ad
with one click.
Pay 250.
Your ads on tv.
Your future customers arewatching.
What do you want them to see?
Speaker 2 (02:02):
wow, 250 bucks and I can put a business on connected
television.
Speaker 3 (02:09):
Yep.
Speaker 2 (02:10):
That's crazy.
Well, why don't we start here?
I always like to give ourguests an opportunity to tell
the origin story.
I think it's really importanthow did you get here?
So why don't you tell us a bitabout Streamer AI, the genesis
of the company and yourbackground, and then we'll dive
(02:33):
in?
Speaker 3 (02:34):
Great, great, yeah.
So I've been in the CTV ad techbusiness for over 12 years at
places like Yahoo Whirl Ilaunched one of the first fast
channels ever for people, peopleTV, yahoo Finance, et cetera.
I took a brief stint out of adtech to lead the video and music
(02:55):
businesses at Shutterstock andeven prior to that I started
actually my career as a producer.
So I've been in the video spaceon different levels for over 12
, 14 years.
So Streamr.
So my last startup I was at webuilt an ads manager.
The goal was to build theeasiest ads manager in CTV and I
(03:18):
do believe we did that.
We made it 10 steps toadvertise all the different
competitors out there like 25plus steps.
Even meta, which is like theHoly grail of ads managers for
SMBs, is like 21 steps.
So we made it 10 steps toadvertise.
We're like we're going to upendthe market and we start meeting
with business owners pizzashops, restaurants they all had
(03:41):
the same two problems.
Number one you'd be like, allright, upload your video ad and
they'd be like I don't have avideo ad.
Speaker 1 (03:49):
What video ad.
Speaker 3 (03:51):
Oh Jesus, all right, here we go, we'll give you an ad
credit.
We have the cheapest creativeservices team in the world.
For 250 bucks, we'll make you avideo ad.
Let's meet in a week, come backand launch this campaign.
And they'd go.
Well, I only wanted to spend250 bucks.
Like, okay, well, we'll giveyou an ad credit.
(04:12):
All right, here we go.
So we come back, we meet in aweek and we get them in the
campaign setup and we're like 10steps, this is going to be so
easy for them.
And they go what the hell is aDMA?
And I'm like, oh good point.
No one that's not in wonkyadvertising knows what a DMA is.
A pizza shop owner DMA?
No.
So we're like, geez, this isstill too complicated.
(04:36):
That was about last year when wesaid, wow, what if we use this
Gen AI craze to reduce thefriction of not having a
creative and build a video adand then use AI to actually do
the campaign setup and launch.
So it literally takes less thantwo minutes.
So basically, that's whenStreamr was born and we started
(04:58):
launching beta campaigns and ourkind of pinnacle case study was
this fitness influencer reachedout all excited.
He was able to generate an adand launch it on TV in 30
seconds while training acustomer at the gym, and we were
like, okay, we definitely builtthe easiest way to advertise.
However, just quickly to getinto where we are today, at the
(05:22):
same time, I had a DSP, a demandside platform, reach out.
I know the founder really well,ceo, and he said, hey, streamr
looks really cool, but we don'twant to push our advertisers to
you guys.
However, we do have a problemwhere, five to six times per
week, we have advertisers whowant to try CTV but don't have a
video ad.
Do you have an API?
(05:42):
And I was like, of course wehave an API.
We did not.
Speaker 1 (05:46):
We did not have an API at the?
Speaker 2 (05:48):
time, but our engineering team is fantastic.
That's a great startup.
Founder response.
Speaker 3 (05:52):
I love it.
Speaker 1 (05:53):
Yes, yes, yes and yes .
Speaker 2 (05:54):
Yeah yeah, let me send over to the docs, you know,
a couple of days.
We're just putting thefinishing touches on them.
Speaker 3 (06:07):
Of course, 100%.
So the good news is I have,like unicorn, 100x level, 10x
engineers.
I have 100X engineers Built itin a weekend, closed that deal
and ever since then we've beenoff to the races with what we
call Gen AI solutions, which isbasically arming broadcasters,
ad tech platforms and agencieswith Gen AI tech to democratize
CTV ad buying.
So I can talk for days.
That's that's it for me, rightthere.
What's going?
Speaker 2 (06:26):
on.
Yeah yeah, yeah, I love it.
I love the journey, the story,the opportunity.
I think it's fascinating.
A couple of things that you said, you know you, there's the
setup problem, right, and any ofus you know who are in
marketing or worked aroundmarketing, who've tried to set
(06:47):
up even just like a Facebook ad,you know, I mean, I'm talking
just a Facebook ad, you know,forget, you know, video on a
connected television, I mean it,it, it is hard, you know, it
really is, and it's and it'soverwhelming.
So so you had that hurdle, youknow, and so you brought what
(07:08):
was it?
You said like 21 steps, whichwas kind of the the best in
class meta, right, and youbrought it down to 10.
So it's like, oh, you know, wecut in half, more than half Um.
But then there's the contentproblem, and that really
resonates because video is just,it's a different beast, right,
(07:31):
you know, as much as we allcarry awesome cameras around and
maybe some of us are a littlemore comfortable than others,
you know, sort of flipping itaround and going on video or
whatever it's like.
Still, even that you know, likeyou don't create ads that way,
at least most of us don't.
So, wow, it's amazing.
(07:53):
Well, why don't we I think agreat place to you know launch
the discussion and also to givethe audience you know the next
level of detail behind the realproblem you're solving, and then
that'll be a great point for usto talk about some of the
technology you've developed andyou know.
(08:14):
And then what this can reallydo is you brought a couple
slides right, and so why don'twe bring those up?
And in case anybody's like, ohno, here we go, here's the pitch
.
I promise you, jonathan is notpitching.
We do not pitch on Voices ofVideo, promise you.
(08:34):
So hang in there, don't hang uphere.
But Jonathan put together somereally good talking point slides
that I thought, hey, these arethe points we're going to cover
anyway, so why not put them upon the screen?
So yeah, so tell us about theproblem around advertising 100%,
yeah.
Speaker 3 (08:53):
So just to double click on the extent of the
problem, a video ad on averagecan cost at least minimum
$10,000 to produce and takeweeks teams of creatives.
But you know, 10,000 is likethe minimum.
A Superbowl commercial, forgetabout it.
Millions of dollars.
Speaker 2 (09:10):
Can you give us an idea like that?
I mean, just for those of usthat don't really know, that
$10,000, is that kind of at thelevel of you know, the local
like roofing company that youknow throws together something
with a couple of guys up on theroof with hammers and you know,
is it that, or is it somethingthat looks a little more like,
(09:30):
you know, commercial televisionor what?
Speaker 3 (09:34):
Even at a local level , if you're working with an
agency and you know a videoproducer and editor again, I did
those jobs.
I was there.
Um, I I really uh lovecreatives and you know
they'reives and the bar isreally high, but yeah, even that
will cost thousands of dollars.
Speaker 2 (09:52):
Amazing.
Speaker 3 (09:53):
And then the other part that, once you have that
video ad, that may be evenharder is the campaign setup and
launch.
So when you get into thesetools like you said, even meta,
which billions of dollars flowthrough the pipes, tens of
billions of dollars and SMBs areusing it.
It's really also agencies usingit on behalf of a small
(10:15):
business owner.
Even that's complicated.
I've worked in the industry forover a decade and I used to run
Facebook, facebook campaignsand even that set up 21 steps,
super complicated the userexperience, lots of buttons and
widgets.
You're like I'm spending money,where am I going to go wrong?
So we simplify that bybasically using the AI to know
(10:39):
where your business is.
If it's a retail location,we'll automatically target the
zip codes around you.
We know the context of yourbusiness.
So, let's say it's a retaillocation, we'll automatically
target the zip codes around you.
We know the context of yourbusiness.
So, let's say it's a burgerjoint, we can extract metadata
(11:05):
from your website, from yourGoogle business profile, and
automatically target burgerenthusiasts, foodies, locals in
your area, fine diners, and addthat to the campaign targeting
automatically.
So really, you're justbasically putting in your credit
card info, picking your budgetand you're launching on premium
CTV.
Speaker 2 (11:19):
That's true, there we go.
Speaker 3 (11:22):
Oh, solution.
Yeah, so, like I said, thesolution we generate the
commercials and you can go tostreamerai right now.
We have some strategicpartnerships with a lot of the
different Gen AI companies, likeOpenAI.
We use ChatGPT for Omni.
On the back end we have 11 Labs.
They're a big AI unicorn audiocompany.
(11:44):
We use them as well.
We use suno, another.
These all these companies raisethese crazy 100 million, 200
million dollars.
Uh, you know, and what we do iswe call it like streamer gpt.
We're basically like the bestai dj remixing uh, ctv business
commercials and we make itextremely easy to generate the
(12:06):
commercial and launch.
So if you go to streamerairight now, put in your favorite
bar or restaurant, you couldgenerate and see it for yourself
.
We don't charge you to previewit, we don't even have you sign
up and maybe we should, but welike to make it really
frictionless, like cut to thebone.
Get all the friction out ofthere.
Speaker 2 (12:26):
Wow, yep.
Speaker 3 (12:46):
I think I advanced here.
Oh no, okay, there we go.
There are 400 million globalsmall businesses spending over
$700 billion in digital ad spend.
It's going to be about a $200billion streaming specific
market by 2030.
So there's just a ton ofopportunity.
And right now, tv on average wecould go to the next slide might
(13:08):
really drive this point home.
So on average, tvs had like 500advertisers.
Your big box, procter Gamble,l'oreal, et cetera dominate TV
ad spend and that was great fora really long time.
But we know linear is on amajor decline, whereas 9 million
(13:30):
plus advertisers advertise onsearch and social spending over
300 billion plus per year.
Our vision is to enablemillions of new businesses to
advertise on streaming by makingit simple and affordable.
Accelerate that by basicallypartnering with big broadcasters
(13:57):
, ad tech platforms and agencieswho already have a lot of ad
spend going through theirplatforms.
We're basically giving them GenAI tools to attract SMBs so we
can realize this vision fasterand democratize CTV ad buying.
We have a huge launch thisMonday.
My my Slack's probably goingcrazy right now while we're
talking about this really biglaunch of a new Gen AI powered
(14:17):
ads manager.
Speaker 2 (14:18):
So very exciting, yeah, yeah, the opportunity is
very clear to me anyway, and Ithink, a lot of people, so it's
great that you're going afterthis.
I am super fascinatedpersonally, but I think many
people in the industry areespecially on the creative side
(14:39):
around Gen AI use of, or Gen AItechnology, you know, to build
creatives, build assets.
So tell us you know how youbuilt this and what exactly
you're doing.
Like, you mentioned 11 labs, sodoes that mean that you're
cloning the voice of, forexample, the business owner and
(15:02):
that gives like a narrationtrack?
