April 7, 2025 • 46 mins
Text us your thoughts on the episode or the show!
What happens when all your digital breadcrumbs—personal emails, professional accounts, device IDs, and physical addresses—get connected into a single identity? Jeremy Katz, SVP of Product Solutions, Identity and Data at Merkle, takes us deep into the fascinating world of identity resolution and first-party data.
Starting with his unconventional path from English major to data analytics leader, Jeremy shares the pivotal moments that shaped his understanding of how organizations can build comprehensive customer views. He breaks down complex concepts like identity graphs, customer data platforms, and the technical challenges of defining seemingly simple terms like "customer." Through real-world examples from his experience implementing enterprise-wide customer data hubs, Jeremy illustrates how companies struggle with and ultimately solve the puzzle of connecting fragmented customer information.
Looking toward the future, Jeremy identifies emerging trends that will reshape how organizations manage and leverage identity—from AI-driven audiences and synthetic data to retail media networks and data collaboration through clean rooms. Whether you're a marketing operations professional trying to implement a CDP or a business leader trying to understand the strategic value of your first-party data, this episode offers crucial insights into one of marketing's most foundational yet complex challenges.
Episode Brought to You By MO Pros
The #1 Community for Marketing Operations Professionals
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Hello, welcome to another episode of OpsCast
brought to you byMarketingOpscom, powered by all
the MoPros out there.
I'm your host, michael Hartman,flying solo today.
We'll get Naomi and or Mikeback soon, I'm sure, and I know
Mike is getting ready for SpringFling 2025 coming up in a
little for a month.
So if you haven't already donethat, go check it out and sign
(00:23):
up, if you can go.
But joining me today is JeremyKatz.
Jeremy is currently SVP ofProduct Solutions, identity and
Data at Merkle.
Jeremy is a data analytics andmarketing technology leader with
15 plus years of experiencedelivering value and building
high performing teams.
He is skilled at solvingcomplex challenges at the
intersection of technology,strategy and storytelling.
(00:45):
His background spans digitalmarketing leadership and ad tech
slash market tech acrossmultiple organizations prior to
joining Merkle.
So, jeremy, thanks for joiningme today.
Speaker 2 (00:54):
Thanks for having me Excited to be here.
Speaker 1 (00:56):
Yeah, this is going to be fun and you know, our
topic is kind of in the realm offirst-party data and things
like that that we're going tofocus on.
But before we, before we getinto that, I'd love for you to,
because you have an interestinguh as I would say is the case
for most of our guests, right ainteresting career journey, and
you recently joined merkle,which is uh, for those who don't
(01:18):
know a large consultancy agency, but prior to that, you spent
most of your working career inin-house at brands.
So once you maybe walk througha little more of your career and
how you know what led to youjoining merkle, and while you're
doing that this is something wedidn't plan for, but I
typically throw this in therelike I would love to know if
(01:40):
there were any sort of key,pivotal moments, you know,
decisions you had to make thatled you one way or another, or
key people, right, who played anoutsized role in your career
trajectory.
Speaker 2 (01:53):
Yeah, great, and that could be one of two ways right.
It could be somebody who Ilearned from and learn what to
emulate, or it could be things Ilearned to mistakes.
Speaker 1 (02:01):
I learned not to repeat as I got out of my career
, things I don't want to do,things I will not do when I'm in
that position, kind of thing.
Speaker 2 (02:08):
Exactly so.
I started my career sort ofwhere many people in data and
analytics start as an Englishand journalism major.
That's a little bit tongue incheek, but when I came into the
corporate world I was reallytrying to find where do I add
value?
And I had a background injournalism and English and I
ended up starting incommunications.
(02:29):
So communications sort ofquickly turned into marketing
and I worked in a lot ofmarketing roles and digital
marketing roles early on in mycareer and a lot of it was
getting thrown in the fire.
So for example, hey, do you wantto learn how to run an email
program?
I'm like sure, I can't codeHTML.
I don't know how to do any ofthat, but I'll learn.
So I got thrown in the mix alot early on.
(02:50):
And yeah, I mean you talk aboutpivotal moments my first job my
boss quit two weeks in.
So all of a sudden I wasleading, you know, marketing for
North America at UTC UnitedTechnologies in Connecticut, for
North America at UTC UnitedTechnologies in Connecticut, and
doing PR, marketing, all kindsof stuff that I was not really
ready for.
But I don't know if you'reready for a lot of things early
(03:12):
in your career.
You got to learn to go.
So learning this marketing as apractitioner and marketing
automation.
I'm not sure we even called itback then, but a lot of it was
hey, you got to figure out.
Speaker 1 (03:24):
It was e-marketing when I was doing it.
Speaker 2 (03:26):
Yeah, E-marketing website marketing Right A lot of
terms that are no longer beingused.
I was able to kind of, like Isay, hop from channel to channel
and really kind of startputting a picture together of
how these things worked inharmony and you know, eventually
I ended up moving into theexciting world of insurance,
which I actually say is veryinteresting when you get into it
(03:48):
, but from the outside it can bea little dry.
And when I did that, Ieventually moved into sort of
more of a support role, I wouldcall it at the center of the
organization.
So now you know, all mycustomers were marketers, like I
, had been up to that point.
So now you know, all mycustomers were marketers, like I
, had been up to that point, andwe had basically built out sort
of a centralized MarTechfunction at that time.
(04:10):
And but I kind of stumbled intothat.
I would say I'm always beeninterested in the technology
aspects and kind of how thingswork.
And it was back in the day whenthe MarTech conference there
was one in Boston used to be inperson.
They're all digital now, butback then it was in person.
Speaker 1 (04:25):
Well, not all of them .
I mean Muff's Blues is not.
Speaker 2 (04:27):
That's right.
There you go.
But I had gone to thatconference and just pulled
together some slides and I cameback to my my CMO.
She wanted to know what what Ilearned and I said, well,
there's a thing called a CDP.
I'm not sure if you've heardabout it, but it's kind of a big
deal.
It's growing and eventuallythat led me into the role to
(04:49):
build a MarTech team.
You know, on board a CDP in theorganization and you know.
So a lot of my career has beenluck and timing, but always
being ready, always stayinghungry and learning and trying
new things and ultimately Ithink that's kind of what led me
to Merkle.
You know I've gone to date onthe analytics, you know, sort of
(05:10):
by accident, not by accident,but you know, as my path has
taken me and you know I decidedto kind of challenge, take on a
new challenge, sort of reinventmyself and sit in a different
perspective yet again, which isnow I get to work with, you know
, companies and clients acrossmany different verticals and I'm
learning a ton again.
I feel like I'm almost back atthe beginning of my career in
some ways with the amount oflearning that's happening.
But yeah, I'm excited at whatI'm doing and I feel like if
you're not learning and ifyou're sort of not progressing
(05:33):
forward, you're falling behind,especially in this climate today
.
So yeah, that's kind of a quicksnapshot.
Speaker 1 (05:40):
So you mentioned you did communications and then into
marketing, and so I'm alwayscurious.
So I have my idea of, like,when I think about
communications versus marketing,what it is.
But how do you differentiatethose?
Speaker 2 (05:54):
Well, I have a very direct take of how I got into, I
would say, communications,which was, you know.
I used to write articles fornewspapers back when they were
printed on paper and deliveredon paper to write articles for
newspapers.
Back when they were printed onpaper and delivered on paper,
and I remember creating articles600 word articles,
investigative sort of journalismin nature and I remember I got
like $100 an article orsomething like that and they
(06:15):
asked me would you want to dolike work for a business profile
?
I think I got paid like 3x theamount for the business profile
and I was like, oh okay, this isa way to take my journalism
skills and maybe apply it forcompanies.
And so my first job, I actuallybrought a portfolio of my
articles in that I had writtenat the time and they're like oh
(06:36):
well, you can write for ourinternal company blog and
actually created an internalsocial media site at the time.
So I was kind of like takingthe same skillset a social media
site at the time.
So I was kind of like takingthe same skill set and I would
call that more like internalcomms, which led into pr and
then supporting our, ourleadership team too.
So that's kind of how I thinkabout the communications aspect.
Speaker 1 (06:54):
But okay, that's kind of how I do too.
I just I was like it's likeinternal communications, pr, ir,
executive, executivepresentations and things like
that, right, okay, yeah, yeah,yeah, yeah, that's as opposed to
marketing, which is morecustomer like, directly customer
focused in general, right,whether it's demand gen or
you're doing sales enablement orproduct marketing, yeah, okay,
(07:15):
interesting.
So my guess is you're the guywho, when you're an English
major, that you like diagrammingsentences because you know you
like the structure, oh, I got anerd out on diagramming
sentences.
Speaker 2 (07:29):
I still would.
I don't do it anymore, but Iyeah.
I love diagramming sentences.
I like breaking down things totheir essence and putting them
together and seeing how theywork.
Speaker 1 (07:38):
So it's so funny because I'm the engineer here in
this call and I'm a data guyand I nearly failed out of
seventh grade english because Ijust like this diagramic sense
thing is the stupidest thingI've ever heard of and I, like I
, I did the bare minimum, like I, literally I think it's the
worst grade I've had all myentire life and I really was
(07:58):
very close to failing that class.
So, um, so irony there, you know, the data nerd, the guy who
likes structure, is the one whodidn't want to diagram sentences
.
All right, so you mentionedthat you've been doing a lot
with data and analytics andrecently in your career you've
been spending time with firstparty data, identity resolution
(08:21):
and you mentioned CDPs, identityresolution and you mentioned
CDPs.
So, yeah, we've had guests onand have talked about CDPs and
we've had guests on and talkedabout first party data.
I don't think we've talked toanyone about identity resolution
.
But maybe what we start with,what are your working
definitions of those and ifthere's any other terms that are
(08:41):
relevant for our conversation,like what's the working
definition of those and how arethey related and different?
Speaker 2 (08:48):
Sure, yeah, and I think you know first party data
is probably one of the easierplaces to start.
You know it's data that you'recollecting, you own, and it
could be current customers.
It could be you know prospectswho fill out a form on your site
but don't convert.
But it's data that you'vecollected as an organization.
You have the rights to thatdata, you own it and you can
(09:09):
leverage that for acquisition,for cross-sell, up-sell, et
cetera, versus like a secondparty where it's a partner
sharing their data with you, ora third party where maybe you're
buying or renting data that youcan use to decorate.
We call it that first partydata Decorate okay.
Which you know.
It's an old-school term, maybea little bit, but I still like
(09:29):
the term decorate.
It makes it sound fancy, Iguess.
But ID resolution to me isthere's all these pieces of
information around who you are.
I'll use myself as an example.
I live in Chicago.
I live in an apartment building, so I have an address
associated with that building.
