In this episode
Matt Spiegel, EVP of TruAudience Growth Strategy at TransUnion, argues that AI doesn’t route around a marketer’s data problems; it multiplies them. New TransUnion research found 89% of marketers are increasing AI investment while only 13% have fully scaled it, and 64% are confident they’ll hit their AI goals even though just 36% rate their data and process readiness as high. Spiegel walks through why pilots stall before enterprise integration (data, process, and people), what “data ready” requires in practice, starting with a common identity key, and how to build an AI measurement framework a board will accept when 65% of marketers still judge AI mainly on time and cost savings.
Key takeaways
- A credit bureau is an identity resolution business at its core. Deciding whether someone qualifies for a loan means stitching disparate signals into one view of a person, and TransUnion applied that same capability to marketing identity, audiences, and measurement.
- Most early AI energy in marketing goes into creative. Spiegel sees creative as the natural starting point because it is usually more internalized than audience development or media buying, which agencies often manage.
- Pilots aren’t failing; scaling them is the hard part. He calls it the classic innovator’s challenge and names three blockers: data, process, and people.
- AI is only as good as the data you feed it. When the CRM is separate from the e-commerce database and neither is linked to media exposure, the model gets piecemeal information and returns incremental insights rather than game-changing ones.
- CMOs are slightly overconfident. Roughly two-thirds expect to reach their AI goals, while only about a third say they are ready on data and on people and org design. Spiegel suspects even the 36% data-ready figure is overstated.
- Data readiness starts with a common identity key. “Matt” and “Matthew” should not be two people in your database, and the same key should link digital and physical exposure, past purchases, activation, and measurement.
- Fragmented data compounds like a misaligned tile floor. Being off by inches in the first corner leaves you feet off across the room, and modeling on top of a model widens the error further. It shows up first in audience segmentation and discovery.
- There is no universal set of AI metrics, but every company needs a model. Pick a couple of hero metrics at the C-suite level, define the tactical metrics beneath them, and hold the approach for at least a year.
- Keep measurement intelligence on your side of the aisle. Media companies and technology aggregators each claim too much credit from inside their own walled gardens; cross-channel measurement you own is what drives better decisions.
Chapters
- 01:23 — The gap between AI investment and AI value
- 03:12 — Matt Spiegel’s background and role at TransUnion
- 04:22 — Why a credit bureau is an identity resolution business
- 06:18 — What a CMO means by “scaling AI,” and why creative goes first
- 08:36 — Only 13% have fully scaled AI: why pilots stall
- 11:17 — Siloed functions block cross-functional AI learnings
- 12:52 — AI exposes the people and process problems you already had
- 16:56 — What data readiness for AI looks like: a common identity key
- 20:56 — What breaks first when AI runs on fragmented data
- 23:05 — Building an AI measurement framework a board will accept
- 26:45 — Why walled-garden metrics over-claim credit
- 27:26 — How organizations doing this well are restructuring
- 30:23 — One year from now, and how Matt stays agile
Why do AI pilots stall before enterprise scale?
Spiegel doesn’t think pilots are going poorly. The difficulty is the classic innovator’s problem: an initiative is separated out to learn, and then someone has to build it into the organization and roll it out at scale. Companies also tend to chase “AI” itself rather than the specific business problems they want the technology to solve. The three recurring blockers he sees are data that isn’t in order, processes built around functional silos, and the change management required to get people working differently.
Why AI makes old silos more visible, not less
Marketing organizations are still typically split by function: someone owns search, someone owns social, someone owns analytics, CRM, and on-site. The learnings from making work “agentified” are cross-functional, so without redesigning the org and its go-to-market processes, teams hit friction before they can capture the value. Spiegel compares the moment to the early internet, when companies chased a “digital strategy” instead of asking how the technology could make the business better.
What does data readiness for AI actually look like?
It starts with a common view of identity: one key that connects touchpoints, digital and physical exposure, and past purchases into a single view of a person. From there, link as many data sets as possible to that key, perhaps starting with CRM and e-commerce transactions, then adding media exposure and on-site personalization data. Spiegel favors doing this inside the cloud platform a company already uses, such as Snowflake, AWS, GCP, or Databricks, so data doesn’t have to move, and he stresses partnering with the technology organization, whose priorities sit largely outside marketing.