Or maybe you can go in a bitdeeper into exactly how you're
deploying those tools and whatthat output kind of looks?
Speaker 3 (15:14):
yeah, yeah, I usually at this point do a real-time
demo.
No, you're not supposed to justdo it live, but I love just
doing it live.
However, I'll definitely talkabout it in.
Uh that are watching.
Speaker 2 (15:26):
Go to streamerai go to your website, right, yeah?
Speaker 3 (15:29):
but, uh, basically all you do is search or enter
the business URL search for yourbusiness name or enter the
business URL, and then on thebackend, we'll generate a script
based on your website, yourGoogle business profile.
There's a few more places we'relooking to pull from in the
future, maybe like yourInstagram, because Instagram is
(15:50):
really like the storefront forSMBs these days, and also.
YouTube as well.
So we'll do that, generate thescript, and then we basically
built we call it a mood relevantmatching service on top of 11
labs and on Suno.
So let's say you're a yogastudio, we will automatically
(16:12):
write a script that makes sensefor a yoga studio, but then also
get a soothing voice from 11labs so that matches what you
would expect from a yogi, andthen also a soothing meditation
style soundtrack for Suno.
Suno is the AI music and bringthat all together generate the
(16:33):
commercial.
Um, and yeah, that's how it'sdone and so are you.
Speaker 2 (16:37):
Are you using stock?
Um, you know video and, uh, I Iguess perhaps some still images
, but mostly video.
Is that, then, what the actualcreative consists of?
And then you're overlaying themusic, the custom music from
Suno, as well as the narrationtrack.
(16:58):
Is that?
Are those the elements of thoseare definitely building blocks.
Speaker 3 (17:03):
We also you can click a button, we have edit with AI
so you can upload some of yourown images and frames of your
business.
I mean even video as well andyou can basically remix it
yourself if you're not happywith the out of the box ad
that's generated.
And stock assets.
I ran the video and musicbusinesses at Shutterstock.
(17:24):
We're definitely exploring thatarea.
Sometimes it's not perfect.
There's a good example.
There's this great cheeseburgerjoint here and I'm in Santa
Monica, California a HiHocheeseburger, and they have very
specific branding around theirWagyu beef burger and then if
you then try to use a stockburger being flipped, it just
(17:47):
doesn't sell the brand.
Speaker 2 (17:49):
So it's almost better to have an image that is
curated for that brand than a uh, you know, generic, uh burger
flipping yeah, that's um, youknow that's fascinating because
of course I, you know I'm amarketer and um, the the biggest
problem with using stock images, stock video, anything, because
(18:11):
sort of stock is not anymore itused to be.
Sometimes quality was a littleinconsistent.
Now like it all basically looksgood.
I mean it has to to be in thelibrary.
So, you know, it's like theimage is great.
Usually the model or, you know,whatever the subject is, you
know, is, you know, looks good,everything's good, but there's
just something in the framethat's out of sync with what the
(18:33):
message is.
You know, and and I always, youknow, for those that work with
me, work around me, you know, Ialways point out, you know, the.
The problem is is that there'sa segment of the audience that
won't even notice that, but ourbrains just pick up on those
things.
There's these super subtle cuesthat just don't seem right, and
(18:58):
so the problem is that, becauseoften I'll get feedback like,
oh, no one will really noticethat, or maybe we can sort of
crop it out a little bit andit's like, yeah, but you know
what, know what the brain is,the person viewing it it's not
even going to know why they feelit's not right, but there's
going to be, but they're goingto feel it's not right.
(19:18):
You know with.
You know whatever the words areon the screen or on the, you
know on the image or thesoundtrack going along with it,
the narration, you know um,whatever the context is.
So that's one of the challenges.
Speaker 3 (19:33):
A hundred percent.
Speaker 2 (19:35):
What about our?
Have you explored, you know,these image creation tools and
these video creation tools?
I imagine they're quiteexpensive and they also are
fairly hard to control.
Like, sometimes you get amazingoutput.
Sometimes you're like oh,what's that?
Eight fingers?
(19:55):
We've all seen plenty of those.
Like, everything's perfectexcept the person has three
fingers.
You're like, hmm, Yep.
Speaker 3 (20:04):
But yeah, I'll give you a.
Speaker 2 (20:05):
What are you doing there?
Speaker 3 (20:07):
I'll give you a little peek into our roadmap.
So we're going to release whatwe call Gen AI 2.0 before the
end of the year, which is goingto make our creatives a lot
better more templates, moretransitions, text on screen
that's more dynamic, really makethe videos pop, and still with
the same speed.
It takes two to three secondsto generate it, so that's going
(20:31):
to be, really great.
In the meantime, we're alwaysexperimenting and I talked to
the founder of pika labs, allthese companies.
They've raised like 100 million, uh, runway ml.
Oh yeah, talk to them becauseopen ai is a partner.
We've talked to them aboutgetting early sora beta access.
I think everyone wants thatSora.
Speaker 2 (20:51):
They were used.
Speaker 3 (20:52):
There was like a Toys R Us commercial that was
generated.
So the reality of all thosetools as you said I've talked to
a lot of there's these reallygreat Gen AI video editors that
are messing around with thisstuff.
You know Emmy award winningeditors.
I've gone to a few conferencesand they all say the same thing
(21:12):
there's no one today.
That's just like prompting thatToys R Us commercial.
Let's stick on that for asecond.
It's not just like build me theToys R Us commercial.
What happened?
is they are doing, prompting,thousands of prompts, pulling
the assets.
They're getting out of it intopost-production tools that
everyone knows and love.
I was a pro guy back in the day, you know Final Cut Avid what
(21:34):
have you?
Speaker 1 (21:34):
into.
Speaker 3 (21:35):
After Effects.
They're putting tons of workinto it, tons of labor, human
labor, and then they are alsobasically the GPU cost to train
those models and run it.
These are getting to the samelevels that we were talking
about before, whereas 10,000plus to produce.
So it's not there yet.
(21:56):
The Gen AI video from Runway,et cetera, is not there yet.
That being said, we're alwaystesting, we're always looking to
integrate.
Our kind of master plan andtalking to these folks is we
will continue to integrate thebest of Gen AI, large language
models and as they get better,we'll integrate with their APIs
(22:17):
and our creative will get better.
But our goal is always to makeit really fast and frictionless
and then showcase the best ofGen AI on streaming TV.
So they're not doing the adtech piece, but we're clearly in
the ad tech domain.
So that's kind of the long-termvision is once one of these is
faster, cheaper and morereliable, we'll be there to
(22:40):
integrate and showcase it.
Speaker 2 (22:41):
Yeah, amazing, yeah, amazing.
So let's pivot the conversationfrom creation, which I would
love to spend another hour justtalking about that, because I
think it's super interesting andit's very relevant to the
reasons that we specified.
But I'm not sure that everybodyis familiar.
(23:03):
You know, earlier you mentioneddemand side platform and you
know DSP.
I think anybody working invideo, at least if you're around
entertainment content that isadvertising as a component
probably has heard DSP and someof these other acronyms, which
(23:27):
I'll let you explain.
But why don't you give a two orthree minute primer on how the
ecosystem works, how it'sstructured, who the players are
and just how these platformseven work?
Because I think this also isvery important.
A lot of discussion today and Ilike that.
(23:48):
You came, you have experiencebuilding like Whirl, for example
, so there's a lot of talk onthe.
You know, in the streaming sidevendors and you know analysts,
observers, that you know fast isgoing to like be this huge boom
.
You know it's going to save theindustry, you know it's this
(24:09):
panacea and and yet you know, weknow that you know some
companies are, are doing welland they're, they're making a go
of it.
Others are, you know theeconomics aren't quite there yet
and everybody has their theoryas to why.
And half the time theconversations I'm in, I'm like
(24:30):
maybe we better explain how allthis actually works.
Yeah, because it's you know.
The implication is, you knowand I'm intensely not going to
name names, but you know,platform XYZ.
You know they just need tocharge more and then they're
going to be profitable, you knowor they and it's like you know,
(24:51):
let's look at how, let's lookat who all the players are and
how the money flows, and so giveus a primer on this, because I
think the audience it'd behelpful and I'm sure there's
somebody out there who's likeyeah, tell me exactly how this
works.
Speaker 3 (25:05):
Totally.
I just want to make it clearthat you asked me to explain ad
tech, which is a daunting task.
The most acronyms per squareinch in any industry.
Speaker 2 (25:18):
So, nate, you know there's the 40,000 foot view.
Why don't you do about the80,000 foot view?
Speaker 3 (25:26):
Well, I think the problem we're trying to solve
does actually help explain.
So we had this great piece thatwas written in a TV news check
that I contributed some quotesabout, and it talked about the
embarrassingly low fill rateproblem in CTV.
So what's really fascinatingright now is we have record
audience watching CTV andeveryone's getting in the ad
(25:49):
supported game, Even Netflix,who for years was like we're
never going to do advertising.
Speaker 2 (25:54):
Now they have advertising.
Speaker 3 (25:56):
There's rumors now Apple is doing it, they are
going to do it, everyone isdoing advertising, so that's
great.
That's one reason to be veryexcited when you're an ad tech
guy in the industry.
So the reality, though, isthere's so many ways my brain's
going.
With Linear, you got your moneyfrom advertising and from
(26:18):
retrans fees from the operatorswho kind of distributed your
content.
For now, at least, there's noclear path where those retrans
fees will carry over into CTV,so you're losing a major source
of revenue if you're likeComcast or NBCU, et cetera.
So we really need to rely onthe advertising ecosystem, the
(26:41):
CTV advertising ecosystem beingrobust.
However, the reality is that,on average, the programmatic
fill rate so programmatic is,you know, buying and selling via
kind of auction styleadvertising on Netflix, hulu, et
cetera, through a demand sideplatform, the biggest kind of
(27:02):
independent one being the tradedesk, the Walt, the Walt gardens
being Google and DV 360.
And there's some reallyinteresting antitrust stuff
going on there.
Google and DV360, and there'ssome really interesting
antitrust stuff going on there.
Speaker 2 (27:13):
Yes, there is.
Speaker 3 (27:15):
And then Meta as well , but Meta's not for CTV, meta's
social and mobile andprogrammatic's kind of the
growing way to transact.
There's also you can buy directto the platforms and usually
that's the bigger guys will getthese big direct spends.
They have big sales teams etcetera.
(27:37):
So we're trying to figure outwith broadcaster partners how do
we help them solve for thatsuper low 38% average fill rate.
And one way to do that is toattract millions of SMB
advertisers into the ecosystemthat are looking for
alternatives to just mobile andsearch, because now there's more
(27:58):
ways to target in CTV and we'redoing that by reducing the
friction with Gen AI and thenthose are direct spends on their
platform, which kind ofbypasses all of the problems
with the programmatic auction.
