Latest / The Yield Doctor with James Deaker / Matt Gay Interview - Ad Operations driving Yield improvements

Transcript

Matt Gay Interview - Ad Operations driving Yield improvements · The Yield Doctor with James Deaker

0:00
my guest today is Matt gay and we're
0:03
going to be talking about how good at
0:05
operations drives good yield management
0:08
and some specifics that you need to be
0:10
thinking about if you're a publisher
0:13
Matt and I first met over 10 years ago
0:16
when he was at Martha Stewart Living on
0:18
the publisher side Matt has deep
0:21
experience in ad operations having been
0:24
the former senior vice president at
0:26
operative a company that specialized in
0:29
ad operation for digital media he has a
0:32
unique perspective having worked in
0:35
technology and adops he's also spent
0:37
time on the publisher side at LinkedIn
0:40
he's worked on the agency side at ipg
0:43
and he has worked in consulting firms
0:46
like Accenture and deoe Matt it is so
0:49
good to have you here today it's great
0:51
to be here James thank you for having me
0:53
excellent Now Matt let's start with some
0:56
of the different types of challenges for
0:58
for ad operations at end of the day
1:01
buyers and sellers both want campaigns
1:04
to be running effectively and and ads to
1:06
be served to people hopefully the right
1:09
people but they value very different
1:11
things so so let's start with the buy
1:14
side what is the most painful thing for
1:17
for advertisers as they're thinking
1:19
about their interaction with the adops
1:21
experience yeah that's a really good
1:23
question and there actually they're um
1:26
they're very different uh it can be very
1:28
different things on the on the bu side
1:30
on the you know the agency side there
1:32
are so many layers of marketing you know
1:35
within a company there's marketing
1:38
operations um there's you know marketing
1:40
planning there's obviously brand
1:41
marketers um there's direct response
1:43
marketing um the marketing organization
1:46
is highly Matrix matrixed um in
1:49
companies and you know the agency is
1:51
kind of left at the end you know with a
1:53
deadline to actually deliver so um what
1:56
often happens is it's a big timing
1:59
Challenge on the agency side because
2:01
things get backed up you know not
2:03
because anybody's obviously trying
2:04
everybody's got the same goals to get
2:05
the campaign you know live you know and
2:07
effective um but there are so many
2:09
moving parts and the agency is you know
2:12
at the end of things um so the biggest
2:15
challenge you know on the agency side
2:17
oftentimes um is to be able to provide a
2:19
lot of analytics um and a lot of you
2:22
know reporting and provide a lot of
2:24
added value um to the campaign U because
2:28
everything is usually always such a to
2:30
get it live that everybody kind of just
2:31
breathes a sigh of relief uh you know
2:33
once it uh you know once it gets um you
2:36
know once it gets live um now and that's
2:39
that's where the fun begins on the you
2:41
know on the publisher side um so now you
2:43
know their job um is to is kind of
2:46
twofold actually and that's where we get
2:48
into you know yield a little bit um you
2:50
know first and foremost and you know the
2:52
umbrella is you know making sure that
2:54
that campaign delivers you know
2:55
according to its um goals and sometimes
2:59
sellers might have oversold a little bit
3:01
uh you know in the you know in the sales
3:03
process um and you know targeting might
3:05
be you know a little bit different the
3:07
demographic might be a little bit
3:08
different um you know the content you
3:10
know may be you know a little bit
3:11
different um and obviously just uh you
3:13
know human response you know might be
3:15
different so on the publisher side you
3:18
know the challenge is making sure that
3:19
that campaign you know delivers always
3:22
uh you know in in full and then the Dual
3:25
challenge is is you know making sure
3:27
it's got the optimum yield you know for
3:29
the public right that's how Publishers
3:31
you know stay um you know stay in
3:33
business so it's also taking a look at
3:35
the campaign and noticing you know what
3:38
might be um you know popping you know a
3:40
little bit more you know where things
3:42
might be a little uh sluggish um and
3:45
maybe you want to up an investment um in
3:47
a you know particular area again
3:50
helping the uh you know the campaign
3:53
achieve its uh achieve its goals and
3:55
that's a big belief that I've always had
3:57
is that you know yield management and
3:59
campaign performance you know don't have
4:00