I've got my name and my firstname and last name on my postal
(09:50):
box downstairs.
I've got an email address thatI use, a Gmail address, and I've
got some Yahoo stuff I use forfancy football, old email
addresses.
I have all these pieces ofinformation.
Call it terrestrially me as aperson that lives in an address.
And then I go on the Internetat my house and I have an IP
(10:11):
address.
You know the user agent pingsagainst my device and knows my
device ID, all these digitalfragments and pieces of
information out there.
And to me, the resolution or theresolving piece is
understanding that all thatbelongs to Jeremy Katz and bring
that down to a single identityRight, and usually that's
(10:31):
through an identity graph whichis looking at the relationship
between all those things and Iwould say, like the distance
between those points, thestronger they are, it's a
shorter distance.
So it's funny because, like, myemail address is the same for
the last 20 plus years, the oneI use, you know, every day.
Sure, my address has changedthree times in the last five
(10:51):
years, so you know it depends onwhich piece of information is
updated and how frequently it'supdated.
But understanding who peopleare is really the foundation of
identity resolution and usuallyyou're engaging, you know, an
outside organization to do that,and that's sort of what we do
in my current role.
There's a number of companiesthat do parts of that.
So now is that is that so?
Speaker 1 (11:16):
yeah, I think we talked about this like my.
Where I cut my teeth in themarketing domain was database
marketing and I was building.
I built a 50 million householddatabase and the whole concept
of householding was new to me atthe time.
It was mostly aconsumer-oriented kind of idea.
But is that related to this, oris it just a separate thing
(11:38):
altogether, or how does that fitin?
Speaker 2 (11:41):
That's a good question.
I think entity resolution isthe broader, you know domain
identity resolutions containedwithin the entity.
So we think about grains and I'mtrying to get too nerdy too
quickly here but grains of data.
Right, like me, as anindividual, I roll up to a
household and it's even morecomplex than that.
Right, I mentioned living in anapartment building, so we have,
(12:02):
like, the family household, theTV household that I have inside
my walls here, and then I livein a unit, a physical unit.
Right, all of that is a grainthat goes up from me as an
individual.
And it's the same on the B2Bside.
Right, like you work at acompany, that's an entity, you
have a role within that companyand you probably have a business
(12:23):
profile.
You know that's the stuff thatyour business email, you're
going on LinkedIn, your LinkedInaccount, you know, has business
content on it.
So it's like those differentgrains.
I think about them as, like thewhole the old school, you know,
grandparent, parent, childrelationship.
You know when you think aboutit and know it could be one to
many.
It could be, you know one toone, but those grains are kind
(12:45):
of what we talk about in in thatentity resolution space
interesting, and my guess is,some of this gets out into the
uh for lack of better term darkweb.
Speaker 1 (12:55):
Because it's funny, because, like literally may have
been today, maybe today oryesterday, I got an email that
was sent to four like one email,but was sent to four different
email addresses that are alllinked to me, two of which I
don't really use.
I think one was from my youknow, my, my right I graduated
(13:19):
college from, and one was likean old hotmail account that I
don't really use anymore and itwas weird.
I was like what did like?
So I was like an old hotmailaccount that I don't really use
anymore and it was weird.
I was like what did?
So?
I was like, oh, somebodysomehow found all this stuff
about me and I probably deletedit because I was like this is
probably not somebody I want tobe dealing with, that they're
going out there and scrapingstuff that's out there, like
that.
So okay, so that all makessense.
(13:44):
I see the connection to mydatabase marketing days when I
was building households.
I get the.
I actually think thecomplexities for B2B are more
significant because of thecomplexity one bit, which is,
you know, not only are you partof an organization but maybe
you're part of a specificbusiness unit or a location,
(14:06):
right.
And if they have multiplelocations or multiple business
units I even worked for onecompany where one person could
be we had sort of separate thisgoes down to CRM, right.
We had separate accounts fordifferent almost like not quite
team level, but kind of liketeam levels for specific things.
(14:26):
We sell this product to thisgroup and this other product to
this other group, and sometimessomebody could be in more than
one place so we'd have likeintentional duplicates too.
So you're dealing with all thatkind of noise.
So maybe before we get into whatyou're doing now, but like when
you were in-house, do you haveany project?
(14:48):
Like, give me an example of howthis concept of these different
things and how did that gethappen?
What were some of the bigchallenges?
And then, how was that?
I guess this is a technologysolution at the end of the day,
right, just bringing this stufftogether.
How did that get leveraged tohelp the business?
Speaker 2 (15:08):
Yeah, and I'll try to tie it back to the CDP example
too, because I think that'swhere I first really got deep
into this.
You know we actually createdyou know for all intents and
purposes a customer database,you know to start here, and it
was called Customer Hub.
That Customer Hub, if you will,had sort of multiple lines of
business, all the information wehad on our customers brought
(15:29):
into one physical.
At the time it was anon-premise database, one place.
Now, when you do that, eveninside an organization, you may
have a customer that sits in twobusiness lines.
To your point, that may havebought two different products.
So we didn't even try to resolveit at first, we just brought it
all together and we said, okay,you know, Michael Hartman has
(15:50):
three email addresses.
What's the best one to use?
If he's given us three emailaddresses, what's the one that
he usually transacts with us on?
And so I think that was thefirst foray was kind of building
an internal asset that we coulduse for customer information
across our organization.
The uses of that are relativelystraightforward, right, Like I
(16:10):
want to send you mail or I wantto send you an email.
You shouldn't have theexperience you had where you get
, you know, three emails sent toyour different email addresses,
as if you're a different personin each one.
So I think there were somereally basic use cases to start.
And then when you bring a CDPon top of that, it's like how do
you ingest all of that?
And some light resolutionhappens in a CDP it's all your
(16:32):
first party data, but then thatgoes out to multiple channels.
So that might go out to emailsite personalization etc.
And you're able to know it'sone person you're orchestrating
from a central place and thendistributing that out across
multiple channels.
So I know CDPs have evolvedquite a bit since the early days
.
Almost everything is called aCDP now and it's very different.
Speaker 1 (16:56):
I was trying not to laugh when you said CDPs, you
know old school, because theyhaven't really been around that
long, right yeah, in the grandscheme of things.
Five, maybe 10 years.
Speaker 2 (17:07):
Yeah, it feels like in the digital marketing data
tech space.
That's like somehow way longer.
But being able to tie thosethings together really helped me
understand all the components.
And then the question was allright, you've got all your
customer data together, you'vegot a CDP, you're able to market
to people.
What about your website?
(17:28):
What about all the differentexperiences happening in your
website?
And we at the time I think wehad like four or five different
portals for different parts ofthe business with different
logins, and I mean just makingsense of that is a relatively
complex challenge.
But then, at the end of the day, you've got data coming in and
you've got to figure out how tomake it all connect.
(17:48):
And I think that's ultimatelywhere it started going deeper
into identity resolution anddigital signals and like
bringing that all back to thesame profile right right, um,
and it's a.
It's a never-ending battle.
Right to get that right sure,okay.
Speaker 1 (18:03):
So I have a couple questions.
So, on the like you brought allthis stuff back together in CDP
and you mentioned pushing itout, I think you said pushing it
I'm paraphrasing pushing itback out to some of the maybe
frontline systems, right,whether it's CRM, Was it being
used to maybe informtransactional stuff?
(18:29):
So, hey, I'm a salesperson, Ican see this activity of key
people in my account and youhave that as something that you
inform, like if I have to make acall or send an email.
Was it being used that way?
Or was it being used for, say,website personalization or
targeting with social based onwhether it's lookalikes or
(18:52):
whatever like how was it?
How was that used once that youdid that kind of brought
everything together?
Speaker 2 (18:59):
Yeah, and that was the fun part for me being a
marketer and knowing whateveryone wanted to do across the
Oregon.
I think I had 180 use cases tostart.
They had to prioritize and youknow, knock some off that list.
So one of the cool things we didwas we used a different service
for identifying people in thephone centers.
(19:21):
So when somebody calls in, itwould pop, you know the phone
number and get their profilestarted.
So when you greet somebody youknow who they are.
So we were really able to takethat to a different level,
improve the match rates andimprove the ability for our
sales reps to be able to knowwho was calling in, in in on the
service side too, and then alsokind of pop a little profile of
(19:41):
that customer.
What's their most recent fouror five interactions, you know.
But there was a lot of thingswe did too.
We were able to help with oursurvey programs for NPS programs
.
Speaker 1 (19:51):
Okay.
Speaker 2 (19:51):
Just by kind of like bringing all that data together
for the deployment.
Speaker 1 (19:54):
So for our listeners and watchers who NPS, if you're
not familiar, net promoter scoreright so.
Speaker 2 (20:05):
Yeah, A lot of acronyms in this space.
Thanks for spelling that outfor me.
And then you know, COVID isactually a really good example
of, you know, having to send anemail out to your entire
customer population a couple oftimes, especially in the
insurance space.
You know, we actually peopleweren't driving All of a sudden.
One day people are driving andthe next day no one's driving.
So a lot of companies, likeours at the time, gave back
(20:26):
premium at that time.
So we had to send a lot ofcommunications to customers and
it was timely because we hadbuilt this capability.
We were able to be very precisein our mailings, understand who
was getting emails, who was,you know, kind of more
traditional paper based and beprecise in how we mailed and
communicated with our entirecustomer population.
You're talking, you know,millions and millions of
(20:48):
customers and relatively largecompany and it was wasn't
seamless, wasn't perfect, but itwas pretty smooth, based on
what we had built.
So just basic communication.
Speaker 1 (20:57):
Yeah, and it was also .
So you're kind of hitting onthe next part of follow up I had
, which is I've been atdifferent places, only one that
I did tried to do the exerciseof defining the characteristics
of this is a customer or this isa person.
What is the minimum?
How do we define that?
(21:18):
Did you go through that kind ofexercise of having clear
definitions about that?
I'm just curious A how long didthat take and how did you?
Because I'm going to guess, ifyou get enough people in the
room, you get 10 people andyou're going to get 10 different
opinions about what is acustomer.
Speaker 2 (21:39):
Yes.
Speaker 1 (21:40):
I joke all the time.
The hardest thing as amarketing ops person is when
somebody comes to you and sayswe need to send an email to all
our customers.
Yeah, what do you mean?
Speaker 2 (21:50):
It seems like that should be an easier answer than
it is.
Speaker 1 (21:54):
It absolutely sounds like it should be a no-brainer,
but it is not.
Speaker 2 (21:58):
Yeah.
So I guess I'll stick with thebecause it's fresh in my mind.
I'll stick with the insuranceexample Independent insurance
agents.
Are those our customers.