You can think of programmaticas the stock market.
It's like buying stocks andthere's lots of transactions,
(28:21):
billions of transactions and adrequests going on in real time
and there's just issues with thepipes sometimes and we're
trying to help with a solutionto solve that.
Speaker 2 (28:31):
Yeah, so a whole lot more to unpack.
But you made the reference andsort of skipped over and I'm not
sure everybody understands whatyou meant by low fill rates,
fill rates.
And then you haven't mentionedyet CPM't mentioned yet cpms.
And in the conversations I'm apart of, I find that that a
(28:56):
surprisingly high number offolks in the video streaming
world don't really have anunderstanding of even how cpm
relates to the profitabilitypotential of a fast service and
and and they're almost likethese decoupled things and so
it's like advertising.
(29:17):
Is this?
You know, 400 billion whateverthe number is I I don't know,
you know, but you know, fortraditional TV advertising.
And if we just capture this,you know some percentage of it,
we're all going to be.
And if we just capture this,you know some percentage of it
we're all going to begajillionaires.
You know, that's like the.
That's the tenor of theconversation, right?
Then you kind of go hold onhold on folks, cpms, let's do a
(29:41):
little math here.
You know like the numbers don'tpencil out.
So, with that set up, explain,go into more detail about bill
rates and then talk to us abouthow that's related and then and
then cpms and and then how allthis fits in to why, um um, you
(30:03):
know, traditional linear tv,that whole ecosystem is still
printing piles of cash and, andyou know, the fast services are
not, are not quite at theprofitability targets that you
know they certainly would liketo be.
Speaker 3 (30:19):
Yeah.
So the good news that I thinkeveryone watching probably does
watch TV, so they've likely beenannoyed by this analogy.
That will explain fill rate.
So you know, you have a certainamount of ad breaks per hour.
If you're Netflix and now youhave ads in your content and
usually there's like two minutesish per break, and you have
(30:42):
maybe like one, you know, fourads, 30 seconds or 15 seconds,
and the fill rate problem iswhen, instead of showing an ad
during those breaks, what youlikely see you don't see dead
air, you don't just see black.
You'll likely see what we callan ad slate, which is something
like we will be right back, wewill be right back.
(31:03):
Everyone hates we will be rightback.
So we're trying to get fewer.
We will be right backs in place.
Increase the fill rate to putmore businesses there that are
hopefully relevant to you.
You're searching for a new pairof sneakers on Instagram and
then we could actually retargetyou with that same thing on CTV.
(31:26):
Now there's a lot more digitalaudience targeting capabilities
in CTV, so that's the goal is toincrease those fill rates,
which increases thebroadcaster's revenue and
hopefully makes it actually abetter user experience instead
of seeing either the same adfrequency.
Speaker 2 (31:44):
Well, that's, I think , the problem that at least that
I can relate to.
It's like OK, I like the JeepCherokee, but I'm not in the
market for a car right now and Isee the Jeep Cherokee ad on
every single.
You know like come on, guys,show me something different you
know and I'm just picking onJeep I you know that hasn't been
(32:05):
the most recent one, but for awhile it was 100%.
Speaker 3 (32:09):
So, yeah, that's the frequency problem of you know.
There's only so manyadvertisers buying, and so
they're just showing that samead over and over again, which
obviously is a horrible userexperience.
Speaker 2 (32:21):
It, it is.
It is, yeah, exactly so,programmatic.
Um, you know, I think everybodyhas certainly heard the term
and understands kind of theconcept.
Like you said, it's sort of thestock market.
You know, there's this auctionsystem, buying and selling.
Um, maybe you could explain alittle bit more about what the
(32:41):
components are like.
Who are the players in a in aprogrammatic transaction?
Um, yeah, you know, maybe evenname a couple of the key
companies or players that areinvolved, because there's
multiple players involved inthis, you know, yeah, and then
and then, something that I'mcurious about is is the?
(33:06):
Is the ability to do thehyper-personalized?
Is the ability to do thehyper-personalized,
hyper-targeted, just like whatyou said?
So you geolocate and you say,hey, mark is in the Phoenix
metro area, so they're probablygoing to show something, phoenix
metro.
But how cool would it be if andsome people may not think this
(33:28):
is cool, but I actually thinkit's quite useful, you know if
not only was I shown someadvertisement by a business in
the Phoenix metro area, butsomebody who was in the same zip
code as me, you know so arestaurant that was in, you know
.
So it's within a.
You know, I don't know.
You know, mile, two mile, threemile radius, however big the
(33:49):
zip code is, I don't really know, but it's not that big, but
yeah, so explain first of allprogrammatic what does that
ecosystem look like?
And then what's required to dothis?
You know, hyper personalized,yeah.
Speaker 3 (34:04):
Yeah, I think it's actually I don't know if it's a
good thing or a bad thing thatI'm so in the weeds in
programmatic ad tech that I'mprobably going to keep assuming
too much.
Speaker 2 (34:15):
Well, I'll ask you to clarify, because I'm not in the
weeds All right, I'll give it ago.
Speaker 3 (34:21):
So let's take like a broadcaster.
I don't know why NBCU keepscoming up like Peacock or
something.
I don't know why NBCU keepscoming up like Peacock or
something.
And then they're going to workwith a lot of SSPs supply side
platforms, which haveconnections to a lot of supply
(34:41):
that will actually put their adrequests up for auction in the
programmatic marketplace.
And there's a lot of SSPs.
And then there's DSPs,demand-side platforms, who kind
of aggregate demand, and that'sthe different advertisers that
will go in and looking to buythat supply.
Speaker 2 (35:01):
So sorry to jump in, but just to clarify for me and
for the audience.
So an SSP would be a publisherin the traditional sense or a
streaming service.
Right, that would be an SSP.
Speaker 3 (35:14):
Well, an SSP basically aggregates.
So DSP and SP are both kind ofaggregators.
Speaker 2 (35:19):
They represent.
Sorry.
They're not the platform.
They're the front door to thoseplatforms, I guess.
Speaker 3 (35:25):
Yeah, ssps represent the publishers and aggregate up
the supply.
Speaker 2 (35:29):
That's what I meant, yeah.
Speaker 3 (35:30):
Yep, so like Pubmatic , magnite those are some of the
biggest ones and then DSPsrepresent and aggregate up the
demand.
So the trade desk is like thebiggest open market one, and
then they're kind of all settinga floor price for a CPM.
So let's say like a $25 averageCTV CPM.
(35:53):
And then you can, a DSP can bidon buying if it's Peacock I
don't want to hurt theirfeelings if I'm putting their
CPMs too low or what have you.
And then- Anyway, $2,500 forPeacock, yeah, they can buy
exactly they can buy uh the adrequests that peacock is selling
(36:13):
through maybe their ssp,pubmatic, magnite, etc.
Um working in the industry.
The reality is a lot ofbroadcasters, the publishers, um
end up thinking, okay, if Ijust keep adding more ssps that
can connect me to more demand,I'll increase my fill rates.
(36:34):
And unfortunately that's notalways the case uh.
So now some of these publishersare working with like 12 to 14
ssps and a lot of them are justlike reselling other uh
impressions.
So it's just taking a cut of acut of a cut.
What we're looking to do withour ads manager is give uh the
(36:55):
publisher a direct way for anadvertiser to be able to buy
their supply.
So it's outside of theprogrammatic auction, so it's
guaranteed to deliver Um andit's a more direct relationship
with the advertiser.
They can now email, market tothem, they can offer them
discounts, et cetera.
(37:17):
So that's kind of our way tohelp the ecosystem increase
revenue and fill rates.
Speaker 2 (37:23):
So you are primary, so your primary relationship is
with the DSP.
Speaker 3 (37:39):
I don't want to pigeonhole us as a demand-side
platform.
That's likely the best analogywhen you're buying through us in
Ads Manager.
We're Gen AI tech, we're, we'regen AI tech, but yes, we're.
We do basically aggregate upthe advertisers to buy through
us.
We work with ad servers then toand SSPs to then display the
(38:02):
ads on publishers yeah, yeah,okay, got it.
Speaker 2 (38:07):
And, and then the advertisers.
You were advertisers.
I jumped in so, as you'restarting to talk about the
advertising side, becausethere's a whole set of companies
and technologies that representthe advertisers right
aggregator.
Speaker 3 (38:27):
Obviously there are self-serve platforms like us who
work directly with theadvertiser, mainly SMBs, and
then there's also agencies thatwill like kind of aggregate up a
smaller pool of advertisers andhelp buy advertising on behalf
(38:47):
of those advertisers across,maybe, various demand side
platforms or do direct buys witha broadcaster.
So really direct buys are likethe Holy Grail.
That's what the upfront TVupfronts are.
Speaker 2 (39:02):
Exactly.
Speaker 3 (39:03):
And it's the big guys that take advantage there, sure
.
Speaker 2 (39:07):
Because the commitments have to be huge to
do a direct buy I mean I meantens of millions or even more.
I mean, what give us a scale?
Speaker 3 (39:17):
yeah, yeah, tens of millions.
Uh, amazon's now a new bigplayer with their prime video
offering and they had like oneof their first up fronts and had
, like you know, record success,but maybe you know vague with
the numbers, but great success,great, great great success Well
what does that mean?
Speaker 2 (39:34):
It was amazing.
Blew away our expectationsExactly.
Speaker 3 (39:38):
Exactly so those really big platforms or
broadcasters are still going tohave their big sales teams
getting those direct campaigns.
What we're really trying to dois give you know maybe those big
platforms will give them Gen AI, but also medium-sized
broadcasters, fast channels.
(39:58):
They can have their own adsmanager in self-serve way to
have direct spend on theirplatform.
So they're still going to plugin to the programmatic ecosystem
.
Just, the reality is at leastright now and there's companies
solving for this the fill ratesare like 38%.
It's not good.
It's not enough money to make aviable business.
(40:18):
So you need other ways to havea direct relationship with the
advertisers and get your fillrates up.
Speaker 2 (40:25):
Yeah, yeah, interesting, yeah, yeah,
interesting Boy.
I just I, I can.
I I'm always amazed at how manyhands are in the cookie jar
when it comes to advertising.
It's just so.
It's such a um, uh, complicatedecosystem and uh, yeah, but
(40:46):
it's amazing.
Well, uh, this has been areally wonderful conversation
and thank you for sharing whatyou've shared, and I really
can't wait to see the launch ofthe Gen AI tools and just the
evolution of the platform.
When we first met I thinkwithin five or 10 minutes of you
(41:10):
explaining what, what you weredoing, I remember saying we need
to have you on voices of video,and this is something that just
even personally, I'm I'm superinterested in and so really see
a lot of potential.
So best of luck, you know, toyou and the team.