to be at odds I think they actually
4:02
serve each other very well and I think
4:04
the buy side and sell side uh you know
4:06
can serve each other you know very well
4:08
because at the end of the day you're
4:09
looking for the best experience for the
4:11
consumer um you know which both are
4:12
responsible for doing when you've worked
4:15
with adops teams either on the publisher
4:18
side or or the buy side um what are the
4:22
the skills that you think make for a
4:24
great ad operations team and and you
4:27
know I'll throw some out at you and I
4:29
won't let you say all of the above
4:31
so is not a proper response all the you
4:35
know there are challenges around things
4:37
like knowing the technology being very
4:41
process FOC focused and make sure that
4:43
the campaign went live doing a check-in
4:45
after a few days there is sort of being
4:49
client focused there's a whole range of
4:51
skills that go into making a good adops
4:53
team from your perspective what's you
4:56
know top top of that list for me it's
5:00
curiosity um because I believe that
5:04
curiosity is tough to train in somebody
5:07
all the things that you mentioned are
5:09
very important um but are all things
5:11
that you can train you can train people
5:13
on the technology um you can train
5:16
people on the process yes some are going
5:18
to be better or good or what have you
5:20
and there's going to be a little bit of
5:21
investment of time um but ultimately in
5:23
my opinion what makes a great adops
5:25
professional is a sense of curiosity a
5:28
sense of curiosity
5:30
what if we change the targeting a little
5:33
bit on this campaign um what if we rais
5:37
the frequency cap or reduce the
5:38
frequency cap or what if we open this
5:41
campaign up to um you know different
5:44
types of you know content or why did the
5:47
seller sell this the way they sold it so
5:51
all of these things you know in my
5:53
opinion ultimately make for a better um
5:56
a better campaign and a better exper
5:59
experience for the marketer and The
6:01
Advertiser because now you're adding
6:03
value right now when you go back and you
6:04
talk to the agency you know Andor the
6:07
marketer you have um perspective and you
6:10
know you are bringing something to the
6:12
table other than yes your campaign you
6:14
know ran and full you spent the whole
6:16
budget we hit all the kpis all that's
6:18
terrific um but honestly anymore that's
6:20
what's expected and in order to
6:22
differentiate yourself um you need the
6:24
people on the front lines the people in
6:26
the campaigns um to be able to pick
6:28
their head up and be like H you know
6:31
what if we did this or what if we uh you
6:33
know what if we did that perfect hey um
6:37
do you have an example of from all of
6:40
your time working in the industry of
6:43
where an Advertiser or a publisher has
6:46
behaved inefficiently due to either the
6:49
systems or the processes that are in
6:52
place yeah
6:55
um there's so you're right there are so
6:59
many um what I I guess if I were to sort
7:02
of generalize you know them um because
7:05
they're all very you know similar is
7:07
when I've come into add up situations uh
7:12
on either the buy side or the sell side
7:14
and you start to be curious and ask the
7:16
questions why um the answer often comes
7:19
back is because that's the way we've
7:21
always done it or that's the way we
7:23
that's the way we do things and what I
7:25
started to do was you know what if we
7:28
stop doing this or what if we start
7:30
doing that or what if we change this um
7:34
and you know usually you get met a
7:36
little bit with the well that's not
7:38
going to work because and that's not
7:39
going to work because and we've tried
7:40
that before and you know um you hum me
7:44
you know try you know try it again let's
7:45
try it in pieces you know let's try it
7:47
in a you know in a different way you
7:49
know I'm not a big proponent of blowing
7:51
things up so to speak um but certainly
7:54
with taking pieces um and uh you know
7:58
and and and starting and stopping you
7:59
know doing you know particular things
8:01
because I've just found that there are
8:03
so many things so many things that are
8:06
being done today um that don't actually
8:09
need to be done um and the the second
8:12
part around that is around automation um
8:16
often you know we've been able to
8:17
automate um a lot of tasks either you
8:20
know within the ad server or within the
8:22
campaign um you know management system
8:25
outside of the ad server analytics
8:28
obviously right we're getting into the
8:29
age um you know of AI and different
8:32