The actual insureds that wesupport are those customers
Non-customers who are engaged inan accident with our customers
claimants.
There's roles for each of thesethings and so I think, like I
(22:21):
tried to expand the definitionof a customer to at least
include anybody who is a currentwe call them enforced customer,
has a current policy with us,as well as anybody who is a
current we call them enforcedcustomer, has a current policy
with us, as well as anybody whowe've interacted with.
You know, in the last weactually pulled 10 years of data
in to start that we haveinformation on they're either a
former customer they quoted withus, you know they had some
(22:42):
interaction with us and we hadthem as a core customer, and
then we kind of had the agentsas a separate group.
But think about all theattributes and interactions and
roles you're now attaching toall those profiles, and that's
where I think like it's less,it's less just science, there's
some art to it and getting allthe business people in a room
and making sure they they can,you know, have their input into
(23:04):
how you define those things.
But ultimately, ultimately yougot to keep it simple, right,
it's got to still be usable.
So you can't have, you know, 55versions of, maybe a data
element.
You've got to try to keep itsomewhat grouped logically, so
it's usable.
Speaker 1 (23:18):
So in my experience, I agree with you that we should
keep it simple.
But in my experience, whattends to happen is people keep
thinking of I don't even knowthat I would call them outlier
cases.
But take a small example right,one of your customers over the
lifetime of them being there,like at any point in time they
may be, you know someone who'sin the process of engaging with
(23:40):
you.
They're not currently a payingcustomer, what you call enforced
, right?
Uh, then they become that, uh,they end up being a claimant in
another case, like with someoneelse who's an enforced guest.
So it could easily go like ohwell, this person, at any point
in time or in a given situation,has a different status.
Like, how do you, did you dealwith that, did you?
(24:00):
Oh, yeah, yeah, okay, how didyou go about getting to
resolution on those kinds ofseemingly simple but challenging
questions?
Speaker 2 (24:14):
Yeah, we actually had some really smart people on the
data science side to who helpedme.
In fact, I think we actuallygot a patent on one of the data
models that we built out and itwas really it kind of brought
together time time as a constant, like place and time.
The individual, who they are,doesn't change as much.
Right, the attributes aroundthem change, but who that person
(24:34):
is doesn't change.
And then some of the roles andthe kind of like you know I
won't get into all the detail,but the aspects around the data
model.
So you may be a customer at onepoint, a prospect, at another
point, a claimant, you could bean agent, you could be an
independent agent selling ourproduct and the customer, and so
the ability to sort of build aflexible data model that
(24:55):
reflected those changes overtime is also important.
And now there's a wholepractice around customer journey
analytics and understanding thecustomer journey.
This is in the early stages ofwhen that was kind of picking up
.
So that's kind of what led meinto analytics is they're like
all right now we want you tohelp build out that journey
analytics perspective for ourclaims and operations area.
(25:16):
So yeah, we dealt with thatquite a bit and I think it's not
linear, right, it could change.
It goes back and forth inbetween those situations and, to
be honest, to this day I stillremember the most valuable
customer segment in ourprospecting was always former
customers, because they've seenyou, they've engaged with you,
they know you, so the responserates are always higher.
So, very interesting in how youthink about you know the loop,
(25:41):
as opposed to like a linear youknow flow of how people interact
with you.
Speaker 1 (25:45):
Yeah, and I wanted to drill down into those because I
suspect you know, given what Iknow about the profile of our
listeners and our audience, isthere's going to be enough
people who have probably nevergone through this.
And would you know if they weresaid, hey, let's go, we want to
bring in a CDP or we want tobuild a data warehouse or data
lake that brings all ourcustomer data together?
(26:07):
It will be more complicatedthan you think it will be
because of all these things thatwill come up.
Customer is a great example ofthe kind of word that I talk
about a lot with people, where,if you say that word and you're
(26:29):
in a room of people likeeveryone's hearing it from their
own context and they're goingto have an idea of what that
means and they may overlap withthe person next to them or the
other people in the room, butthere's probably not a hundred
percent overlap.
And, yeah, you get that into abroader context of you know,
just within marketing, youprobably don't have it If you go
(26:49):
marketing and customer supportor we go marketing and sales,
and it starts to get reallychallenging to, to to get to a
decision on that, and so I'mwith you.
You want to keep it simple,which probably also affects the
(27:10):
ability of whatever technologyyou're using to handle volume
and things like that, and thenalso affects usability, but also
you want it to be as completeas you need for the use cases,
like it's not a small task.
So appreciate you sharing that.
Speaker 2 (27:31):
Yeah, if I could opine for one second on CDP.
Speaker 1 (27:33):
Please do, I just did .
Speaker 2 (27:34):
A lot of them.
It's like just load your datain and magically it'll all be
cleaned up and you'll haveperfect profiles, and that's
something that I've learnedthrough the pain of going
through this.
It's not that easy.
You have to do some prep andget your data in order.
When you bring a tool like thatand that goes for any tool
right, that can go for amarketing automation platform
too.
There's upfront work that'llmake your life a lot easier, and
(27:57):
it's not the fun work sometimes, but it's the work that is
going to help you be successfulat the next stage.
Speaker 1 (28:03):
Well, I think this is .
It falls into the category tosome degree.
It falls into the category ofwhy I tell people, like, don't
(28:33):
go down the path of going fromno reporting to say we're going
to build a dashboard becauseyou're going to and also are
issues, and then you can go backand address them right, whether
it's a process issue or asystem issue or a people issue,
and usually some combination.
So like, don't, like, I want tomake sure what I didn't hear
from you is don't do the cdpthing until you get your data in
order, because if you do that,you'll never do a cdp thing,
which is fine.
Speaker 2 (28:47):
But take a step.
Speaker 1 (28:48):
That's how you learn yeah, yeah, yeah, no, I'm a big
believer like, expose this stuffand it brings light on it, and
then you can make improvementsand it becomes a bit of a uh, a
fly.
We can get a flywheel effectright.
You find a problem, you solvethe problem, it gets better.
You find the next problem, yousolve it, you get, you know, and
so I'm a big believer in that.
And you know the problem, itgets better.
You find the next problem, yousolve it, you get, you know, and
so I'm a big believer in that.
(29:08):
And you know, reporting CDP hasprobably fallen into a similar
category.
Okay, so I want to get back intothis identity resolution stuff,
because something that you andI talked about as we were kind
of getting ready for this isthat you've now started work I
don't know, I can't remember ifit was at Merkle or where it was
before, where you were, um, youwere actually connecting
(29:33):
professional profiles andpersonal profiles together about
a person and somehow leveragingthat.
So a number of questions likehey, did I understand that right
?
And B?
Um, how does that generallywork?
And then C, are there anyimplications in terms of privacy
or you know that kind ofconcerns there that you have to
(29:54):
deal with when you're doing that.
Speaker 2 (29:57):
Yeah, I'll try to take the first two and then the
privacy one.
We could go, we could talk alot about that, because there's
obviously that's a that's achallenge in a lot of components
of this, even on the consumerside as well.
You know, and again I Imentioned earlier I think about
how I I got exposed to a lot ofthings very early on in my
career and, like the exampleI've already given, you know, we
had small business, mediumbusinesses, consumers, sort of
(30:21):
all in the in the insurancespace.
We had all those differentcategories, um, so it was
something that we had to, I hadto understand relatively early,
and it's could be differentpartners you work with too.
So it's not always the casethat you're going to work with
somebody who's really good atthe identity resolution for a
consumer and data for a consumer.
You may have to work with adifferent company who's really
(30:43):
good at the business side, right, and oftentimes they approach
it from the other direction.
So when I think of aconsumer-based identity
resolution, you're starting withpeople.
You're starting with theirpersonal profile, like I
mentioned earlier, all theinformation about me as an
individual.
On the business side, you get alot of companies that are
really good at understanding thecompany dimension first.
So they kind of go in inverse.
(31:04):
They go from what's theindustry SIC code and what is
the types of things they selland they kind of start with the
company and then they ladderdown to divisions and
organizations at the top,decision makers, and they kind
of go that way down to theindividual.
I think the other thing I wouldjust mention is it's not that
hard if you think about abusiness email like you don't
(31:27):
have to be that smart to thinkabout.
If I know the format of abusiness email like you don't
have to be that smart to thinkabout.
If I know the format of abusiness email, is it first name
, dot, last name at company dotcom.
It's relatively easy, probablyfor a lot of people that are in
the more in the sales space orB2B space, to understand how to
reach individuals in thatcompany.
The challenge is it's soethereal like your role could
change your company changes.
Role could change your companychanges.
(31:52):
So like the durability of thatbusiness email is not as
powerful as like your physicaladdress or your personal email,
but it's the linkages thathappen across both.
I think LinkedIn is a greatplace you know as an example
right, most people are onLinkedIn.
They have their personal emailaddress is how they sign up for
the platform.
But they have their company andthey have information about
their.
You know their job, you know,probably, history of their jobs,
(32:14):
you know in that profile.
So I'm just going to use thatexample and, you know, think
about how, how LinkedIn is alsomarketing right to individuals
and enabling advertisers toreach those individuals on
platform.
You know it's it's really basedon those business
characteristics.
Once they go off platform,you're still marketing to people
.
So that linkage of like, say,email, your login information
(32:38):
that they're able to aggregateup and that gets to the privacy
piece.
Right, usually it's done in theaggregated way, it's not in an
individual way to build audiencepools for targeting.
And that connectivity issomething that my current
company does really well.
It's an interesting kind ofapproach, which is we have all
this depth of knowledge onpeople.
We've also built this secondary, you know, kind of view of
(33:00):
businesses and businessinformation and now we're trying
to bridge that gap.
So we've connected on those youknow elements, right, those
identification elements like apersonal email and business
email.
But at the end of the day, it'sabout reaching those people with
advertising or with messagingor with experiences, so it can
get relatively complex based onthe information you have.
(33:21):
You could also have phonenumbers.
You could have businessaddresses, right, but it's a
similar concept.
It's just you're going to get asmaller pool to start with.
You're going to actually get adrop from the personal
information into the businessside, so your quality is going
to be hopefully better, but youraudience size is going to
shrink and part of that isbecause of the nature of the
(33:43):
identifiers.
They're not as durable as thepersonal side.
But I started touching on it alittle bit.
I think you know privacy andsecurity there's, there's
concerns that you know arewhether you're in the U S,
outside the U S, that that verystate by state in the U S and
certainly you know, in the EU isprobably the only place I can
(34:04):
think of that has a relativelylarge number of companies under
a single framework with GDPR,but a lot of that is focused on
digital.
You know, if you think aboutlike I'm old enough to remember
the phone book, you would get aphone book dropped in your
driveway.
It's got everyone's address andeveryone's phone number.