We do have a couple questionsthat came in, so mind, mind
(41:33):
taking some questions yeah,let's do it I, I think, I, I
think these will be easy tomanage.
So, um, yeah, so it's one of theadvantages for those who are
watching live.
By the way, um, feel free totype into the chat of whichever
platform you're watching on.
I, I know we're streaming on Idon't know how many platforms
(41:53):
right now, but more than one,but go ahead and log in or type
into the chat and we'll try andget to those.
But okay, so, so here's a goodquestion.
So what's the most unusual usecase that can be covered with
Gen AI today?
So now, this is, I think, justwith all of your experience
(42:16):
building using Gen AI tools.
So I don't think this is reallyadvertising specific, but yeah,
I mean, in all of yourexperience as a creative.
I think you'd have a goodperspective on this.
Speaker 3 (42:30):
So yeah, so I actually have a recent case
study that is, I think, I think,really unusual and awesome.
So we have a law firm out herein California that we work with
that's advertised to ourplatform multiple times and one
of the partners texted me he'slike hey, crazy idea, there was
(42:51):
a injury at an Amazon warehouseout here in California.
Could you turn around like aworkers comp ad today that would
target potentially other peoplethat have been injured at
warehouses and blanket their zipcode so we could see if anyone
else will come forward, becausefor them that's worth hundreds
of thousands, if not millions,of dollars and, oh yeah, 250
(43:14):
minimum personal injuryattorneys.
Speaker 2 (43:16):
I, I'm clear with that business.
That's yeah, that's the side oflaw.
Speaker 3 (43:20):
They are printing money or they potentially can
yeah that's so we did, we did it, we did it wow and ran the
campaign.
Uh, turn it around in less than24 hours.
And he's like I'm going to hityou up every time news breaks
and run campaigns.
And I'm like, oh my gosh, youjust literally gave me a new.
You know, I'm supposed to bethe innovator with the great
(43:42):
product roadmap.
I'm like breaking news Gen AIcommercials Like that is.
You have to play to theadvantages and the speed and the
personalized audience targetingare the advantage Gen AI brings
, and if you would, have broughtthat to a creative shop that'd
be like great give us two weeks.
We'll get that back and he'slike two weeks.
(44:02):
It's in the news right now.
We want to show it to peoplethat are going to be like oh wow
, look at that Amazon thing, Ialso got hurt, maybe I should
come forward.
So it I'm like oh, this is awhole new roadmap thing.
I'm like, hey guys, we need tofigure this out.
So it was awesome.
Speaker 2 (44:18):
I, I, I, I love it and um, you know not to be, um,
you know we're, we're, we're,we're not advocating for.
You know, trying to profit off.
You know off.
You know terrible situationsbut a corollary that I recall I
remember hearing years ago and Ihad family a long time back in
(44:40):
Florida and you know, of course,you have.
You know, in the Midwest youhave.
You know tornadoes and you know, in Florida you have hurricanes
and you know tornadoes and youknow, in Florida you have
hurricanes and you know.
So there's weather and you knowone of these comes through and
you know, all of a suddenthere's hundreds of homes that
need new roofs.
There's, you know, there's allthe damage, right, and they
(45:01):
actually used to call theroofers storm chasers and they
were, you know.
Now, it usually was not a goodthing because usually the work
was a little shoddy and you know.
But they swoop into town becausethe insurance companies are
almost just standing therewriting checks yep you know, and
they would come in and you knowyou think like, wow, how
amazing, um, if you could get anad that's hyper relevant for
(45:25):
some you, you know, for some youknow news or something that's
really relevant at that at thatmoment and in that area you know
, um, wow, that is.
I mean, that's a whole newfrontier for advertising.
Speaker 3 (45:39):
So yeah, I thought it was.
Uh.
I love when, like, a customerblows your mind and put
something on your productroadmap.
Speaker 2 (45:46):
I'm like wow that's a cool product market fit.
Yeah, cool.
Well, there's another questionthat came in and I don't know if
you have worked in the Indiamarket, so if you haven't,
though, maybe you can stillcomment.
And the question is just how doyou see CTV advertising working
(46:07):
in like India?
But if you have experiencethere, answer it for India.
If not India, what's happeningin other markets around the
world, or other countries, etcetera?
Speaker 3 (46:21):
100%.
So that's another amazing GenAI specific use case that can be
solved.
So because we partner with 11labs, we have over 32 languages
and even more dialects that wetranslation and dialect.
That's right, so we actuallyhave run campaigns
internationally.
We did one uh, where we uhbasically automatically uh put
(46:44):
it in uh mandarin and indifferent.
It was a company, company inAsia that wanted to hit
different parts of Asia thatspoke different languages.
We're able to create a versionof those commercials with the
different languages anddifferent dialects.
So in India Hindi, tamil.
We can support it.
Speaker 2 (47:03):
Sure, wow, wow, that's cool.
Yeah, I can definitely see thelanguage translation as being
super powerful.
So talk to us then also aboutanother question that came in is
just and this one, I guess, iskind of general.
(47:24):
I feel like we kind of touchedon it, but maybe it's a good way
to wrap things up.
The question is I'd love tolearn about Gen AI revolution in
CTV, and I feel like this wholeinterview, the whole discussion
has been around really centeredon both the applications, the
(47:45):
tools, the use cases and reallythe value that Gen AI can bring.
But maybe where we'll end is,you know where is this going?
You know, and if I could don'tmake it so crystal ball, like
well, in 10 years, you know, butlike maybe the next, like three
years, like where is all thisgoing?
(48:06):
You know, in terms of Gen AItools, what you're working on,
advertising on CTV, just thewhole thing.
Speaker 3 (48:15):
I can't help but plug yet another streamer-specific
thing, and I wish.
Speaker 2 (48:20):
I had the video, if you see the video.
Speaker 3 (48:23):
I think it's a pretty amazing video.
So we have a product in beta.
We call it Streamer Gen AIInfinite amazing video.
So we have a product in beta,we call it Streamer Gen AI
Infinite.
And the idea here this, to me,is the biggest game changer that
Gen AI is going to have onadvertising.
So when I was at Whirl, theparent company app loving
multi-billion dollar mobileperformance demand side platform
(48:43):
, and they would have a big teamof creatives that would for
every campaign let's say,DoorDash is an advertiser they'd
create like eight to 10variations of the creative,
changing subtle things to try todrive an action, and they run
those eight to 10 creatives onsmall percentages of traffic and
whatever one performed the best, they'd shift the budget over
(49:06):
to.
So multivariate testing or A-Btesting.
Speaker 2 (49:09):
A-b Gen testing yeah, AB Gen AI.
Speaker 3 (49:10):
Yeah, gen AI completely upends this.
So we have a product in beta,like I said, gen AI Infinite
where we can take this so manysteps further, where there's an
infinite amount of variations wecan create, basically in real
time and not just testing, likedifferent colors of buttons.
It's literally differentscripts, different call to
(49:31):
actions, a funny ad, a dramaticad, a serious ad, and then
different voices, et cetera, toreally hone in on and
personalize the advertisementfor the user.
Speaker 2 (49:46):
Incredible.
Speaker 3 (49:47):
That's kind of one piece, and then I'll go the holy
grail.
So here's the user Incredible,that's kind of one piece.
And then I'll go the holy grail.
So here's the holy grail.
We're talking to companiesabout doing this and I think
this is going to happen in lessthan 10 years If you're able to,
as a broadcaster or publisherwhich they are pass us the
context of the show you'rewatching.
Say, you're watching House ofDragons on Max.
(50:09):
So you pass us the context inreal time through the bid
request and advertising.
We could then, at the next adbreak, automatically create,
let's say, a Nike commercialwith dragons in it that plays at
that commercial break.
Now that is crazy.
Speaker 2 (50:31):
That is crazy and I love that we're ending there
because, yeah, I just think, boy, you know there's.
You know we can live our lives,you know, with sort of this
general, you know with optimism,with you know a feeling
abundance, and we approacheverything that way.
Or we can live in fear, right,and I know, around AI in
(50:57):
particular, it seems to beeither one it's like a zero or
one.
It's either zero we're alltoast the world's over, you know
, like I'm gonna go buy a bunkersomewhere and never see the sun
or it's like this is just goingto unlock, you know, new
experiences, realities that youknow we could never even have
(51:19):
dreamed of.
And I'm very much in the lattercamp.
And you know, and a lot of it isjust one comment that I was
going to make early on is I'veseen many, many, many studies
and you probably even have a lotof the data to back this up
where people actually do nothate advertising.
(51:41):
What they hate is irrelevantadvertising.
They hate the repetitive,irrelevant advertising.
They hate the repetitive.
They hate showing a Jeep ad 18times when I'm not even in the
market for a car.
So it's not like, well, I amlooking for a car I'm not so
much interested in SUV, but it'slike I have no interest in a
(52:03):
car, but I've seen a Jeep ad 18times in the last two hours, or
know, or whatever, whatever thecorollary is.
And so with that, you know,it's just super exciting to me
that all of these tools aregoing to be able to make
messaging so much more relevantto me.
You know, based on what theyknow about my, my geolocation,
(52:27):
or, you know, using all the datathat's available, right, and
they're going to be able totarget, and then I'm going to
feel, wow, that actually wasvaluable and useful, and I'm
really glad that I saw that,because I didn't know that
company existed, or I'd heardabout them, but I forgot about
them, or whatever the situationis.
So, yeah, that's cool.
(52:49):
Well, let's end there, jonathan.
Thank you for coming on.
Voices of Video.
Thank you, mark, and we will.
I'm sure you know we'll betracking your progress and we'll
have to have you back as youlaunch, you know, the new
platform.
So Great, cool.
Speaker 3 (53:05):
Sounds like a plan.
Cheers, have a good one.
Yeah, you bet, great Cool.
Sounds like a plan.
Cheers, have a good one.
Speaker 1 (53:13):
Yeah, you bet.
This episode of Voices of Videois brought to you by NetInt
Technologies.
If you are looking forcutting-edge video encoding
solutions, check out NetInt'sproducts at netintcom.
Voices of Video.
Voices of Video.
Voices of Video.
Voices of Video.
Speaker 2 (00:18):
Well, good morning everyone and welcome to this new
edition, next edition of Voicesof Video.
Hey, I'm Mark Donegan and I amhere with Jonathan Moffey.
Jonathan, welcome to Voices ofVideo, mark thanks for having me
.
Yeah.
(00:38):
So you know, we're going tojump in and have a great
conversation about CTVadvertising, which you know.
I don't think there's anybodywho's working in video today who
doesn't care about that.
Right Agreed.
Speaker 3 (00:52):
For sure yeah absolutely so.
Speaker 2 (00:58):
I know you built the company Streamr, but before you
know, I let you talk aboutStreamr, what you guys are doing
, and then we have aconversation around um, you know
, I let you talk about streamer,what you guys are doing, and
then we have a conversationaround uh, you know what's
happening in the world of ctvadvertising.