What breaks first when AI runs on fragmented data
Spiegel’s analogy is laying tile: a couple of inches off in the first corner becomes feet off by the other side of the room. Segments that look highly predictive convert only nominally better, and a model built on top of a model compounds the error. Incomplete or unlinked performance data leads to wrong inferences about what worked. The result is that expected lift in paid or owned channels doesn’t appear, and teams conclude AI doesn’t work, when the real issue is the data and the test cycles.
How to build an AI measurement framework a board will accept
There isn’t a single right answer for the metrics, Spiegel says, but there is a clear need for a model, meaning an approach. What a company sells and where it sits in the market determine whether brand awareness, consideration, favorability, or conversion matters most. Start with the CMO and C-suite peers agreeing on a couple of hero metrics, define which metrics apply at the tactical level, then stick with that approach for at least a year. And keep the intelligence in-house and cross-channel rather than relying on the metrics each media platform reports about itself.
What organizations getting it right have changed
Most organizations are changing rather than changed, Spiegel notes. The examples that excite him are companies that redesigned a workflow, such as a creative brief whose first step is now an LLM, with people connected in parallel instead of in sequence, and companies that have brought their data sets into one common platform. The pattern: a learning agenda, milestones, chasing solutions to business problems rather than the technology, and a willingness to re-architect roles and processes.
FAQ
What percentage of marketing leaders have fully scaled AI? According to the TransUnion research discussed in this episode, only 13% of marketing leaders report that they have fully scaled AI, even though adoption is high.
Why do AI pilots fail to scale in marketing organizations? Matt Spiegel says pilots usually aren’t failing; organizations stall on three things: data that isn’t unified, processes and org designs built around functional silos, and the people-side change required to work differently.
What does “data ready” for AI mean? It means having a common identity key that resolves the same person across records and links touchpoints, media exposure, purchases, activation, and measurement, ideally within the cloud data platform the company already uses.
How should marketers measure the ROI of AI? Agree at the C-suite level on a couple of hero metrics suited to the business, define the tactical metrics beneath them, hold the approach for at least a year, and measure cross-channel on your own side rather than relying on each platform’s self-reported results.
What does TransUnion’s TruAudience do? TruAudience is TransUnion’s suite of marketing data products, covering identity resolution, audience data and segmentation, and measurement through both marketing mix modeling and attribution.
About Matt Spiegel
At TransUnion, Matt drives the growth strategy for TruAudience, the company’s suite of marketing data products. He partners with product and market leaders to align client needs with innovation, ensuring strategic synergy across teams. His leadership stems from deep experience: he previously led TransUnion’s Marketing Solutions and Media & Entertainment businesses, and has spent decades at the intersection of marketing, media, data, and technology.
Before joining TransUnion in 2018, Matt was SVP at MediaLink, where he managed Fortune 100 client relationships and built their Data and Technology Practice. Earlier, he founded Resolution Media, Omnicom’s search marketing agency, and as Global Digital CEO of Omnicom Media Group, he launched the company’s programmatic buying capabilities.
A recognized business builder and change agent, Matt has held roles as an entrepreneur, agency executive, and strategic consultant. He’s a Crain’s Chicago Business 40 Under 40 honoree and currently advises startups and young entrepreneurs. He also serves on the board of The KWOE Foundation. Matt lives in Chicago with his wife, Marnie, and their two children, Leo and Penny.
Resources
TruAudience by TransUnion: https://www.transunion.com/solution/truaudience
The Agile Brand podcast is brought to you by TEKsystems.
The hardest part of martech right now is telling which tools are actually earning their keep. That’s the question running through the MarTech Summit Boston agenda on October 27. We’re attending MarTech Summit Boston on October 27 as a media partner. If you want to join, the code AGILE15 saves you 15%.
We’re proud to be a media partner for #MAICON26 – Oct. 13-15! Learn how AI can power your marketing and business and help you grow smarter. Use code AGILE150 to save!
Need to evaluate martech, AI, or CX platforms and don’t know where to start? The Agile Brand Guide List is now in beta. We’re building the most useful, free-to-use, and up-to-date way to assess what truly matters when evaluating platforms.
Reach your customers with Reddit. Spend $500 in ad spend, get $500 back in ad credit!