Like the nature of howterrestrial data has has evolved
(34:25):
over the last 25 years, 30years, has evolved over the last
25 years, 30 years.
It's pretty easy to find.
You know people's like phonenumbers and addresses and things
.
Even now you know out there onthe web.
So I think there's like almostlike we've been conditioned to
expect that in some ways, andthere's services out there where
you can like send people mailright to their house.
Digital is kind of a newfrontier though.
(34:46):
Digital it's like when you signon to a site now and the pop up
comes up, you know, do youaccept cookies and privacy
policies and all that?
Speaker 1 (34:54):
Yeah, these are terms and conditions which we all
agree to, but nobody ever readsJust click and move past it.
Speaker 2 (34:59):
Yeah, I mean, usually , if you stopped and look at
what they're putting in there,it's going to say things like
you're allowing us to use yourinformation for aggregated
advertising or differentsolutions.
And I'm not saying don't clickon that, that's your own
decision to make.
But you saw the whole Googlething over the last couple of
years.
Right, cookies are going to goaway.
Now they're going to kind ofput that choice back in the
(35:21):
hands of consumers.
Do they want to click yes or noand accept cookies or not?
So it's a rapidly evolvingspace.
And accept cookies or not?
So it's a rapidly evolvingspace.
What I would say is expect thatyou're going to get aggregated
pools of people to target, butat some point they're based on
real identity.
It could be a seed.
You know that it's based on andlookalike models and other
(35:42):
things kind of boost thoseaudiences up and that's what.
Google's doing.
All these walled gardens aredoing is building kind of inside
their, their walls.
You know they're buildingtargeting pools based on
interest.
You know, like I love sneakers,I get targeted with sneaker ads
all the time.
I'm sure I'm part of some poolsall over the place that say you
know sneaker lover, or you knowjordan fan, that, um, that
(36:05):
enables me to get targetedadvertising as I'm crossing
throughout the web, even if theydon't know who I am at an
individual level.
Speaker 1 (36:12):
Interesting, yeah, okay, but that's I think it's
fascinating because I thinkthat's and maybe this is all
related to tied to the COVIDstuff too I think there was a
more of a push to connectpersonal and professional
profiles because people at leasta large chunk of, say, office
workers were now working fromhome.
(36:33):
Right, they were no longer inan office identifying where they
were and who they are.
You know it's, it was a new,new challenge, but it sounds
like there's progress on thatfor better or worse.
Right, it's, it's like I'm.
I'm like you, like I remember atime.
I remember when I was doingdatabase marketing and, um,
(36:56):
there is one question I have foryou because you touched on it.
But, like, I hired third partiesto provide data and then
process data because the volumeswere so big that we didn't have
the internal technology to beable to process it at the volume
it was.
But I remember, even at thattime, right being just sort of
(37:17):
dumbfounded about how much datawas either directly about people
or could be inferred aboutpeople based on things like
their five-digit zip code, right.
And when you got a nine-digitzip code, it's like in the US,
right is even more granular injust how accurate.
It was in a general sense, andnow you've just got gobs more
data because technical term,yeah.
(37:39):
So, uh, do you still?
Do you still find that you haveto when you were doing cdp
stuff and all that, like, um, Iguess now all this is
cloud-based.
This that wasn't the case backwhen I was doing this, so is
there still a lot of dataprocessors that are doing this
for some of the companies you'veworked with, or what's that
(38:01):
market like?
Speaker 2 (38:02):
Yeah, there are.
I mean thinking back again tothings I wish I knew at the time
.
Speaker 1 (38:08):
Right.
Speaker 2 (38:09):
Thinking back again to things I wish I knew at the
time.
We ended up building thecustomer hub asset in a legacy
on-prem database and thenimmediately moved it to the
cloud like a year later, andthat took like a year and a half
, so these things are notovernight.
Luckily, if you start in thecloud today like Snowflake is a
good example there's a lot ofcomposable capabilities now
where you can separate storageand compute.
(38:29):
So now it's like I only pay forthe data that I store and then
when I want to call it or queryit, I'm only paying for that
compute.
I don't have to buy a bunch ofspace ahead of time.
I can kind of like match it tomy supply and demand, which has
been a pretty big game changer.
One of the things that I thinkis fascinating you hear the word
composability everywhere nowLike we're actually taking a lot
(38:51):
of our identity services andinstead of saying, send us your
data in a very secure SFTPsecure way, we'll do the
processing and append data andsend it back to you, we're
saying we're going to package upour apps and then we're going
to containerize them and let youput them in your database so
you can kind of run thoseservices as a native app inside
your environment.
(39:12):
It's almost like bring your ownright, you can bring your own
identity, bring your own datainto your ecosystem.
I will say that's leveled theplaying field a little bit,
where you don't have to now goout and buy a huge set of
on-prem hardware and build thosedatabases.
But you still have thechallenge of it's expensive when
you do massive scale.
So it kind of depends on likewhat industry you're in, what's
(39:33):
your target market.
Are you talking about 5 millionpeople or are you talking about
, like you know, a couplehundred million?
It can scale quickly, but Ithink that's less of a challenge
now than it used to be, largelybecause of the rise of some of
these cloud providers.
Speaker 1 (39:47):
Yeah, so you touched on the volume, which is, I think
you and I both have worked atB2C, b2b you more than me,
probably in both worlds, but Imean my view is generally I
think B2C the big challenge isvolume and B2B the big challenge
is volume and B2B the bigchallenge is the complexity of
(40:10):
the data and the structures andthe lack of controls about the
data, particularly when youstart getting into sales realm,
right, sales and customersuccess, because you've got
people who are trying to get youknow.
They've got incentives notnecessarily to focus on quality,
data quality and it's not theirfault, right, it's just the
reality.
So, like in your experience,what do you see as things that
(40:32):
are similar common challenges inthe B2C space?
Based on the B2B space, do yousee other major differences or
am I off my rocker about theones I saw?
Speaker 2 (40:44):
No, you're right.
I think the volume challengeusually means you have to start
with, let's say, yourfirst-party data, to go back to
that and build a lookalike modeland scale it Right.
Yeah, a couple hundred millionpeople.
We have, I think, like 270million or so people in our
graph.
It's like the 18 pluspopulation of the US.
Like how much of that market doyou want to go after?
(41:04):
And you can kind of scale upfrom there.
The business side is a lot morecomplex, though.
You get down into the all thoserelationships right.
Like you want to reach adecision maker.
You want to reach somebodywho's responsible for buying
this type of software, if you'reselling to them in that regard.
So quality becomes an importantaspect.
But also it's more expensive.
I mean the scale on theconsumer side means you can get
(41:26):
cheaper inventory.
You know at scale you're payinglower CPMs.
Anyone who's advertised onLinkedIn versus advertised like
in Google right, you'll knowlike it's scale on one side and
lower CPMs and a lot tighterscale and it's like more
expensive per.
So I think it actually makes iteven more important that you
know what your target is in theB2B space and you're able to
(41:46):
kind of use the insights you cangather around somebody to make
sure you're, you know, puttingputting your uh, your eggs in
the right basket and making sureyou can drive the results you
want, um, especially when youget to like frequency and other
aspects.
So um yeah, that's the onlything I would say to you is the,
the uh, the spend side of it isdifferent.
Speaker 1 (42:09):
The spend side of it is different.
Spend side is different.
It's higher on the B2B sidebecause of the complexity.
Okay yeah, more costly Kind ofmakes sense to me, right,
Because you're trying to solve alittle more complex network
problem, right?
Connections, all thesedifferent things, Okay, Okay.
(42:30):
Okay, we've covered a ton ofground and we're going to
probably need to wrap up heresoon, but is there anything that
we haven't covered here thatyou like?
Hey, what is?
These are the major trends thatare happening in CDPs or
identity resolution.
You know that people should beaware of as they're working with
their organizations.
Speaker 2 (42:51):
Yeah, so this will be our little future.
We have to use the word AI in aconversation.
We can't talk about data and notmention AI Just ai at the end
of everything.
So I guess the one thing I'llmention is AI-driven audiences,
and synthetic data is a hugething.
So you're getting a lot of likethink about the globe, the
(43:11):
entire world.
It's hard to get addressableIDs on the entire globe,
especially with this patchworkof privacy.
So one of the things that'shappening now is AI driven
audiences or synthetic audiences, where I can like model out a
whole country's population.
I can test things in that,instead of doing like a survey
panel, I can test things outbefore it happens, right.
(43:31):
So that's a really interestingand fascinating space for
anybody who's interested in AIand synthetic data.
I would say retail medianetworks are fascinating.
I'm getting more into that nowthat I'm on the agency side
Retail media networks.
Speaker 1 (43:46):
what does that mean?
Speaker 2 (43:48):
So that means I am a retailer or a company and I have
my first party data.
I'm actually going to become anadvertiser now.
I'm going to use thatinformation that I have for my
customers and I'm going to kindof there's kind of two paths
right, and usually you do alittle bit of both.
I can create audiences of myfirst party customers and can
serve those to other secondparties to use for targeting.
(44:11):
Think about like Walmart andamazon and all those big
companies.
The other side of it is like myown space.
If I'm like a grocer, maybe Ihave like advertisement screens
in my aisles and I can.
Cpg brands can use that spaceto advertise.
Speaker 1 (44:25):
So it's kind of what I thought, yeah, it's kind of
what I thought, yeah, and it'skind of a space that really only
large publishers could dobefore Right.
Speaker 2 (44:33):
Yeah, it's blowing up now.
Yeah.
Speaker 1 (44:34):
Yeah, okay, interesting, and then anything
else.
Speaker 2 (44:39):
Last thing I'll say is data collaboration.
It used to be like you know howdo I use my data to drive my
outcomes, but now it's like howdo I partner with another
company and overlap our data andsee where there's opportunities
and that's where you get into,like clean rooms and other
aspects?
Speaker 1 (44:52):
so just to drop one more buzzword at the end there
I'll clean rooms, yeah, meaning,meaning where we combine our
data, but we don't let like wedon't uh no one's share it.
Yeah, yeah okay, okay, that wasalways used for measurement
before now it's used for.
Speaker 2 (45:11):
Let's actually overlap and see you know what,
what shared audiences we have orwhat unique audiences we have,
and um, a lot of partnershipsare happening now around the
ability to do that at scale now.
So interesting yeah, I'll.
I won't say any more buzzwordsthat'll, otherwise I'll be here
yeah, no, it's okay.
Speaker 1 (45:29):
Um, this has been fascinating and it's a little
bit of me remembering where Istarted my marketing journey.
So very interesting stuff.
Jeremy, appreciate it.
If folks want to connect withyou, learn more about what
you're doing or just follow upwith you.