I think you brought a videoright to tee up this whole thing
.
Speaker 3 (01:17):
Let's check it out, our first promo video oh, that's
awesome.
Speaker 2 (01:20):
yeah, it's like one minute long, so let's roll the
video.
Speaker 1 (01:25):
Business owner add on TV for free Well, not free
cheap.
Already interested?
Visit streamerai If you needmore info.
Keep watching $250, two minutesyour ad on TV.
You can show up to our websitewith nothing at all and we'll
create an ad for your businessin less than a minute with
nothing at all and we'll createan ad for your business in less
(01:47):
than a minute.
Let's summarize visitstreamerai, upload or create ad
with one click.
Pay 250.
Your ads on tv.
Your future customers arewatching.
What do you want them to see?
Speaker 2 (02:02):
wow, 250 bucks and I can put a business on connected
television.
Speaker 3 (02:09):
Yep.
Speaker 2 (02:10):
That's crazy.
Well, why don't we start here?
I always like to give ourguests an opportunity to tell
the origin story.
I think it's really importanthow did you get here?
So why don't you tell us a bitabout Streamer AI, the genesis
of the company and yourbackground, and then we'll dive
(02:33):
in?
Speaker 3 (02:34):
Great, great, yeah.
So I've been in the CTV ad techbusiness for over 12 years at
places like Yahoo Whirl Ilaunched one of the first fast
channels ever for people, peopleTV, yahoo Finance, et cetera.
I took a brief stint out of adtech to lead the video and music
(02:55):
businesses at Shutterstock andeven prior to that I started
actually my career as a producer.
So I've been in the video spaceon different levels for over 12
, 14 years.
So Streamr.
So my last startup I was at webuilt an ads manager.
The goal was to build theeasiest ads manager in CTV and I
(03:18):
do believe we did that.
We made it 10 steps toadvertise all the different
competitors out there like 25plus steps.
Even meta, which is like theHoly grail of ads managers for
SMBs, is like 21 steps.
So we made it 10 steps toadvertise.
We're like we're going to upendthe market and we start meeting
with business owners pizzashops, restaurants they all had
(03:41):
the same two problems.
Number one you'd be like, allright, upload your video ad and
they'd be like I don't have avideo ad.
Speaker 1 (03:49):
What video ad.
Speaker 3 (03:51):
Oh Jesus, all right, here we go, we'll give you an ad
credit.
We have the cheapest creativeservices team in the world.
For 250 bucks, we'll make you avideo ad.
Let's meet in a week, come backand launch this campaign.
And they'd go.
Well, I only wanted to spend250 bucks.
Like, okay, well, we'll giveyou an ad credit.
(04:12):
All right, here we go.
So we come back, we meet in aweek and we get them in the
campaign setup and we're like 10steps, this is going to be so
easy for them.
And they go what the hell is aDMA?
And I'm like, oh good point.
No one that's not in wonkyadvertising knows what a DMA is.
A pizza shop owner DMA?
No.
So we're like, geez, this isstill too complicated.
(04:36):
That was about last year when wesaid, wow, what if we use this
Gen AI craze to reduce thefriction of not having a
creative and build a video adand then use AI to actually do
the campaign setup and launch.
So it literally takes less thantwo minutes.
So basically, that's whenStreamr was born and we started
(04:58):
launching beta campaigns and ourkind of pinnacle case study was
this fitness influencer reachedout all excited.
He was able to generate an adand launch it on TV in 30
seconds while training acustomer at the gym, and we were
like, okay, we definitely builtthe easiest way to advertise.
However, just quickly to getinto where we are today, at the
(05:22):
same time, I had a DSP, a demandside platform, reach out.
I know the founder really well,ceo, and he said, hey, streamr
looks really cool, but we don'twant to push our advertisers to
you guys.
However, we do have a problemwhere, five to six times per
week, we have advertisers whowant to try CTV but don't have a
video ad.
Do you have an API?
(05:42):
And I was like, of course wehave an API.
We did not.
Speaker 1 (05:46):
We did not have an API at the?
Speaker 2 (05:48):
time, but our engineering team is fantastic.
That's a great startup.
Founder response.
Speaker 3 (05:52):
I love it.
Speaker 1 (05:53):
Yes, yes, yes and yes .
Speaker 2 (05:54):
Yeah yeah, let me send over to the docs, you know,
a couple of days.
We're just putting thefinishing touches on them.
Speaker 3 (06:07):
Of course, 100%.
So the good news is I have,like unicorn, 100x level, 10x
engineers.
I have 100X engineers Built itin a weekend, closed that deal
and ever since then we've beenoff to the races with what we
call Gen AI solutions, which isbasically arming broadcasters,
ad tech platforms and agencieswith Gen AI tech to democratize
CTV ad buying.
So I can talk for days.
That's that's it for me, rightthere.
What's going?
Speaker 2 (06:26):
on.
Yeah yeah, yeah, I love it.
I love the journey, the story,the opportunity.
I think it's fascinating.
A couple of things that you said, you know you, there's the
setup problem, right, and any ofus you know who are in
marketing or worked aroundmarketing, who've tried to set
(06:47):
up even just like a Facebook ad,you know, I mean, I'm talking
just a Facebook ad, you know,forget, you know, video on a
connected television, I mean it,it, it is hard, you know, it
really is, and it's and it'soverwhelming.
So so you had that hurdle, youknow, and so you brought what
(07:08):
was it?
You said like 21 steps, whichwas kind of the the best in
class meta, right, and youbrought it down to 10.
So it's like, oh, you know, wecut in half, more than half Um.
But then there's the contentproblem, and that really
resonates because video is just,it's a different beast, right,
(07:31):
you know, as much as we allcarry awesome cameras around and
maybe some of us are a littlemore comfortable than others,
you know, sort of flipping itaround and going on video or
whatever it's like.
Still, even that you know, likeyou don't create ads that way,
at least most of us don't.
So, wow, it's amazing.
(07:53):
Well, why don't we I think agreat place to you know launch
the discussion and also to givethe audience you know the next
level of detail behind the realproblem you're solving, and then
that'll be a great point for usto talk about some of the
technology you've developed andyou know.
(08:14):
And then what this can reallydo is you brought a couple
slides right, and so why don'twe bring those up?
And in case anybody's like, ohno, here we go, here's the pitch
.
I promise you, jonathan is notpitching.
We do not pitch on Voices ofVideo, promise you.
(08:34):
So hang in there, don't hang uphere.
But Jonathan put together somereally good talking point slides
that I thought, hey, these arethe points we're going to cover
anyway, so why not put them upon the screen?
So yeah, so tell us about theproblem around advertising 100%,
yeah.
Speaker 3 (08:53):
So just to double click on the extent of the
problem, a video ad on averagecan cost at least minimum
$10,000 to produce and takeweeks teams of creatives.
But you know, 10,000 is likethe minimum.
A Superbowl commercial, forgetabout it.
Millions of dollars.
Speaker 2 (09:10):
Can you give us an idea like that?
I mean, just for those of usthat don't really know, that
$10,000, is that kind of at thelevel of you know, the local
like roofing company that youknow throws together something
with a couple of guys up on theroof with hammers and you know,
is it that, or is it somethingthat looks a little more like,
(09:30):
you know, commercial televisionor what?
Speaker 3 (09:34):
Even at a local level , if you're working with an
agency and you know a videoproducer and editor again, I did
those jobs.
I was there.
Um, I I really uh lovecreatives and you know
they'reives and the bar isreally high, but yeah, even that
will cost thousands of dollars.
Speaker 2 (09:52):
Amazing.
Speaker 3 (09:53):
And then the other part that, once you have that
video ad, that may be evenharder is the campaign setup and
launch.
So when you get into thesetools like you said, even meta,
which billions of dollars flowthrough the pipes, tens of
billions of dollars and SMBs areusing it.
It's really also agencies usingit on behalf of a small
(10:15):
business owner.
Even that's complicated.
I've worked in the industry forover a decade and I used to run
Facebook, facebook campaignsand even that set up 21 steps,
super complicated the userexperience, lots of buttons and
widgets.
You're like I'm spending money,where am I going to go wrong?
So we simplify that bybasically using the AI to know
(10:39):
where your business is.
If it's a retail location,we'll automatically target the
zip codes around you.
We know the context of yourbusiness.
So, let's say it's a retaillocation, we'll automatically
target the zip codes around you.
We know the context of yourbusiness.
So, let's say it's a burgerjoint, we can extract metadata
(11:05):
from your website, from yourGoogle business profile, and
automatically target burgerenthusiasts, foodies, locals in
your area, fine diners, and addthat to the campaign targeting
automatically.
So really, you're justbasically putting in your credit
card info, picking your budgetand you're launching on premium
CTV.
Speaker 2 (11:19):
That's true, there we go.
Speaker 3 (11:22):
Oh, solution.
Yeah, so, like I said, thesolution we generate the
commercials and you can go tostreamerai right now.
We have some strategicpartnerships with a lot of the
different Gen AI companies, likeOpenAI.
We use ChatGPT for Omni.
On the back end we have 11 Labs.
They're a big AI unicorn audiocompany.
(11:44):
We use them as well.
We use suno, another.
These all these companies raisethese crazy 100 million, 200
million dollars.
Uh, you know, and what we do iswe call it like streamer gpt.
We're basically like the bestai dj remixing uh, ctv business
commercials and we make itextremely easy to generate the
(12:06):
commercial and launch.
So if you go to streamerairight now, put in your favorite
bar or restaurant, you couldgenerate and see it for yourself
.
We don't charge you to previewit, we don't even have you sign
up and maybe we should, but welike to make it really
frictionless, like cut to thebone.
Get all the friction out ofthere.
Speaker 2 (12:26):
Wow, yep.
Speaker 3 (12:46):
I think I advanced here.
Oh no, okay, there we go.
There are 400 million globalsmall businesses spending over
$700 billion in digital ad spend.
It's going to be about a $200billion streaming specific
market by 2030.
So there's just a ton ofopportunity.
And right now, tv on average wecould go to the next slide might
(13:08):
really drive this point home.
So on average, tvs had like 500advertisers.
Your big box, procter Gamble,l'oreal, et cetera dominate TV
ad spend and that was great fora really long time.
But we know linear is on amajor decline, whereas 9 million
(13:30):
plus advertisers advertise onsearch and social spending over
300 billion plus per year.
Our vision is to enablemillions of new businesses to
advertise on streaming by makingit simple and affordable.
Accelerate that by basicallypartnering with big broadcasters
(13:57):
, ad tech platforms and agencieswho already have a lot of ad
spend going through theirplatforms.