Chaser is the only Slack-native project management platform that helps teams turn messages into tracked tasks, automate follow-ups, and maintain team-wide visibility, without adopting another tool. Now integrated with Claude and other GenAI tools. Learn more at trychaser.com and use code AGILEBRAND for a 3-month free trial (normal trial is 14 days).
Start building your own apps with Replit and get $20 off.
Enjoyed the show? Tell us more and give us a rating so others can find the show.
Connect with Greg on LinkedIn.
Don’t miss a thing: get the latest episodes, sign up for our newsletter, and check out Martechipedia, the wiki for marketing technology.
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op.
Transcript
[00:01:23] Greg Kihlström: Hi, I’m Greg Kihlström, host of The Agile Brand, and here’s a question for you. What if the hardest part of scaling AI and marketing isn’t the technology at all, but the fact that you can’t prove what it’s actually doing? Agility isn’t about how quickly you can adopt a tool, it’s about whether your data, your teams, and your measurement can keep up with what the tool is doing once it’s running at scale. Today we’re talking about the widening gap between AI investment and AI value. Specifically, we’re gonna cover why near universal AI adoption hasn’t produced near universal business impact, the readiness problems in data, in process, in ownership that keep organizations parked in the pilot phase, and how to measure AI’s contribution to growth rather than settling for time saved and cost cut. To help me discuss this topic, I’d like to welcome Matt Spiegel, EVP True Audience Growth Strategy at TransUnion. Matt, welcome to the show.
[00:03:03] Matt Spiegel: Hey Greg, thanks for having me. Good to be here.
[00:03:06] Greg Kihlström: Yeah, looking forward to diving in here. Before we do, though, why don’t you give a little background on yourself and your, and your role at TransUnion.
[00:03:12] Matt Spiegel: Yeah, sure. I- I’ve been in the world of digital media marketing technology and data for a few decades now. lots of different sides of the aisle. Sold ads for a while, was in agency days, search agency, was in ad tech, did some management consulting. I’ve been with TransUnion now for about eight years. I joined when we really started the journey to be in this marketing category, and joined when I saw that the wave of identity-backed, identity resolution-backed marketing products was gonna be ever more important. and through the journey at TransUnion, I’ve helped build a lot of our capabilities, ran the product group for a while, helping lead some of the acquisitions we’ve done, run some of our go-to-market teams. And, and today I’m, in this growth strategy role. I really sit as a liaison between our internal teams and our external teams and clients and partners in
[00:03:57] Matt Spiegel: helping to do strategic partnerships and develop relationships and, do things like this.
[00:04:02] Greg Kihlström: Nice. Nice. Love it. And so a lot of listeners may know TransUnion as a credit bureau, but could you frame the work you do, so related to, to s- the rest of TransUnion, but the, the work that you do, with True Audience. what’s its core focus for the enterprise customers that you’re helping?
[00:04:22] Matt Spiegel: Yeah, for sure. Yeah, so the, the link real quick I think is important to understand. If you look at the core of what being a credit bureau is, it’s a really, detailed identity resolution problem. In order to judge whether a bank should give us a loan, a credit card, or anything of the similar, any financial product, you’ve gotta take a lot of disparate signals about who we are as people and stitch it together to make a risk assessment to whether we deserve that, again, financial product. So really if you look at that core, that’s a, again, an identity resolution data analytics problem, and so it was natural for TransUnion to say, “Where else can we apply that type of thinking and those capabilities?” And the marketing world over the last, 15 years was really moving towards that, need and direction. So what we’ve done is we’ve built a business on top of really great identity resolution
[00:05:07] Matt Spiegel: and a great identity graph. And, and at the core what we do for companies is really three core things. We help people understand who are you currently working with. So that is a set of identity resolution products. How do we stitch together different signals to give you that view? how do we help you think about who you wanna reach next? So that’s a set of different types of data attributes, data modeling, data distribution, and really a segment and insights building type view. And then the third thing we do is help you figure out whether those investments in the first two, understanding who I’m working with and understanding who I wanna reach, is it actually working, and we do that through a set of measurement analytics products, both of top-down, marketing mix measurement models, as well as bottoms-up attribution. And so those are our, our core three product lines, and we work with, as quote-unquote, all sides of the aisle.