What's the best way for them todo that?
Speaker 2 (45:47):
I'd say LinkedIn is probably the best way.
Look me up on LinkedIn.
I think it's JeremyCCatz atLinkedIncom, or something very
basic.
Speaker 1 (45:54):
Got it, got it Okay, perfect.
Well, thank you, jeremy.
Again, thanks to our audiencefor continuing to support us and
, as always, if you havesuggestions for topics or guests
or you want to be a guest, feelfree to reach out to Naomi,
mike or me through LinkedIn orthrough the marketingopscom
community and we'd be happy totalk to you about that.
(46:16):
Until next time, bye, everybody.
Hello, welcome to another episode of OpsCast
brought to you byMarketingOpscom, powered by all
the MoPros out there.
I'm your host, michael Hartman,flying solo today.
We'll get Naomi and or Mikeback soon, I'm sure, and I know
Mike is getting ready for SpringFling 2025 coming up in a
little for a month.
So if you haven't already donethat, go check it out and sign
(00:23):
up, if you can go.
But joining me today is JeremyKatz.
Jeremy is currently SVP ofProduct Solutions, identity and
Data at Merkle.
Jeremy is a data analytics andmarketing technology leader with
15 plus years of experiencedelivering value and building
high performing teams.
He is skilled at solvingcomplex challenges at the
intersection of technology,strategy and storytelling.
(00:45):
His background spans digitalmarketing leadership and ad tech
slash market tech acrossmultiple organizations prior to
joining Merkle.
So, jeremy, thanks for joiningme today.
Speaker 2 (00:54):
Thanks for having me Excited to be here.
Speaker 1 (00:56):
Yeah, this is going to be fun and you know, our
topic is kind of in the realm offirst-party data and things
like that that we're going tofocus on.
But before we, before we getinto that, I'd love for you to,
because you have an interestinguh as I would say is the case
for most of our guests, right ainteresting career journey, and
you recently joined merkle,which is uh, for those who don't
(01:18):
know a large consultancy agency, but prior to that, you spent
most of your working career inin-house at brands.
So once you maybe walk througha little more of your career and
how you know what led to youjoining merkle, and while you're
doing that this is something wedidn't plan for, but I
typically throw this in therelike I would love to know if
(01:40):
there were any sort of key,pivotal moments, you know,
decisions you had to make thatled you one way or another, or
key people, right, who played anoutsized role in your career
trajectory.
Speaker 2 (01:53):
Yeah, great, and that could be one of two ways right.
It could be somebody who Ilearned from and learn what to
emulate, or it could be things Ilearned to mistakes.
Speaker 1 (02:01):
I learned not to repeat as I got out of my career
, things I don't want to do,things I will not do when I'm in
that position, kind of thing.
Speaker 2 (02:08):
Exactly so.
I started my career sort ofwhere many people in data and
analytics start as an Englishand journalism major.
That's a little bit tongue incheek, but when I came into the
corporate world I was reallytrying to find where do I add
value?
And I had a background injournalism and English and I
ended up starting incommunications.
(02:29):
So communications sort ofquickly turned into marketing
and I worked in a lot ofmarketing roles and digital
marketing roles early on in mycareer and a lot of it was
getting thrown in the fire.
So for example, hey, do you wantto learn how to run an email
program?
I'm like sure, I can't codeHTML.
I don't know how to do any ofthat, but I'll learn.
So I got thrown in the mix alot early on.
(02:50):
And yeah, I mean you talk aboutpivotal moments my first job my
boss quit two weeks in.
So all of a sudden I wasleading, you know, marketing for
North America at UTC UnitedTechnologies in Connecticut, for
North America at UTC UnitedTechnologies in Connecticut, and
doing PR, marketing, all kindsof stuff that I was not really
ready for.
But I don't know if you'reready for a lot of things early
(03:12):
in your career.
You got to learn to go.
So learning this marketing as apractitioner and marketing
automation.
I'm not sure we even called itback then, but a lot of it was
hey, you got to figure out.
Speaker 1 (03:24):
It was e-marketing when I was doing it.
Speaker 2 (03:26):
Yeah, E-marketing website marketing Right A lot of
terms that are no longer beingused.
I was able to kind of, like Isay, hop from channel to channel
and really kind of startputting a picture together of
how these things worked inharmony and you know, eventually
I ended up moving into theexciting world of insurance,
which I actually say is veryinteresting when you get into it
(03:48):
, but from the outside it can bea little dry.
And when I did that, Ieventually moved into sort of
more of a support role, I wouldcall it at the center of the
organization.
So now you know, all mycustomers were marketers, like I
, had been up to that point.
So now you know, all mycustomers were marketers, like I
, had been up to that point, andwe had basically built out sort
of a centralized MarTechfunction at that time.
(04:10):
And but I kind of stumbled intothat.
I would say I'm always beeninterested in the technology
aspects and kind of how thingswork.
And it was back in the day whenthe MarTech conference there
was one in Boston used to be inperson.
They're all digital now, butback then it was in person.
Speaker 1 (04:25):
Well, not all of them .
I mean Muff's Blues is not.
Speaker 2 (04:27):
That's right.
There you go.
But I had gone to thatconference and just pulled
together some slides and I cameback to my my CMO.
She wanted to know what what Ilearned and I said, well,
there's a thing called a CDP.
I'm not sure if you've heardabout it, but it's kind of a big
deal.
It's growing and eventuallythat led me into the role to
(04:49):
build a MarTech team.
You know, on board a CDP in theorganization and you know.
So a lot of my career has beenluck and timing, but always
being ready, always stayinghungry and learning and trying
new things and ultimately Ithink that's kind of what led me
to Merkle.
You know I've gone to date onthe analytics, you know, sort of
(05:10):
by accident, not by accident,but you know, as my path has
taken me and you know I decidedto kind of challenge, take on a
new challenge, sort of reinventmyself and sit in a different
perspective yet again, which isnow I get to work with, you know
, companies and clients acrossmany different verticals and I'm
learning a ton again.
I feel like I'm almost back atthe beginning of my career in
some ways with the amount oflearning that's happening.
But yeah, I'm excited at whatI'm doing and I feel like if
you're not learning and ifyou're sort of not progressing
(05:33):
forward, you're falling behind,especially in this climate today
.
So yeah, that's kind of a quicksnapshot.
Speaker 1 (05:40):
So you mentioned you did communications and then into
marketing, and so I'm alwayscurious.
So I have my idea of, like,when I think about
communications versus marketing,what it is.
But how do you differentiatethose?
Speaker 2 (05:54):
Well, I have a very direct take of how I got into, I
would say, communications,which was, you know.
I used to write articles fornewspapers back when they were
printed on paper and deliveredon paper to write articles for
newspapers.
Back when they were printed onpaper and delivered on paper,
and I remember creating articles600 word articles,
investigative sort of journalismin nature and I remember I got
like $100 an article orsomething like that and they
(06:15):
asked me would you want to dolike work for a business profile
?
I think I got paid like 3x theamount for the business profile
and I was like, oh okay, this isa way to take my journalism
skills and maybe apply it forcompanies.
And so my first job, I actuallybrought a portfolio of my
articles in that I had writtenat the time and they're like oh
(06:36):
well, you can write for ourinternal company blog and
actually created an internalsocial media site at the time.
So I was kind of like takingthe same skillset a social media
site at the time.
So I was kind of like takingthe same skill set and I would
call that more like internalcomms, which led into pr and
then supporting our, ourleadership team too.
So that's kind of how I thinkabout the communications aspect.
Speaker 1 (06:54):
But okay, that's kind of how I do too.
I just I was like it's likeinternal communications, pr, ir,
executive, executivepresentations and things like
that, right, okay, yeah, yeah,yeah, yeah, that's as opposed to
marketing, which is morecustomer like, directly customer
focused in general, right,whether it's demand gen or
you're doing sales enablement orproduct marketing, yeah, okay,
(07:15):
interesting.
So my guess is you're the guywho, when you're an English
major, that you like diagrammingsentences because you know you
like the structure, oh, I got anerd out on diagramming
sentences.
Speaker 2 (07:29):
I still would.
I don't do it anymore, but Iyeah.
I love diagramming sentences.
I like breaking down things totheir essence and putting them
together and seeing how theywork.
Speaker 1 (07:38):
So it's so funny because I'm the engineer here in
this call and I'm a data guyand I nearly failed out of
seventh grade english because Ijust like this diagramic sense
thing is the stupidest thingI've ever heard of and I, like I
, I did the bare minimum, like I, literally I think it's the
worst grade I've had all myentire life and I really was
(07:58):
very close to failing that class.
So, um, so irony there, you know, the data nerd, the guy who
likes structure, is the one whodidn't want to diagram sentences
.
All right, so you mentionedthat you've been doing a lot
with data and analytics andrecently in your career you've
been spending time with firstparty data, identity resolution
(08:21):
and you mentioned CDPs, identityresolution and you mentioned
CDPs.
So, yeah, we've had guests onand have talked about CDPs and
we've had guests on and talkedabout first party data.
I don't think we've talked toanyone about identity resolution
.
But maybe what we start with,what are your working
definitions of those and ifthere's any other terms that are
(08:41):
relevant for our conversation,like what's the working
definition of those and how arethey related and different?
Speaker 2 (08:48):
Sure, yeah, and I think you know first party data
is probably one of the easierplaces to start.
You know it's data that you'recollecting, you own, and it
could be current customers.
It could be you know prospectswho fill out a form on your site
but don't convert.
But it's data that you'vecollected as an organization.
You have the rights to thatdata, you own it and you can
(09:09):
leverage that for acquisition,for cross-sell, up-sell, et
cetera, versus like a secondparty where it's a partner
sharing their data with you, ora third party where maybe you're
buying or renting data that youcan use to decorate.
We call it that first partydata Decorate okay.
Which you know.
It's an old-school term, maybea little bit, but I still like
(09:29):
the term decorate.
It makes it sound fancy, Iguess.
But ID resolution to me isthere's all these pieces of
information around who you are.
I'll use myself as an example.
I live in Chicago.
I live in an apartment building, so I have an address
associated with that building.
I've got my name and my firstname and last name on my postal
(09:50):
box downstairs.
I've got an email address thatI use, a Gmail address, and I've
got some Yahoo stuff I use forfancy football, old email
addresses.
I have all these pieces ofinformation.
Call it terrestrially me as aperson that lives in an address.
And then I go on the Internetat my house and I have an IP
(10:11):
address.
You know the user agent pingsagainst my device and knows my
device ID, all these digitalfragments and pieces of
information out there.
And to me, the resolution or theresolving piece is
understanding that all thatbelongs to Jeremy Katz and bring
that down to a single identityRight, and usually that's
(10:31):
through an identity graph whichis looking at the relationship
between all those things and Iwould say, like the distance
between those points, thestronger they are, it's a
shorter distance.