kinds of you know analytics you're able
8:33
to you know automate a lot of things so
8:35
the first thing is is to you know stop
8:38
doing or clean out kind of the you know
8:40
the dead wood of of of process and then
8:42
the next layer is to automate what needs
8:45
to be done but what can be you know
8:47
automated which I haven't really seen
8:50
that put many people out of jobs yet
8:52
I've seen that actually the way I
8:54
describe it is you start to work with
8:55
your head and not your hands um your
8:58
adops people now and going back to what
9:00
I was talking about earlier your adops
9:02
people now can add value um you know
9:04
they can actually start to analyze and
9:06
start to think about things and
9:07
everybody should be thinking about yield
9:09
and everybody should be thinking about
9:10
forecasting and inventory optimization I
9:13
know is the yield doctor that's music to
9:14
your ears that everybody should be
9:16
thinking about y management but in in
9:18
you know advertising operations that's
9:20
really what it's all about what about
9:22
have you seen examples where someone
9:25
sort of focused on the wrong part of the
9:28
equation Maybe you know they should be
9:30
focused on multi-touch and they're using
9:32
last touch or other examples like that
9:36
yeah um I there's a there's one
9:39
particular example that comes to mind um
9:42
very you know relatively large you know
9:45
marketing or company with a marketing
9:47
organization spends hundreds of millions
9:49
of dollars a year in um you know in
9:51
digital advertising and I was on the you
9:54
know the publisher side you know adops
9:56
you know yield and you know sales kind
9:58
of you know Al together and this company
10:03
was when again remember I just said they
10:05
spend hundreds of millions of dollars in
10:07
digital marketing uh they were still
10:08
working off a last touch attribution um
10:11
last click and that took me a little bit
10:14
by surprise because this was only you
10:16
know about a year or two ago and uh the
10:18
response we got back was well it's
10:20
easier it's easier to use less click and
10:24
it's tough to argue with to some degree
10:26
it is certainly easier and cleaner um
10:29
but you're actually missing the story um
10:32
and you know at the end of the day
10:34
you're actually wasting money um because
10:38
you can be a lot more efficient with
10:41
your ad dollars if you are looking at
10:44
the whole uh customer Journey or
10:48
multi-touch attribution and to take it
10:51
one step further their system of record
10:54
um worked off of Left touch attribution
10:56
and low cost um and that's what they to
10:59
buy based off of but we had other data
11:03
and they had other data that actually
11:05
showed that they could be much more
11:08
efficient um if they bought differently
11:11
but their processes and their policies
11:14
um and even their systems were such that
11:17
it didn't allow them to actually um you
11:19
know operate you know differently so to
11:21
the tune of wasting you know I don't
11:23
know how much but you know a whole lot
11:24
of money on marketing a lot of
11:26
inefficient um you know campaigns and it
11:29
just kind of goes to show you James the
11:32
the the part that humans really still
11:35
play you know and will continue to play
11:37
you know in uh you know in these roles
11:40
um yes we have you know AI yes we have
11:43
Automation and that is going to automate
11:45
many many different things but at the
11:46
end of the day this inefficiency you
11:48
know was actually caused by and can be
11:51
solved by it actually can be solved by
11:53
people thinking about the whole
11:56
marketing process and actually
11:57
understanding you know the whole
11:59
marketing process so you know I like to
12:01
think of you know technology as a Helper
12:05
and helping me do an automate some of
12:07
the things that I can do uh that it can
12:10
do so that I can then use my time um you
12:12
know thinking more about analyses and
12:14
and and bigger business problems so when
12:16
you think about the evolution of the
12:18
industry at the moment we've got a
12:20
situation where the tools should be
12:22
getting easier to use right we should be
12:25
getting more standardized processes but
12:27
in some respects continue to have the
12:30
industry get more and more complex due
12:34
to new formats digital out of home CTV
12:38
ETA who who's winning in that
12:42
tradeoff um what do you mean who's
12:44
winning who is the complexity
12:46
overwhelming the systems is is life
12:49
getting simpler I see um I think uh I
12:52
think it's you know one for complexity