We're basically giving them GenAI tools to attract SMBs so we
can realize this vision fasterand democratize CTV ad buying.
We have a huge launch thisMonday.
My my Slack's probably goingcrazy right now while we're
talking about this really biglaunch of a new Gen AI powered
(14:17):
ads manager.
Speaker 2 (14:18):
So very exciting, yeah, yeah, the opportunity is
very clear to me anyway, and Ithink, a lot of people, so it's
great that you're going afterthis.
I am super fascinatedpersonally, but I think many
people in the industry areespecially on the creative side
(14:39):
around Gen AI use of, or Gen AItechnology, you know, to build
creatives, build assets.
So tell us you know how youbuilt this and what exactly
you're doing.
Like, you mentioned 11 labs, sodoes that mean that you're
cloning the voice of, forexample, the business owner and
(15:02):
that gives like a narrationtrack?
Or maybe you can go in a bitdeeper into exactly how you're
deploying those tools and whatthat output kind of looks?
Speaker 3 (15:14):
yeah, yeah, I usually at this point do a real-time
demo.
No, you're not supposed to justdo it live, but I love just
doing it live.
However, I'll definitely talkabout it in.
Uh that are watching.
Speaker 2 (15:26):
Go to streamerai go to your website, right, yeah?
Speaker 3 (15:29):
but, uh, basically all you do is search or enter
the business URL search for yourbusiness name or enter the
business URL, and then on thebackend, we'll generate a script
based on your website, yourGoogle business profile.
There's a few more places we'relooking to pull from in the
future, maybe like yourInstagram, because Instagram is
(15:50):
really like the storefront forSMBs these days, and also.
YouTube as well.
So we'll do that, generate thescript, and then we basically
built we call it a mood relevantmatching service on top of 11
labs and on Suno.
So let's say you're a yogastudio, we will automatically
(16:12):
write a script that makes sensefor a yoga studio, but then also
get a soothing voice from 11labs so that matches what you
would expect from a yogi, andthen also a soothing meditation
style soundtrack for Suno.
Suno is the AI music and bringthat all together generate the
(16:33):
commercial.
Um, and yeah, that's how it'sdone and so are you.
Speaker 2 (16:37):
Are you using stock?
Um, you know video and, uh, I Iguess perhaps some still images
, but mostly video.
Is that, then, what the actualcreative consists of?
And then you're overlaying themusic, the custom music from
Suno, as well as the narrationtrack.
(16:58):
Is that?
Are those the elements of thoseare definitely building blocks.
Speaker 3 (17:03):
We also you can click a button, we have edit with AI
so you can upload some of yourown images and frames of your
business.
I mean even video as well andyou can basically remix it
yourself if you're not happywith the out of the box ad
that's generated.
And stock assets.
I ran the video and musicbusinesses at Shutterstock.
(17:24):
We're definitely exploring thatarea.
Sometimes it's not perfect.
There's a good example.
There's this great cheeseburgerjoint here and I'm in Santa
Monica, California a HiHocheeseburger, and they have very
specific branding around theirWagyu beef burger and then if
you then try to use a stockburger being flipped, it just
(17:47):
doesn't sell the brand.
Speaker 2 (17:49):
So it's almost better to have an image that is
curated for that brand than a uh, you know, generic, uh burger
flipping yeah, that's um, youknow that's fascinating because
of course I, you know I'm amarketer and um, the the biggest
problem with using stock images, stock video, anything, because
(18:11):
sort of stock is not anymore itused to be.
Sometimes quality was a littleinconsistent.
Now like it all basically looksgood.
I mean it has to to be in thelibrary.
So, you know, it's like theimage is great.
Usually the model or, you know,whatever the subject is, you
know, is, you know, looks good,everything's good, but there's
just something in the framethat's out of sync with what the
(18:33):
message is.
You know, and and I always, youknow, for those that work with
me, work around me, you know, Ialways point out, you know, the.
The problem is is that there'sa segment of the audience that
won't even notice that, but ourbrains just pick up on those
things.
There's these super subtle cuesthat just don't seem right, and
(18:58):
so the problem is that, becauseoften I'll get feedback like,
oh, no one will really noticethat, or maybe we can sort of
crop it out a little bit andit's like, yeah, but you know
what, know what the brain is,the person viewing it it's not
even going to know why they feelit's not right, but there's
going to be, but they're goingto feel it's not right.
(19:18):
You know with.
You know whatever the words areon the screen or on the, you
know on the image or thesoundtrack going along with it,
the narration, you know um,whatever the context is.
So that's one of the challenges.
Speaker 3 (19:33):
A hundred percent.
Speaker 2 (19:35):
What about our?
Have you explored, you know,these image creation tools and
these video creation tools?
I imagine they're quiteexpensive and they also are
fairly hard to control.
Like, sometimes you get amazingoutput.
Sometimes you're like oh,what's that?
Eight fingers?
(19:55):
We've all seen plenty of those.
Like, everything's perfectexcept the person has three
fingers.
You're like, hmm, Yep.
Speaker 3 (20:04):
But yeah, I'll give you a.
Speaker 2 (20:05):
What are you doing there?
Speaker 3 (20:07):
I'll give you a little peek into our roadmap.
So we're going to release whatwe call Gen AI 2.0 before the
end of the year, which is goingto make our creatives a lot
better more templates, moretransitions, text on screen
that's more dynamic, really makethe videos pop, and still with
the same speed.
It takes two to three secondsto generate it, so that's going
(20:31):
to be, really great.
In the meantime, we're alwaysexperimenting and I talked to
the founder of pika labs, allthese companies.
They've raised like 100 million, uh, runway ml.
Oh yeah, talk to them becauseopen ai is a partner.
We've talked to them aboutgetting early sora beta access.
I think everyone wants thatSora.
Speaker 2 (20:51):
They were used.
Speaker 3 (20:52):
There was like a Toys R Us commercial that was
generated.
So the reality of all thosetools as you said I've talked to
a lot of there's these reallygreat Gen AI video editors that
are messing around with thisstuff.
You know Emmy award winningeditors.
I've gone to a few conferencesand they all say the same thing
(21:12):
there's no one today.
That's just like prompting thatToys R Us commercial.
Let's stick on that for asecond.
It's not just like build me theToys R Us commercial.
What happened?
is they are doing, prompting,thousands of prompts, pulling
the assets.
They're getting out of it intopost-production tools that
everyone knows and love.
I was a pro guy back in the day, you know Final Cut Avid what
(21:34):
have you?
Speaker 1 (21:34):
into.
Speaker 3 (21:35):
After Effects.
They're putting tons of workinto it, tons of labor, human
labor, and then they are alsobasically the GPU cost to train
those models and run it.
These are getting to the samelevels that we were talking
about before, whereas 10,000plus to produce.
So it's not there yet.
(21:56):
The Gen AI video from Runway,et cetera, is not there yet.
That being said, we're alwaystesting, we're always looking to
integrate.
Our kind of master plan andtalking to these folks is we
will continue to integrate thebest of Gen AI, large language
models and as they get better,we'll integrate with their APIs
(22:17):
and our creative will get better.
But our goal is always to makeit really fast and frictionless
and then showcase the best ofGen AI on streaming TV.
So they're not doing the adtech piece, but we're clearly in
the ad tech domain.
So that's kind of the long-termvision is once one of these is
faster, cheaper and morereliable, we'll be there to
(22:40):
integrate and showcase it.
Speaker 2 (22:41):
Yeah, amazing, yeah, amazing.
So let's pivot the conversationfrom creation, which I would
love to spend another hour justtalking about that, because I
think it's super interesting andit's very relevant to the
reasons that we specified.
But I'm not sure that everybodyis familiar.
(23:03):
You know, earlier you mentioneddemand side platform and you
know DSP.
I think anybody working invideo, at least if you're around
entertainment content that isadvertising as a component
probably has heard DSP and someof these other acronyms, which
(23:27):
I'll let you explain.
But why don't you give a two orthree minute primer on how the
ecosystem works, how it'sstructured, who the players are
and just how these platformseven work?
Because I think this also isvery important.
A lot of discussion today and Ilike that.
(23:48):
You came, you have experiencebuilding like Whirl, for example
, so there's a lot of talk onthe.
You know, in the streaming sidevendors and you know analysts,
observers, that you know fast isgoing to like be this huge boom
.
You know it's going to save theindustry, you know it's this
(24:09):
panacea and and yet you know, weknow that you know some
companies are, are doing welland they're, they're making a go
of it.
Others are, you know theeconomics aren't quite there yet
and everybody has their theoryas to why.
And half the time theconversations I'm in, I'm like
(24:30):
maybe we better explain how allthis actually works.
Yeah, because it's you know.
The implication is, you knowand I'm intensely not going to
name names, but you know,platform XYZ.
You know they just need tocharge more and then they're
going to be profitable, you knowor they and it's like you know,
(24:51):
let's look at how, let's lookat who all the players are and
how the money flows, and so giveus a primer on this, because I
think the audience it'd behelpful and I'm sure there's
somebody out there who's likeyeah, tell me exactly how this
works.
Speaker 3 (25:05):
Totally.
I just want to make it clearthat you asked me to explain ad
tech, which is a daunting task.
The most acronyms per squareinch in any industry.
Speaker 2 (25:18):
So, nate, you know there's the 40,000 foot view.
Why don't you do about the80,000 foot view?
Speaker 3 (25:26):
Well, I think the problem we're trying to solve
does actually help explain.
So we had this great piece thatwas written in a TV news check
that I contributed some quotesabout, and it talked about the
embarrassingly low fill rateproblem in CTV.
So what's really fascinatingright now is we have record
audience watching CTV andeveryone's getting in the ad
(25:49):
supported game, Even Netflix,who for years was like we're
never going to do advertising.
Speaker 2 (25:54):
Now they have advertising.
Speaker 3 (25:56):
There's rumors now Apple is doing it, they are
going to do it, everyone isdoing advertising, so that's
great.
That's one reason to be veryexcited when you're an ad tech
guy in the industry.
So the reality, though, isthere's so many ways my brain's
going.
With Linear, you got your moneyfrom advertising and from
(26:18):
retrans fees from the operatorswho kind of distributed your
content.
For now, at least, there's noclear path where those retrans
fees will carry over into CTV,so you're losing a major source
of revenue if you're likeComcast or NBCU, et cetera.
So we really need to rely onthe advertising ecosystem, the
(26:41):
CTV advertising ecosystem beingrobust.
However, the reality is that,on average, the programmatic
fill rate so programmatic is,you know, buying and selling via
kind of auction styleadvertising on Netflix, hulu, et
cetera, through a demand sideplatform, the biggest kind of
(27:02):
independent one being the tradedesk, the Walt, the Walt gardens
being Google and DV 360.
And there's some reallyinteresting antitrust stuff
going on there.