So it's funny because, like, myemail address is the same for
the last 20 plus years, the oneI use, you know, every day.
Sure, my address has changedthree times in the last five
(10:51):
years, so you know it depends onwhich piece of information is
updated and how frequently it'supdated.
But understanding who peopleare is really the foundation of
identity resolution and usuallyyou're engaging, you know, an
outside organization to do that,and that's sort of what we do
in my current role.
There's a number of companiesthat do parts of that.
So now is that is that so?
Speaker 1 (11:16):
yeah, I think we talked about this like my.
Where I cut my teeth in themarketing domain was database
marketing and I was building.
I built a 50 million householddatabase and the whole concept
of householding was new to me atthe time.
It was mostly aconsumer-oriented kind of idea.
But is that related to this, oris it just a separate thing
(11:38):
altogether, or how does that fitin?
Speaker 2 (11:41):
That's a good question.
I think entity resolution isthe broader, you know domain
identity resolutions containedwithin the entity.
So we think about grains and I'mtrying to get too nerdy too
quickly here but grains of data.
Right, like me, as anindividual, I roll up to a
household and it's even morecomplex than that.
Right, I mentioned living in anapartment building, so we have,
(12:02):
like, the family household, theTV household that I have inside
my walls here, and then I livein a unit, a physical unit.
Right, all of that is a grainthat goes up from me as an
individual.
And it's the same on the B2Bside.
Right, like you work at acompany, that's an entity, you
have a role within that companyand you probably have a business
(12:23):
profile.
You know that's the stuff thatyour business email, you're
going on LinkedIn, your LinkedInaccount, you know, has business
content on it.
So it's like those differentgrains.
I think about them as, like thewhole the old school, you know,
grandparent, parent, childrelationship.
You know when you think aboutit and know it could be one to
many.
It could be, you know one toone, but those grains are kind
(12:45):
of what we talk about in in thatentity resolution space
interesting, and my guess is,some of this gets out into the
uh for lack of better term darkweb.
Speaker 1 (12:55):
Because it's funny, because, like literally may have
been today, maybe today oryesterday, I got an email that
was sent to four like one email,but was sent to four different
email addresses that are alllinked to me, two of which I
don't really use.
I think one was from my youknow, my, my right I graduated
(13:19):
college from, and one was likean old hotmail account that I
don't really use anymore and itwas weird.
I was like what did like?
So I was like an old hotmailaccount that I don't really use
anymore and it was weird.
I was like what did?
So?
I was like, oh, somebodysomehow found all this stuff
about me and I probably deletedit because I was like this is
probably not somebody I want tobe dealing with, that they're
going out there and scrapingstuff that's out there, like
that.
So okay, so that all makessense.
(13:44):
I see the connection to mydatabase marketing days when I
was building households.
I get the.
I actually think thecomplexities for B2B are more
significant because of thecomplexity one bit, which is,
you know, not only are you partof an organization but maybe
you're part of a specificbusiness unit or a location,
(14:06):
right.
And if they have multiplelocations or multiple business
units I even worked for onecompany where one person could
be we had sort of separate thisgoes down to CRM, right.
We had separate accounts fordifferent almost like not quite
team level, but kind of liketeam levels for specific things.
(14:26):
We sell this product to thisgroup and this other product to
this other group, and sometimessomebody could be in more than
one place so we'd have likeintentional duplicates too.
So you're dealing with all thatkind of noise.
So maybe before we get into whatyou're doing now, but like when
you were in-house, do you haveany project?
(14:48):
Like, give me an example of howthis concept of these different
things and how did that gethappen?
What were some of the bigchallenges?
And then, how was that?
I guess this is a technologysolution at the end of the day,
right, just bringing this stufftogether.
How did that get leveraged tohelp the business?
Speaker 2 (15:08):
Yeah, and I'll try to tie it back to the CDP example
too, because I think that'swhere I first really got deep
into this.
You know we actually createdyou know for all intents and
purposes a customer database,you know to start here, and it
was called Customer Hub.
That Customer Hub, if you will,had sort of multiple lines of
business, all the information wehad on our customers brought
(15:29):
into one physical.
At the time it was anon-premise database, one place.
Now, when you do that, eveninside an organization, you may
have a customer that sits in twobusiness lines.
To your point, that may havebought two different products.
So we didn't even try to resolveit at first, we just brought it
all together and we said, okay,you know, Michael Hartman has
(15:50):
three email addresses.
What's the best one to use?
If he's given us three emailaddresses, what's the one that
he usually transacts with us on?
And so I think that was thefirst foray was kind of building
an internal asset that we coulduse for customer information
across our organization.
The uses of that are relativelystraightforward, right, Like I
(16:10):
want to send you mail or I wantto send you an email.
You shouldn't have theexperience you had where you get
, you know, three emails sent toyour different email addresses,
as if you're a different personin each one.
So I think there were somereally basic use cases to start.
And then when you bring a CDPon top of that, it's like how do
you ingest all of that?
And some light resolutionhappens in a CDP it's all your
(16:32):
first party data, but then thatgoes out to multiple channels.
So that might go out to emailsite personalization etc.
And you're able to know it'sone person you're orchestrating
from a central place and thendistributing that out across
multiple channels.
So I know CDPs have evolvedquite a bit since the early days
.
Almost everything is called aCDP now and it's very different.
Speaker 1 (16:56):
I was trying not to laugh when you said CDPs, you
know old school, because theyhaven't really been around that
long, right yeah, in the grandscheme of things.
Five, maybe 10 years.
Speaker 2 (17:07):
Yeah, it feels like in the digital marketing data
tech space.
That's like somehow way longer.
But being able to tie thosethings together really helped me
understand all the components.
And then the question was allright, you've got all your
customer data together, you'vegot a CDP, you're able to market
to people.
What about your website?
(17:28):
What about all the differentexperiences happening in your
website?
And we at the time I think wehad like four or five different
portals for different parts ofthe business with different
logins, and I mean just makingsense of that is a relatively
complex challenge.
But then, at the end of the day, you've got data coming in and
you've got to figure out how tomake it all connect.
(17:48):
And I think that's ultimatelywhere it started going deeper
into identity resolution anddigital signals and like
bringing that all back to thesame profile right right, um,
and it's a.
It's a never-ending battle.
Right to get that right sure,okay.
Speaker 1 (18:03):
So I have a couple questions.
So, on the like you brought allthis stuff back together in CDP
and you mentioned pushing itout, I think you said pushing it
I'm paraphrasing pushing itback out to some of the maybe
frontline systems, right,whether it's CRM, Was it being
used to maybe informtransactional stuff?
(18:29):
So, hey, I'm a salesperson, Ican see this activity of key
people in my account and youhave that as something that you
inform, like if I have to make acall or send an email.
Was it being used that way?
Or was it being used for, say,website personalization or
targeting with social based onwhether it's lookalikes or
(18:52):
whatever like how was it?
How was that used once that youdid that kind of brought
everything together?
Speaker 2 (18:59):
Yeah, and that was the fun part for me being a
marketer and knowing whateveryone wanted to do across the
Oregon.
I think I had 180 use cases tostart.
They had to prioritize and youknow, knock some off that list.
So one of the cool things we didwas we used a different service
for identifying people in thephone centers.
(19:21):
So when somebody calls in, itwould pop, you know the phone
number and get their profilestarted.
So when you greet somebody youknow who they are.
So we were really able to takethat to a different level,
improve the match rates andimprove the ability for our
sales reps to be able to knowwho was calling in, in in on the
service side too, and then alsokind of pop a little profile of
(19:41):
that customer.
What's their most recent fouror five interactions, you know.
But there was a lot of thingswe did too.
We were able to help with oursurvey programs for NPS programs
.
Speaker 1 (19:51):
Okay.
Speaker 2 (19:51):
Just by kind of like bringing all that data together
for the deployment.
Speaker 1 (19:54):
So for our listeners and watchers who NPS, if you're
not familiar, net promoter scoreright so.
Speaker 2 (20:05):
Yeah, A lot of acronyms in this space.
Thanks for spelling that outfor me.
And then you know, COVID isactually a really good example
of, you know, having to send anemail out to your entire
customer population a couple oftimes, especially in the
insurance space.
You know, we actually peopleweren't driving All of a sudden.
One day people are driving andthe next day no one's driving.
So a lot of companies, likeours at the time, gave back
(20:26):
premium at that time.
So we had to send a lot ofcommunications to customers and
it was timely because we hadbuilt this capability.
We were able to be very precisein our mailings, understand who
was getting emails, who was,you know, kind of more
traditional paper based and beprecise in how we mailed and
communicated with our entirecustomer population.
You're talking, you know,millions and millions of
(20:48):
customers and relatively largecompany and it was wasn't
seamless, wasn't perfect, but itwas pretty smooth, based on
what we had built.
So just basic communication.
Speaker 1 (20:57):
Yeah, and it was also .
So you're kind of hitting onthe next part of follow up I had
, which is I've been atdifferent places, only one that
I did tried to do the exerciseof defining the characteristics
of this is a customer or this isa person.
What is the minimum?
How do we define that?
(21:18):
Did you go through that kind ofexercise of having clear
definitions about that?
I'm just curious A how long didthat take and how did you?
Because I'm going to guess, ifyou get enough people in the
room, you get 10 people andyou're going to get 10 different
opinions about what is acustomer.
Speaker 2 (21:39):
Yes.
Speaker 1 (21:40):
I joke all the time.
The hardest thing as amarketing ops person is when
somebody comes to you and sayswe need to send an email to all
our customers.
Yeah, what do you mean?
Speaker 2 (21:50):
It seems like that should be an easier answer than
it is.
Speaker 1 (21:54):
It absolutely sounds like it should be a no-brainer,
but it is not.
Speaker 2 (21:58):
Yeah.
So I guess I'll stick with thebecause it's fresh in my mind.
I'll stick with the insuranceexample Independent insurance
agents.
Are those our customers.
The actual insureds that wesupport are those customers
Non-customers who are engaged inan accident with our customers
claimants.
There's roles for each of thesethings and so I think, like I
(22:21):
tried to expand the definitionof a customer to at least
include anybody who is a currentwe call them enforced customer,
has a current policy with us,as well as anybody who is a
current we call them enforcedcustomer, has a current policy
with us, as well as anybody whowe've interacted with.
You know, in the last weactually pulled 10 years of data
in to start that we haveinformation on they're either a
former customer they quoted withus, you know they had some
(22:42):
interaction with us and we hadthem as a core customer, and
then we kind of had the agentsas a separate group.