12:54
zero for uh you know Simplicity and you
12:58
know honestly and
12:59
I you know this is this is just another
13:03
Evolution right we go up and we go down
13:06
you know you and I have been at this for
13:07
you know a long time um you know and
13:10
back in the day you know there were
13:11
banner ads and then all of a sudden you
13:13
know video showed up um and then we had
13:16
Dynamic you know ads um and at the end
13:20
of the day I firmly believe that it's
13:24
the humans that complicate things and
13:27
that can also simplify things so what I
13:30
mean by that is yes there are all of
13:34
these new things going on like you
13:37
mentioned digital um you know out of
13:39
home connected TV um and they all have
13:42
their Associated complicated you know
13:44
Technologies you know along with it um
13:47
which is all great but you have to start
13:49
small and you have to build um one of
13:54
the things you know that um um I always
13:58
tell that we always do and I tell other
14:00
adops teams especially when they're
14:02
looking at you know reporting and yield
14:04
is to make sure your foundation is right
14:06
make sure your data is actually correct
14:08
because if you go starting off you know
14:10
with complicated yield analyses and
14:12
forecasting if your base data isn't uh
14:16
right um then your forecast and your
14:18
yield is going to be wrong the same
14:19
things go with all of these new
14:21
technologies if you don't get them right
14:23
in a small and simple format then when
14:25
you scale um things go haywire and then
14:29
yes the machines do take over um and
14:31
complexity wins and simplicity um you
14:34
know does not win so I think it's just
14:36
up to the humans to take control of the
14:38
technology start small get it to work
14:41
and then scale from there and that's
14:42
where I've seen
14:44
success got it uh let's talk about yield
14:47
since this is the the yield doctor
14:49
Channel what are the mistakes that
14:52
Publishers make in add operations that
14:55
negatively impact yield so one I hit on
14:58
a already right which is around just
15:01
understanding what they've got um so
15:04
they may or may not have a rate card and
15:06
I remember from you know one of your
15:08
lessons you know recently um they had
15:10
the wrong rate card or the old years
15:12
rate card pinned up uh in their in their
15:14
cubicle which I'm sure there's a lot of
15:16
people you know uh smirking uh you know
15:19
about that because that's happened to
15:20
them uh you know as well um but the
15:23
first is to understand what is actually
15:25
going on with your business what are you
15:27
charging what is your yield what is your
15:30
floor what is your average what are your
15:34
highs um how did we do quarter over
15:36
quarter year over year week over week
15:39
how are we do it s interrupt but have
15:41
you seen examples where organizations
15:43
don't understand all those things oh
15:45
yeah oh without a doubt I've I've worked
15:47
in I've worked in two of them um and uh
15:50
and I've seen uh you know many more
15:52
where they don't really actually
15:55
understand what is going on against that
15:56
raate card so it's nice to have a rate
15:58
card but all know if you don't actually
16:00
that goes back to my data point if you
16:02
don't have the clean data to match up
16:04
against your rate card to understand
16:06
what's going on hey the rate card might
16:08
be wrong maybe you've got a bad rate
16:10
card maybe you've got bad floors maybe
16:12
they're too high maybe they're too low
16:15
um you know you never you never know
16:17
until you start a dialogue and you start
16:19
a conversation um the second thing the
16:23
second mistake that um you know to your
16:25
question I've dealt with um and talked
16:28
to many others is having an accurate
16:29
forecast you know adops you know um you
16:33
know machine Jokes Aside right but you
16:35
know adops should be able to accurately
16:39
forecast and predict within a certain
16:42
band what's going to happen um yes there
16:45
are things you know worldly Godly events
16:48
that happen um that are out of your
16:50
control um but for the most part
16:53
people are experts and should be able to
16:55
give an idea of forecasting Revenue as
16:59
well as inventory I've also seen a lot
17:01
of people talk about forecast in terms
17:03
of
17:04
inventory nobody really cares about
17:06
inventory in a business level they care
17:07
about Revenue um so if you can't
17:10
monetize that you know revenue and you
17:12
can't forecast what that's going to be
17:15
um that can be um you know quite you