Google and DV360, and there'ssome really interesting
antitrust stuff going on there.
Speaker 2 (27:13):
Yes, there is.
Speaker 3 (27:15):
And then Meta as well , but Meta's not for CTV, meta's
social and mobile andprogrammatic's kind of the
growing way to transact.
There's also you can buy directto the platforms and usually
that's the bigger guys will getthese big direct spends.
They have big sales teams etcetera.
(27:37):
So we're trying to figure outwith broadcaster partners how do
we help them solve for thatsuper low 38% average fill rate.
And one way to do that is toattract millions of SMB
advertisers into the ecosystemthat are looking for
alternatives to just mobile andsearch, because now there's more
(27:58):
ways to target in CTV and we'redoing that by reducing the
friction with Gen AI and thenthose are direct spends on their
platform, which kind ofbypasses all of the problems
with the programmatic auction.
You can think of programmaticas the stock market.
It's like buying stocks andthere's lots of transactions,
(28:21):
billions of transactions and adrequests going on in real time
and there's just issues with thepipes sometimes and we're
trying to help with a solutionto solve that.
Speaker 2 (28:31):
Yeah, so a whole lot more to unpack.
But you made the reference andsort of skipped over and I'm not
sure everybody understands whatyou meant by low fill rates,
fill rates.
And then you haven't mentionedyet CPM't mentioned yet cpms.
And in the conversations I'm apart of, I find that that a
(28:56):
surprisingly high number offolks in the video streaming
world don't really have anunderstanding of even how cpm
relates to the profitabilitypotential of a fast service and
and and they're almost likethese decoupled things and so
it's like advertising.
(29:17):
Is this?
You know, 400 billion whateverthe number is I I don't know,
you know, but you know, fortraditional TV advertising.
And if we just capture this,you know some percentage of it,
we're all going to be.
And if we just capture this,you know some percentage of it
we're all going to begajillionaires.
You know, that's like the.
That's the tenor of theconversation, right?
Then you kind of go hold onhold on folks, cpms, let's do a
(29:41):
little math here.
You know like the numbers don'tpencil out.
So, with that set up, explain,go into more detail about bill
rates and then talk to us abouthow that's related and then and
then cpms and and then how allthis fits in to why, um um, you
(30:03):
know, traditional linear tv,that whole ecosystem is still
printing piles of cash and, andyou know, the fast services are
not, are not quite at theprofitability targets that you
know they certainly would liketo be.
Speaker 3 (30:19):
Yeah.
So the good news that I thinkeveryone watching probably does
watch TV, so they've likely beenannoyed by this analogy.
That will explain fill rate.
So you know, you have a certainamount of ad breaks per hour.
If you're Netflix and now youhave ads in your content and
usually there's like two minutesish per break, and you have
(30:42):
maybe like one, you know, fourads, 30 seconds or 15 seconds,
and the fill rate problem iswhen, instead of showing an ad
during those breaks, what youlikely see you don't see dead
air, you don't just see black.
You'll likely see what we callan ad slate, which is something
like we will be right back, wewill be right back.
(31:03):
Everyone hates we will be rightback.
So we're trying to get fewer.
We will be right backs in place.
Increase the fill rate to putmore businesses there that are
hopefully relevant to you.
You're searching for a new pairof sneakers on Instagram and
then we could actually retargetyou with that same thing on CTV.
(31:26):
Now there's a lot more digitalaudience targeting capabilities
in CTV, so that's the goal is toincrease those fill rates,
which increases thebroadcaster's revenue and
hopefully makes it actually abetter user experience instead
of seeing either the same adfrequency.
Speaker 2 (31:44):
Well, that's, I think , the problem that at least that
I can relate to.
It's like OK, I like the JeepCherokee, but I'm not in the
market for a car right now and Isee the Jeep Cherokee ad on
every single.
You know like come on, guys,show me something different you
know and I'm just picking onJeep I you know that hasn't been
(32:05):
the most recent one, but for awhile it was 100%.
Speaker 3 (32:09):
So, yeah, that's the frequency problem of you know.
There's only so manyadvertisers buying, and so
they're just showing that samead over and over again, which
obviously is a horrible userexperience.
Speaker 2 (32:21):
It, it is.
It is, yeah, exactly so,programmatic.
Um, you know, I think everybodyhas certainly heard the term
and understands kind of theconcept.
Like you said, it's sort of thestock market.
You know, there's this auctionsystem, buying and selling.
Um, maybe you could explain alittle bit more about what the
(32:41):
components are like.
Who are the players in a in aprogrammatic transaction?
Um, yeah, you know, maybe evenname a couple of the key
companies or players that areinvolved, because there's
multiple players involved inthis, you know, yeah, and then
and then, something that I'mcurious about is is the?
(33:06):
Is the ability to do thehyper-personalized?
Is the ability to do thehyper-personalized,
hyper-targeted, just like whatyou said?
So you geolocate and you say,hey, mark is in the Phoenix
metro area, so they're probablygoing to show something, phoenix
metro.
But how cool would it be if andsome people may not think this
(33:28):
is cool, but I actually thinkit's quite useful, you know if
not only was I shown someadvertisement by a business in
the Phoenix metro area, butsomebody who was in the same zip
code as me, you know so arestaurant that was in, you know
.
So it's within a.
You know, I don't know.
You know, mile, two mile, threemile radius, however big the
(33:49):
zip code is, I don't really know, but it's not that big, but
yeah, so explain first of allprogrammatic what does that
ecosystem look like?
And then what's required to dothis?
You know, hyper personalized,yeah.
Speaker 3 (34:04):
Yeah, I think it's actually I don't know if it's a
good thing or a bad thing thatI'm so in the weeds in
programmatic ad tech that I'mprobably going to keep assuming
too much.
Speaker 2 (34:15):
Well, I'll ask you to clarify, because I'm not in the
weeds All right, I'll give it ago.
Speaker 3 (34:21):
So let's take like a broadcaster.
I don't know why NBCU keepscoming up like Peacock or
something.
I don't know why NBCU keepscoming up like Peacock or
something.
And then they're going to workwith a lot of SSPs supply side
platforms, which haveconnections to a lot of supply
(34:41):
that will actually put their adrequests up for auction in the
programmatic marketplace.
And there's a lot of SSPs.
And then there's DSPs,demand-side platforms, who kind
of aggregate demand, and that'sthe different advertisers that
will go in and looking to buythat supply.
Speaker 2 (35:01):
So sorry to jump in, but just to clarify for me and
for the audience.
So an SSP would be a publisherin the traditional sense or a
streaming service.
Right, that would be an SSP.
Speaker 3 (35:14):
Well, an SSP basically aggregates.
So DSP and SP are both kind ofaggregators.
Speaker 2 (35:19):
They represent.
Sorry.
They're not the platform.
They're the front door to thoseplatforms, I guess.
Speaker 3 (35:25):
Yeah, ssps represent the publishers and aggregate up
the supply.
Speaker 2 (35:29):
That's what I meant, yeah.
Speaker 3 (35:30):
Yep, so like Pubmatic , magnite those are some of the
biggest ones and then DSPsrepresent and aggregate up the
demand.
So the trade desk is like thebiggest open market one, and
then they're kind of all settinga floor price for a CPM.
So let's say like a $25 averageCTV CPM.
(35:53):
And then you can, a DSP can bidon buying if it's Peacock I
don't want to hurt theirfeelings if I'm putting their
CPMs too low or what have you.
And then- Anyway, $2,500 forPeacock, yeah, they can buy
exactly they can buy uh the adrequests that peacock is selling
(36:13):
through maybe their ssp,pubmatic, magnite, etc.
Um working in the industry.
The reality is a lot ofbroadcasters, the publishers, um
end up thinking, okay, if Ijust keep adding more ssps that
can connect me to more demand,I'll increase my fill rates.
(36:34):
And unfortunately that's notalways the case uh.
So now some of these publishersare working with like 12 to 14
ssps and a lot of them are justlike reselling other uh
impressions.
So it's just taking a cut of acut of a cut.
What we're looking to do withour ads manager is give uh the
(36:55):
publisher a direct way for anadvertiser to be able to buy
their supply.
So it's outside of theprogrammatic auction, so it's
guaranteed to deliver Um andit's a more direct relationship
with the advertiser.
They can now email, market tothem, they can offer them
discounts, et cetera.
(37:17):
So that's kind of our way tohelp the ecosystem increase
revenue and fill rates.
Speaker 2 (37:23):
So you are primary, so your primary relationship is
with the DSP.
Speaker 3 (37:39):
I don't want to pigeonhole us as a demand-side
platform.
That's likely the best analogywhen you're buying through us in
Ads Manager.
We're Gen AI tech, we're, we'regen AI tech, but yes, we're.
We do basically aggregate upthe advertisers to buy through
us.
We work with ad servers then toand SSPs to then display the
(38:02):
ads on publishers yeah, yeah,okay, got it.
Speaker 2 (38:07):
And, and then the advertisers.
You were advertisers.
I jumped in so, as you'restarting to talk about the
advertising side, becausethere's a whole set of companies
and technologies that representthe advertisers right
aggregator.
Speaker 3 (38:27):
Obviously there are self-serve platforms like us who
work directly with theadvertiser, mainly SMBs, and
then there's also agencies thatwill like kind of aggregate up a
smaller pool of advertisers andhelp buy advertising on behalf
(38:47):
of those advertisers across,maybe, various demand side
platforms or do direct buys witha broadcaster.
So really direct buys are likethe Holy Grail.
That's what the upfront TVupfronts are.
Speaker 2 (39:02):
Exactly.
Speaker 3 (39:03):
And it's the big guys that take advantage there, sure
.
Speaker 2 (39:07):
Because the commitments have to be huge to
do a direct buy I mean I meantens of millions or even more.
I mean, what give us a scale?
Speaker 3 (39:17):
yeah, yeah, tens of millions.
Uh, amazon's now a new bigplayer with their prime video
offering and they had like oneof their first up fronts and had
, like you know, record success,but maybe you know vague with
the numbers, but great success,great, great great success Well
what does that mean?
Speaker 2 (39:34):
It was amazing.
Blew away our expectationsExactly.
Speaker 3 (39:38):
Exactly so those really big platforms or
broadcasters are still going tohave their big sales teams
getting those direct campaigns.
What we're really trying to dois give you know maybe those big
platforms will give them Gen AI, but also medium-sized
broadcasters, fast channels.
(39:58):
They can have their own adsmanager in self-serve way to
have direct spend on theirplatform.
So they're still going to plugin to the programmatic ecosystem
.
Just, the reality is at leastright now and there's companies
solving for this the fill ratesare like 38%.
It's not good.
It's not enough money to make aviable business.
(40:18):
So you need other ways to havea direct relationship with the
advertisers and get your fillrates up.