But think about all theattributes and interactions and
roles you're now attaching toall those profiles, and that's
where I think like it's less,it's less just science, there's
some art to it and getting allthe business people in a room
and making sure they they can,you know, have their input into
(23:04):
how you define those things.
But ultimately, ultimately yougot to keep it simple, right,
it's got to still be usable.
So you can't have, you know, 55versions of, maybe a data
element.
You've got to try to keep itsomewhat grouped logically, so
it's usable.
Speaker 1 (23:18):
So in my experience, I agree with you that we should
keep it simple.
But in my experience, whattends to happen is people keep
thinking of I don't even knowthat I would call them outlier
cases.
But take a small example right,one of your customers over the
lifetime of them being there,like at any point in time they
may be, you know someone who'sin the process of engaging with
(23:40):
you.
They're not currently a payingcustomer, what you call enforced
, right?
Uh, then they become that, uh,they end up being a claimant in
another case, like with someoneelse who's an enforced guest.
So it could easily go like ohwell, this person, at any point
in time or in a given situation,has a different status.
Like, how do you, did you dealwith that, did you?
(24:00):
Oh, yeah, yeah, okay, how didyou go about getting to
resolution on those kinds ofseemingly simple but challenging
questions?
Speaker 2 (24:14):
Yeah, we actually had some really smart people on the
data science side to who helpedme.
In fact, I think we actuallygot a patent on one of the data
models that we built out and itwas really it kind of brought
together time time as a constant, like place and time.
The individual, who they are,doesn't change as much.
Right, the attributes aroundthem change, but who that person
(24:34):
is doesn't change.
And then some of the roles andthe kind of like you know I
won't get into all the detail,but the aspects around the data
model.
So you may be a customer at onepoint, a prospect, at another
point, a claimant, you could bean agent, you could be an
independent agent selling ourproduct and the customer, and so
the ability to sort of build aflexible data model that
(24:55):
reflected those changes overtime is also important.
And now there's a wholepractice around customer journey
analytics and understanding thecustomer journey.
This is in the early stages ofwhen that was kind of picking up
.
So that's kind of what led meinto analytics is they're like
all right now we want you tohelp build out that journey
analytics perspective for ourclaims and operations area.
(25:16):
So yeah, we dealt with thatquite a bit and I think it's not
linear, right, it could change.
It goes back and forth inbetween those situations and, to
be honest, to this day I stillremember the most valuable
customer segment in ourprospecting was always former
customers, because they've seenyou, they've engaged with you,
they know you, so the responserates are always higher.
So, very interesting in how youthink about you know the loop,
(25:41):
as opposed to like a linear youknow flow of how people interact
with you.
Speaker 1 (25:45):
Yeah, and I wanted to drill down into those because I
suspect you know, given what Iknow about the profile of our
listeners and our audience, isthere's going to be enough
people who have probably nevergone through this.
And would you know if they weresaid, hey, let's go, we want to
bring in a CDP or we want tobuild a data warehouse or data
lake that brings all ourcustomer data together?
(26:07):
It will be more complicatedthan you think it will be
because of all these things thatwill come up.
Customer is a great example ofthe kind of word that I talk
about a lot with people, where,if you say that word and you're
(26:29):
in a room of people likeeveryone's hearing it from their
own context and they're goingto have an idea of what that
means and they may overlap withthe person next to them or the
other people in the room, butthere's probably not a hundred
percent overlap.
And, yeah, you get that into abroader context of you know,
just within marketing, youprobably don't have it If you go
(26:49):
marketing and customer supportor we go marketing and sales,
and it starts to get reallychallenging to, to to get to a
decision on that, and so I'mwith you.
You want to keep it simple,which probably also affects the
(27:10):
ability of whatever technologyyou're using to handle volume
and things like that, and thenalso affects usability, but also
you want it to be as completeas you need for the use cases,
like it's not a small task.
So appreciate you sharing that.
Speaker 2 (27:31):
Yeah, if I could opine for one second on CDP.
Speaker 1 (27:33):
Please do, I just did .
Speaker 2 (27:34):
A lot of them.
It's like just load your datain and magically it'll all be
cleaned up and you'll haveperfect profiles, and that's
something that I've learnedthrough the pain of going
through this.
It's not that easy.
You have to do some prep andget your data in order.
When you bring a tool like thatand that goes for any tool
right, that can go for amarketing automation platform
too.
There's upfront work that'llmake your life a lot easier, and
(27:57):
it's not the fun work sometimes, but it's the work that is
going to help you be successfulat the next stage.
Speaker 1 (28:03):
Well, I think this is .
It falls into the category tosome degree.
It falls into the category ofwhy I tell people, like, don't
(28:33):
go down the path of going fromno reporting to say we're going
to build a dashboard becauseyou're going to and also are
issues, and then you can go backand address them right, whether
it's a process issue or asystem issue or a people issue,
and usually some combination.
So like, don't, like, I want tomake sure what I didn't hear
from you is don't do the cdpthing until you get your data in
order, because if you do that,you'll never do a cdp thing,
which is fine.
Speaker 2 (28:47):
But take a step.
Speaker 1 (28:48):
That's how you learn yeah, yeah, yeah, no, I'm a big
believer like, expose this stuffand it brings light on it, and
then you can make improvementsand it becomes a bit of a uh, a
fly.
We can get a flywheel effectright.
You find a problem, you solvethe problem, it gets better.
You find the next problem, yousolve it, you get, you know, and
so I'm a big believer in that.
And you know the problem, itgets better.
You find the next problem, yousolve it, you get, you know, and
so I'm a big believer in that.
(29:08):
And you know, reporting CDP hasprobably fallen into a similar
category.
Okay, so I want to get back intothis identity resolution stuff,
because something that you andI talked about as we were kind
of getting ready for this isthat you've now started work I
don't know, I can't remember ifit was at Merkle or where it was
before, where you were, um, youwere actually connecting
(29:33):
professional profiles andpersonal profiles together about
a person and somehow leveragingthat.
So a number of questions likehey, did I understand that right
?
And B?
Um, how does that generallywork?
And then C, are there anyimplications in terms of privacy
or you know that kind ofconcerns there that you have to
(29:54):
deal with when you're doing that.
Speaker 2 (29:57):
Yeah, I'll try to take the first two and then the
privacy one.
We could go, we could talk alot about that, because there's
obviously that's a that's achallenge in a lot of components
of this, even on the consumerside as well.
You know, and again I Imentioned earlier I think about
how I I got exposed to a lot ofthings very early on in my
career and, like the exampleI've already given, you know, we
had small business, mediumbusinesses, consumers, sort of
(30:21):
all in the in the insurancespace.
We had all those differentcategories, um, so it was
something that we had to, I hadto understand relatively early,
and it's could be differentpartners you work with too.
So it's not always the casethat you're going to work with
somebody who's really good atthe identity resolution for a
consumer and data for a consumer.
You may have to work with adifferent company who's really
(30:43):
good at the business side, right, and oftentimes they approach
it from the other direction.
So when I think of aconsumer-based identity
resolution, you're starting withpeople.
You're starting with theirpersonal profile, like I
mentioned earlier, all theinformation about me as an
individual.
On the business side, you get alot of companies that are
really good at understanding thecompany dimension first.
So they kind of go in inverse.
(31:04):
They go from what's theindustry SIC code and what is
the types of things they selland they kind of start with the
company and then they ladderdown to divisions and
organizations at the top,decision makers, and they kind
of go that way down to theindividual.
I think the other thing I wouldjust mention is it's not that
hard if you think about abusiness email like you don't
(31:27):
have to be that smart to thinkabout.
If I know the format of abusiness email like you don't
have to be that smart to thinkabout.
If I know the format of abusiness email, is it first name
, dot, last name at company dotcom.
It's relatively easy, probablyfor a lot of people that are in
the more in the sales space orB2B space, to understand how to
reach individuals in thatcompany.
The challenge is it's soethereal like your role could
change your company changes.
Role could change your companychanges.
(31:52):
So like the durability of thatbusiness email is not as
powerful as like your physicaladdress or your personal email,
but it's the linkages thathappen across both.
I think LinkedIn is a greatplace you know as an example
right, most people are onLinkedIn.
They have their personal emailaddress is how they sign up for
the platform.
But they have their company andthey have information about
their.
You know their job, you know,probably, history of their jobs,
(32:14):
you know in that profile.
So I'm just going to use thatexample and, you know, think
about how, how LinkedIn is alsomarketing right to individuals
and enabling advertisers toreach those individuals on
platform.
You know it's it's really basedon those business
characteristics.
Once they go off platform,you're still marketing to people
.
So that linkage of like, say,email, your login information
(32:38):
that they're able to aggregateup and that gets to the privacy
piece.
Right, usually it's done in theaggregated way, it's not in an
individual way to build audiencepools for targeting.
And that connectivity issomething that my current
company does really well.
It's an interesting kind ofapproach, which is we have all
this depth of knowledge onpeople.
We've also built this secondary, you know, kind of view of
(33:00):
businesses and businessinformation and now we're trying
to bridge that gap.
So we've connected on those youknow elements, right, those
identification elements like apersonal email and business
email.
But at the end of the day, it'sabout reaching those people with
advertising or with messagingor with experiences, so it can
get relatively complex based onthe information you have.
(33:21):
You could also have phonenumbers.
You could have businessaddresses, right, but it's a
similar concept.
It's just you're going to get asmaller pool to start with.
You're going to actually get adrop from the personal
information into the businessside, so your quality is going
to be hopefully better, but youraudience size is going to
shrink and part of that isbecause of the nature of the
(33:43):
identifiers.
They're not as durable as thepersonal side.
But I started touching on it alittle bit.
I think you know privacy andsecurity there's, there's
concerns that you know arewhether you're in the U S,
outside the U S, that that verystate by state in the U S and
certainly you know, in the EU isprobably the only place I can
(34:04):
think of that has a relativelylarge number of companies under
a single framework with GDPR,but a lot of that is focused on
digital.
You know, if you think aboutlike I'm old enough to remember
the phone book, you would get aphone book dropped in your
driveway.
It's got everyone's address andeveryone's phone number.
Like the nature of howterrestrial data has has evolved
(34:25):
over the last 25 years, 30years, has evolved over the last
25 years, 30 years.
It's pretty easy to find.
You know people's like phonenumbers and addresses and things
.
Even now you know out there onthe web.
So I think there's like almostlike we've been conditioned to
expect that in some ways, andthere's services out there where
you can like send people mailright to their house.
Digital is kind of a newfrontier though.
(34:46):
Digital it's like when you signon to a site now and the pop up
comes up, you know, do youaccept cookies and privacy
policies and all that?
Speaker 1 (34:54):
Yeah, these are terms and conditions which we all
agree to, but nobody ever readsJust click and move past it.