17:17
know quite a challenge um and I'd say
17:19
the third part is having a um a feedback
17:23
loop and a and a and a conversation you
17:26
know in uh you know internally um adops
17:30
should be ingrained in my opinion and
17:32
you know I'm obviously biased but if you
17:34
think of a you know a hub and spoke you
17:36
know I think adops is the middle right
17:38
is that Hub and they get together they
17:41
can pull in the salespeople they can
17:42
pull in the marketers um they can pull
17:45
in finance uh they can pull in you know
17:48
content creation you know editorial you
17:51
know depending on um you know what your
17:54
uh you know platform is all about um and
17:56
feedback what's going on right what are
17:59
users engaging with okay that's
18:01
interesting what are advertisers buying
18:04
because if users are engaging in one
18:06
thing and advertisers are buying another
18:07
thing we have a disconnect but if we're
18:09
not talking to each other we have no
18:11
idea what is actually going on same
18:14
thing with Finance I spent many hours uh
18:17
with our CFO one of our CFOs um
18:20
explaining the digital advertising
18:22
business and how it actually works
18:24
compared to other businesses because
18:25
it's very different so the third mistake
18:28
I see people make is that they get into
18:30
silos um and the senior people and the
18:33
senior people um you know should be
18:36
communicating and having an you know an
18:38
open uh Communication channel with the
18:40
entire business to um exchange all that
18:42
information and you're saying even
18:44
across Publishers from the the content
18:47
generation side and the advertising side
18:49
working together because often you and I
18:51
been in the space where people talk
18:52
about church and state and that
18:54
separation and they literally use
18:57
religious metaphor
18:59
you're saying that those two sides need
19:01
to work together more closely that's to
19:03
be effective they do James and I've seen
19:05
it increase yield um I've seen it
19:07
increase yield and I've seen it increase
19:09
user um engagement and I've seen it
19:12
increase in Revenue uh which means that
19:14
advertisers are coming more and more you
19:17
know and more often and yeah there are
19:19
certain times where you know editorial
19:22
will be like I don't care whether you
19:23
can sell this or not we're doing this
19:25
because our consumers are duranium
19:27
that's fine but there's other situations
19:29
where we had advertisers come um and you
19:32
know want to be against a certain type
19:35
of content or a certain type of
19:36
targeting um and you know the editorial
19:40
team you know if you will um would be
19:42
happy you know to build against that
19:43
they have their own standards and we
19:45
this was not about asking anybody to
19:47
violate any sort of you know standards
19:49
or anything you know like that um it's
19:51
just about I keep coming back to if you
19:53
can make it a better experience for the
19:55
consumer everybody wins excellent Matt
19:58
as we to the end here do you have any
20:00
final thoughts for the audience and and
20:01
what is the best way for for people to
20:03
find you yeah yeah um well the best way
20:07
you know for people to find me um is uh
20:09
just probably on LinkedIn um you know
20:11
Matthew gay on LinkedIn um my uh my
20:14
phone number is up there and uh you know
20:16
my email address is there as well so um
20:20
you know James I think that you know if
20:22
I were to you know leave with you know
20:25
one sort of comment is you know ad is
20:29
really about being curious and I said it
20:31
before but I think that really you know
20:33
sums things up on all levels whether
20:35
you're a trafficker whether you're an
20:37
outsourced operator in India or the
20:40
Philippines um or whether you're the you
20:42
know vice president of um you know
20:45
Revenue operations it's really about
20:47
being curious adops has all the data in
20:51
the world they have they have the buy
20:53
side data they have the sell-side data
20:54
they've got all kinds of data they've
20:56
got all kinds of information um and uh
20:59
they've got all sorts of smarts and
21:00
they've got all sorts of tools you know
21:02
that are available to them um and they
21:04
can really be you know a resource an
21:07
overall resource not just a revenue
21:09
resource but an overall resource for a
21:10
company um you know if uh you know if
21:12
they get used properly so man thank you
21:15
so much for joining me it's an absolute
21:17
pleasure James thank you very much it
21:19
was really nice to talk to you again