Speaker 2 (40:25):
Yeah, yeah, interesting, yeah, yeah,
interesting Boy.
I just I, I can.
I I'm always amazed at how manyhands are in the cookie jar
when it comes to advertising.
It's just so.
It's such a um, uh, complicatedecosystem and uh, yeah, but
(40:46):
it's amazing.
Well, uh, this has been areally wonderful conversation
and thank you for sharing whatyou've shared, and I really
can't wait to see the launch ofthe Gen AI tools and just the
evolution of the platform.
When we first met I thinkwithin five or 10 minutes of you
(41:10):
explaining what, what you weredoing, I remember saying we need
to have you on voices of video,and this is something that just
even personally, I'm I'm superinterested in and so really see
a lot of potential.
So best of luck, you know, toyou and the team.
We do have a couple questionsthat came in, so mind, mind
(41:33):
taking some questions yeah,let's do it I, I think, I, I
think these will be easy tomanage.
So, um, yeah, so it's one of theadvantages for those who are
watching live.
By the way, um, feel free totype into the chat of whichever
platform you're watching on.
I, I know we're streaming on Idon't know how many platforms
(41:53):
right now, but more than one,but go ahead and log in or type
into the chat and we'll try andget to those.
But okay, so, so here's a goodquestion.
So what's the most unusual usecase that can be covered with
Gen AI today?
So now, this is, I think, justwith all of your experience
(42:16):
building using Gen AI tools.
So I don't think this is reallyadvertising specific, but yeah,
I mean, in all of yourexperience as a creative.
I think you'd have a goodperspective on this.
Speaker 3 (42:30):
So yeah, so I actually have a recent case
study that is, I think, I think,really unusual and awesome.
So we have a law firm out herein California that we work with
that's advertised to ourplatform multiple times and one
of the partners texted me he'slike hey, crazy idea, there was
(42:51):
a injury at an Amazon warehouseout here in California.
Could you turn around like aworkers comp ad today that would
target potentially other peoplethat have been injured at
warehouses and blanket their zipcode so we could see if anyone
else will come forward, becausefor them that's worth hundreds
of thousands, if not millions,of dollars and, oh yeah, 250
(43:14):
minimum personal injuryattorneys.
Speaker 2 (43:16):
I, I'm clear with that business.
That's yeah, that's the side oflaw.
Speaker 3 (43:20):
They are printing money or they potentially can
yeah that's so we did, we did it, we did it wow and ran the
campaign.
Uh, turn it around in less than24 hours.
And he's like I'm going to hityou up every time news breaks
and run campaigns.
And I'm like, oh my gosh, youjust literally gave me a new.
You know, I'm supposed to bethe innovator with the great
(43:42):
product roadmap.
I'm like breaking news Gen AIcommercials Like that is.
You have to play to theadvantages and the speed and the
personalized audience targetingare the advantage Gen AI brings
, and if you would, have broughtthat to a creative shop that'd
be like great give us two weeks.
We'll get that back and he'slike two weeks.
(44:02):
It's in the news right now.
We want to show it to peoplethat are going to be like oh wow
, look at that Amazon thing, Ialso got hurt, maybe I should
come forward.
So it I'm like oh, this is awhole new roadmap thing.
I'm like, hey guys, we need tofigure this out.
So it was awesome.
Speaker 2 (44:18):
I, I, I, I love it and um, you know not to be, um,
you know we're, we're, we're,we're not advocating for.
You know, trying to profit off.
You know off.
You know terrible situationsbut a corollary that I recall I
remember hearing years ago and Ihad family a long time back in
(44:40):
Florida and you know, of course,you have.
You know, in the Midwest youhave.
You know tornadoes and you know, in Florida you have hurricanes
and you know tornadoes and youknow, in Florida you have
hurricanes and you know.
So there's weather and you knowone of these comes through and
you know, all of a suddenthere's hundreds of homes that
need new roofs.
There's, you know, there's allthe damage, right, and they
(45:01):
actually used to call theroofers storm chasers and they
were, you know.
Now, it usually was not a goodthing because usually the work
was a little shoddy and you know.
But they swoop into town becausethe insurance companies are
almost just standing therewriting checks yep you know, and
they would come in and you knowyou think like, wow, how
amazing, um, if you could get anad that's hyper relevant for
(45:25):
some you, you know, for some youknow news or something that's
really relevant at that at thatmoment and in that area you know
, um, wow, that is.
I mean, that's a whole newfrontier for advertising.
Speaker 3 (45:39):
So yeah, I thought it was.
Uh.
I love when, like, a customerblows your mind and put
something on your productroadmap.
Speaker 2 (45:46):
I'm like wow that's a cool product market fit.
Yeah, cool.
Well, there's another questionthat came in and I don't know if
you have worked in the Indiamarket, so if you haven't,
though, maybe you can stillcomment.
And the question is just how doyou see CTV advertising working
(46:07):
in like India?
But if you have experiencethere, answer it for India.
If not India, what's happeningin other markets around the
world, or other countries, etcetera?
Speaker 3 (46:21):
100%.
So that's another amazing GenAI specific use case that can be
solved.
So because we partner with 11labs, we have over 32 languages
and even more dialects that wetranslation and dialect.
That's right, so we actuallyhave run campaigns
internationally.
We did one uh, where we uhbasically automatically uh put
(46:44):
it in uh mandarin and indifferent.
It was a company, company inAsia that wanted to hit
different parts of Asia thatspoke different languages.
We're able to create a versionof those commercials with the
different languages anddifferent dialects.
So in India Hindi, tamil.
We can support it.
Speaker 2 (47:03):
Sure, wow, wow, that's cool.
Yeah, I can definitely see thelanguage translation as being
super powerful.
So talk to us then also aboutanother question that came in is
just and this one, I guess, iskind of general.
(47:24):
I feel like we kind of touchedon it, but maybe it's a good way
to wrap things up.
The question is I'd love tolearn about Gen AI revolution in
CTV, and I feel like this wholeinterview, the whole discussion
has been around really centeredon both the applications, the
(47:45):
tools, the use cases and reallythe value that Gen AI can bring.
But maybe where we'll end is,you know where is this going?
You know, and if I could don'tmake it so crystal ball, like
well, in 10 years, you know, butlike maybe the next, like three
years, like where is all thisgoing?
(48:06):
You know, in terms of Gen AItools, what you're working on,
advertising on CTV, just thewhole thing.
Speaker 3 (48:15):
I can't help but plug yet another streamer-specific
thing, and I wish.
Speaker 2 (48:20):
I had the video, if you see the video.
Speaker 3 (48:23):
I think it's a pretty amazing video.
So we have a product in beta.
We call it Streamer Gen AIInfinite amazing video.
So we have a product in beta,we call it Streamer Gen AI
Infinite.
And the idea here this, to me,is the biggest game changer that
Gen AI is going to have onadvertising.
So when I was at Whirl, theparent company app loving
multi-billion dollar mobileperformance demand side platform
(48:43):
, and they would have a big teamof creatives that would for
every campaign let's say,DoorDash is an advertiser they'd
create like eight to 10variations of the creative,
changing subtle things to try todrive an action, and they run
those eight to 10 creatives onsmall percentages of traffic and
whatever one performed the best, they'd shift the budget over
(49:06):
to.
So multivariate testing or A-Btesting.
Speaker 2 (49:09):
A-b Gen testing yeah, AB Gen AI.
Speaker 3 (49:10):
Yeah, gen AI completely upends this.
So we have a product in beta,like I said, gen AI Infinite
where we can take this so manysteps further, where there's an
infinite amount of variations wecan create, basically in real
time and not just testing, likedifferent colors of buttons.
It's literally differentscripts, different call to
(49:31):
actions, a funny ad, a dramaticad, a serious ad, and then
different voices, et cetera, toreally hone in on and
personalize the advertisementfor the user.
Speaker 2 (49:46):
Incredible.
Speaker 3 (49:47):
That's kind of one piece, and then I'll go the holy
grail.
So here's the user Incredible,that's kind of one piece.
And then I'll go the holy grail.
So here's the holy grail.
We're talking to companiesabout doing this and I think
this is going to happen in lessthan 10 years If you're able to,
as a broadcaster or publisherwhich they are pass us the
context of the show you'rewatching.
Say, you're watching House ofDragons on Max.
(50:09):
So you pass us the context inreal time through the bid
request and advertising.
We could then, at the next adbreak, automatically create,
let's say, a Nike commercialwith dragons in it that plays at
that commercial break.
Now that is crazy.
Speaker 2 (50:31):
That is crazy and I love that we're ending there
because, yeah, I just think, boy, you know there's.
You know we can live our lives,you know, with sort of this
general, you know with optimism,with you know a feeling
abundance, and we approacheverything that way.
Or we can live in fear, right,and I know, around AI in
(50:57):
particular, it seems to beeither one it's like a zero or
one.
It's either zero we're alltoast the world's over, you know
, like I'm gonna go buy a bunkersomewhere and never see the sun
or it's like this is just goingto unlock, you know, new
experiences, realities that youknow we could never even have
(51:19):
dreamed of.
And I'm very much in the lattercamp.
And you know, and a lot of it isjust one comment that I was
going to make early on is I'veseen many, many, many studies
and you probably even have a lotof the data to back this up
where people actually do nothate advertising.
(51:41):
What they hate is irrelevantadvertising.
They hate the repetitive,irrelevant advertising.
They hate the repetitive.
They hate showing a Jeep ad 18times when I'm not even in the
market for a car.
So it's not like, well, I amlooking for a car I'm not so
much interested in SUV, but it'slike I have no interest in a
(52:03):
car, but I've seen a Jeep ad 18times in the last two hours, or
know, or whatever, whatever thecorollary is.
And so with that, you know,it's just super exciting to me
that all of these tools aregoing to be able to make
messaging so much more relevantto me.
You know, based on what theyknow about my, my geolocation,
(52:27):
or, you know, using all the datathat's available, right, and
they're going to be able totarget, and then I'm going to
feel, wow, that actually wasvaluable and useful, and I'm
really glad that I saw that,because I didn't know that
company existed, or I'd heardabout them, but I forgot about
them, or whatever the situationis.
So, yeah, that's cool.
(52:49):
Well, let's end there, jonathan.
Thank you for coming on.
Voices of Video.
Thank you, mark, and we will.
I'm sure you know we'll betracking your progress and we'll
have to have you back as youlaunch, you know, the new
platform.
So Great, cool.
Speaker 3 (53:05):
Sounds like a plan.
Cheers, have a good one.
Yeah, you bet, great Cool.
Sounds like a plan.
Cheers, have a good one.
Speaker 1 (53:13):
Yeah, you bet.
This episode of Voices of Videois brought to you by NetInt
Technologies.
If you are looking forcutting-edge video encoding
solutions, check out NetInt'sproducts at netintcom.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!