Speaker 2 (34:59):
Yeah, I mean, usually , if you stopped and look at
what they're putting in there,it's going to say things like
you're allowing us to use yourinformation for aggregated
advertising or differentsolutions.
And I'm not saying don't clickon that, that's your own
decision to make.
But you saw the whole Googlething over the last couple of
years.
Right, cookies are going to goaway.
Now they're going to kind ofput that choice back in the
(35:21):
hands of consumers.
Do they want to click yes or noand accept cookies or not?
So it's a rapidly evolvingspace.
And accept cookies or not?
So it's a rapidly evolvingspace.
What I would say is expect thatyou're going to get aggregated
pools of people to target, butat some point they're based on
real identity.
It could be a seed.
You know that it's based on andlookalike models and other
(35:42):
things kind of boost thoseaudiences up and that's what.
Google's doing.
All these walled gardens aredoing is building kind of inside
their, their walls.
You know they're buildingtargeting pools based on
interest.
You know, like I love sneakers,I get targeted with sneaker ads
all the time.
I'm sure I'm part of some poolsall over the place that say you
know sneaker lover, or you knowjordan fan, that, um, that
(36:05):
enables me to get targetedadvertising as I'm crossing
throughout the web, even if theydon't know who I am at an
individual level.
Speaker 1 (36:12):
Interesting, yeah, okay, but that's I think it's
fascinating because I thinkthat's and maybe this is all
related to tied to the COVIDstuff too I think there was a
more of a push to connectpersonal and professional
profiles because people at leasta large chunk of, say, office
workers were now working fromhome.
(36:33):
Right, they were no longer inan office identifying where they
were and who they are.
You know it's, it was a new,new challenge, but it sounds
like there's progress on thatfor better or worse.
Right, it's, it's like I'm.
I'm like you, like I remember atime.
I remember when I was doingdatabase marketing and, um,
(36:56):
there is one question I have foryou because you touched on it.
But, like, I hired third partiesto provide data and then
process data because the volumeswere so big that we didn't have
the internal technology to beable to process it at the volume
it was.
But I remember, even at thattime, right being just sort of
(37:17):
dumbfounded about how much datawas either directly about people
or could be inferred aboutpeople based on things like
their five-digit zip code, right.
And when you got a nine-digitzip code, it's like in the US,
right is even more granular injust how accurate.
It was in a general sense, andnow you've just got gobs more
data because technical term,yeah.
(37:39):
So, uh, do you still?
Do you still find that you haveto when you were doing cdp
stuff and all that, like, um, Iguess now all this is
cloud-based.
This that wasn't the case backwhen I was doing this, so is
there still a lot of dataprocessors that are doing this
for some of the companies you'veworked with, or what's that
(38:01):
market like?
Speaker 2 (38:02):
Yeah, there are.
I mean thinking back again tothings I wish I knew at the time
.
Speaker 1 (38:08):
Right.
Speaker 2 (38:09):
Thinking back again to things I wish I knew at the
time.
We ended up building thecustomer hub asset in a legacy
on-prem database and thenimmediately moved it to the
cloud like a year later, andthat took like a year and a half
, so these things are notovernight.
Luckily, if you start in thecloud today like Snowflake is a
good example there's a lot ofcomposable capabilities now
where you can separate storageand compute.
(38:29):
So now it's like I only pay forthe data that I store and then
when I want to call it or queryit, I'm only paying for that
compute.
I don't have to buy a bunch ofspace ahead of time.
I can kind of like match it tomy supply and demand, which has
been a pretty big game changer.
One of the things that I thinkis fascinating you hear the word
composability everywhere nowLike we're actually taking a lot
(38:51):
of our identity services andinstead of saying, send us your
data in a very secure SFTPsecure way, we'll do the
processing and append data andsend it back to you, we're
saying we're going to package upour apps and then we're going
to containerize them and let youput them in your database so
you can kind of run thoseservices as a native app inside
your environment.
(39:12):
It's almost like bring your ownright, you can bring your own
identity, bring your own datainto your ecosystem.
I will say that's leveled theplaying field a little bit,
where you don't have to now goout and buy a huge set of
on-prem hardware and build thosedatabases.
But you still have thechallenge of it's expensive when
you do massive scale.
So it kind of depends on likewhat industry you're in, what's
(39:33):
your target market.
Are you talking about 5 millionpeople or are you talking about
, like you know, a couplehundred million?
It can scale quickly, but Ithink that's less of a challenge
now than it used to be, largelybecause of the rise of some of
these cloud providers.
Speaker 1 (39:47):
Yeah, so you touched on the volume, which is, I think
you and I both have worked atB2C, b2b you more than me,
probably in both worlds, but Imean my view is generally I
think B2C the big challenge isvolume and B2B the big challenge
is volume and B2B the bigchallenge is the complexity of
(40:10):
the data and the structures andthe lack of controls about the
data, particularly when youstart getting into sales realm,
right, sales and customersuccess, because you've got
people who are trying to get youknow.
They've got incentives notnecessarily to focus on quality,
data quality and it's not theirfault, right, it's just the
reality.
So, like in your experience,what do you see as things that
(40:32):
are similar common challenges inthe B2C space?
Based on the B2B space, do yousee other major differences or
am I off my rocker about theones I saw?
Speaker 2 (40:44):
No, you're right.
I think the volume challengeusually means you have to start
with, let's say, yourfirst-party data, to go back to
that and build a lookalike modeland scale it Right.
Yeah, a couple hundred millionpeople.
We have, I think, like 270million or so people in our
graph.
It's like the 18 pluspopulation of the US.
Like how much of that market doyou want to go after?
(41:04):
And you can kind of scale upfrom there.
The business side is a lot morecomplex, though.
You get down into the all thoserelationships right.
Like you want to reach adecision maker.
You want to reach somebodywho's responsible for buying
this type of software, if you'reselling to them in that regard.
So quality becomes an importantaspect.
But also it's more expensive.
I mean the scale on theconsumer side means you can get
(41:26):
cheaper inventory.
You know at scale you're payinglower CPMs.
Anyone who's advertised onLinkedIn versus advertised like
in Google right, you'll knowlike it's scale on one side and
lower CPMs and a lot tighterscale and it's like more
expensive per.
So I think it actually makes iteven more important that you
know what your target is in theB2B space and you're able to
(41:46):
kind of use the insights you cangather around somebody to make
sure you're, you know, puttingputting your uh, your eggs in
the right basket and making sureyou can drive the results you
want, um, especially when youget to like frequency and other
aspects.
So um yeah, that's the onlything I would say to you is the,
the uh, the spend side of it isdifferent.
Speaker 1 (42:09):
The spend side of it is different.
Spend side is different.
It's higher on the B2B sidebecause of the complexity.
Okay yeah, more costly Kind ofmakes sense to me, right,
Because you're trying to solve alittle more complex network
problem, right?
Connections, all thesedifferent things, Okay, Okay.
(42:30):
Okay, we've covered a ton ofground and we're going to
probably need to wrap up heresoon, but is there anything that
we haven't covered here thatyou like?
Hey, what is?
These are the major trends thatare happening in CDPs or
identity resolution.
You know that people should beaware of as they're working with
their organizations.
Speaker 2 (42:51):
Yeah, so this will be our little future.
We have to use the word AI in aconversation.
We can't talk about data and notmention AI Just ai at the end
of everything.
So I guess the one thing I'llmention is AI-driven audiences,
and synthetic data is a hugething.
So you're getting a lot of likethink about the globe, the
(43:11):
entire world.
It's hard to get addressableIDs on the entire globe,
especially with this patchworkof privacy.
So one of the things that'shappening now is AI driven
audiences or synthetic audiences, where I can like model out a
whole country's population.
I can test things in that,instead of doing like a survey
panel, I can test things outbefore it happens, right.
(43:31):
So that's a really interestingand fascinating space for
anybody who's interested in AIand synthetic data.
I would say retail medianetworks are fascinating.
I'm getting more into that nowthat I'm on the agency side
Retail media networks.
Speaker 1 (43:46):
what does that mean?
Speaker 2 (43:48):
So that means I am a retailer or a company and I have
my first party data.
I'm actually going to become anadvertiser now.
I'm going to use thatinformation that I have for my
customers and I'm going to kindof there's kind of two paths
right, and usually you do alittle bit of both.
I can create audiences of myfirst party customers and can
serve those to other secondparties to use for targeting.
(44:11):
Think about like Walmart andamazon and all those big
companies.
The other side of it is like myown space.
If I'm like a grocer, maybe Ihave like advertisement screens
in my aisles and I can.
Cpg brands can use that spaceto advertise.
Speaker 1 (44:25):
So it's kind of what I thought, yeah, it's kind of
what I thought, yeah, and it'skind of a space that really only
large publishers could dobefore Right.
Speaker 2 (44:33):
Yeah, it's blowing up now.
Yeah.
Speaker 1 (44:34):
Yeah, okay, interesting, and then anything
else.
Speaker 2 (44:39):
Last thing I'll say is data collaboration.
It used to be like you know howdo I use my data to drive my
outcomes, but now it's like howdo I partner with another
company and overlap our data andsee where there's opportunities
and that's where you get into,like clean rooms and other
aspects?
Speaker 1 (44:52):
so just to drop one more buzzword at the end there
I'll clean rooms, yeah, meaning,meaning where we combine our
data, but we don't let like wedon't uh no one's share it.
Yeah, yeah okay, okay, that wasalways used for measurement
before now it's used for.
Speaker 2 (45:11):
Let's actually overlap and see you know what,
what shared audiences we have orwhat unique audiences we have,
and um, a lot of partnershipsare happening now around the
ability to do that at scale now.
So interesting yeah, I'll.
I won't say any more buzzwordsthat'll, otherwise I'll be here
yeah, no, it's okay.
Speaker 1 (45:29):
Um, this has been fascinating and it's a little
bit of me remembering where Istarted my marketing journey.
So very interesting stuff.
Jeremy, appreciate it.
If folks want to connect withyou, learn more about what
you're doing or just follow upwith you.
What's the best way for them todo that?
Speaker 2 (45:47):
I'd say LinkedIn is probably the best way.
Look me up on LinkedIn.
I think it's JeremyCCatz atLinkedIncom, or something very
basic.
Speaker 1 (45:54):
Got it, got it Okay, perfect.
Well, thank you, jeremy.
Again, thanks to our audiencefor continuing to support us and
, as always, if you havesuggestions for topics or guests
or you want to be a guest, feelfree to reach out to Naomi,
mike or me through LinkedIn orthrough the marketingopscom
community and we'd be happy totalk to you about that.
(46:16):
Until next time, bye, everybody.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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