[00:05:53] Matt Spiegel: we work with big media companies, with big ad tech, with many big marketers, some of the leading retailers, some of the leading financial service institutions. I’m really across the spectrum of things. different clients use us for some things, some clients use us for all. but at the end of the day, we try to provide a consistent view of, of people, both at an individual and a household level, so that you can do great people-based marketing. We wanna be an enabler to that process.
[00:06:18] Greg Kihlström: Okay, great. So yeah, let’s, let’s dive in here and we’re gonna talk about a few things today, but we’re gonna talk specifically about some of TransUnion’s research on, this, what I talked about in the intro, this, this widening gap between AI investment and AI value. And so, before we get too far into the findings, I wanna talk a little bit about, AI is a very broad topic, right? It could, it could mean many, many different things. so, AI-enabled marketing certainly covers a lot of ground. it could be a copywriter using, ChatGPT, a bidding algorithm, system making decisions agentically. when you’re sitting with a CMO, who says that they’re scaling AI hat… How do you f- how do you find out what they’re actually talking about?
[00:07:07] Matt Spiegel: Yeah. Ask a lot of questions.
[00:07:09] Greg Kihlström: [laughs]
[00:07:10] Matt Spiegel: you are right, it, it’s really all over the board. It, it’s interesting, I would say that what I’ve found for the most part today, where most of the early energy is, it is around creative and I think that makes some sense. and by the way, I don’t think that’s anyone’s, full ambition or desire, but I have found that it’s a normal place to start because it is an area where a marketing organization typically has more direct influence than some of the day-to-day decisions that would happen on a, an audience development or a media buying perspective. oftentimes that’s managed by agencies. So I mean, again, and often creative agencies are involved as well, not suggesting they’re not, but there’s something about the creative process which often is more internalized than some of the other processes for a lot of marketing organizations. Again, that’s not true across the board. The industry is far from monolithic, but I’m seeing that a bunch but the ambition when we talk to a lot of companies and when it, whether it’s in my role at TransUnion or just being out in the industry, I sense, and, and if, and I, of course, in my mind, right, we all, I think, see this, is that everyone sees AI, agents, LLMs, how, to your point, the broad spectrum of what this can mean, as fundamentally game-changing, that it should desi- change workflows, it should, change org designs, it should create, create a massive amount of more efficiency and speed and analytical thinking and rigor and uncovering insights. but there’s no question the journey is very, very, very much in the early days.
[00:08:36] Greg Kihlström: Yeah, yeah. And so, the, the research and, and we’ll, we’ll put a link to the, the research in the show notes as well, but the, the research highlights a pretty significant disconnect. So AI adoption is high, but only 13% of marketing leaders report that they’ve fully scaled it. So what’s the strategic reason the organization stall between a working pilot, I hear a lot about this [laughs], there, there’s lots-
[00:09:02] Greg Kihlström: … of pilots going on, but that, the, that gap between the working pilot and enterprise-wide integration, you know-
[00:09:09] Greg Kihlström: … is it the, pilots aren’t successful? Are they not designed properly? What, what, what are some of the things going on here?
[00:09:15] Matt Spiegel: Yeah, no, I don’t, I don’t think it’s that things are going poorly, I think it’s the classic kinda innovator’s challenge, right? Which is you separate an initiative to try to learn some things, but then you gotta figure out how to actually build it in the organization to roll it out at scale and part of that challenge is that, we’re all chasing and collectively AI, when the reality is we wanna leverage the technology to solve different business problems. And one of the hard parts is to figure out, well, which business problems do I wanna most solve and, and what we’ve seen is that the… that there’s really three core challenges that I’d identify. Let’s talk about data, let’s talk about processes, and let’s talk about people, right? So on, on the data front, the most value you’re gonna get from using agents, LLMs to process and create a speed and, and, and an ability to find insights that you weren’t gonna find otherwise, or that an analyst might take you weeks to months to do, maybe not months, but certainly weeks, is you have to actually have all your data house in order, in order to feed the LLM the right information. And so we see a lot of times where companies actually haven’t yet done that. And what, what does that mean? Might mean that the CRM database is separate from your e-commerce transactional database, which might be,
[00:10:31] Matt Spiegel: is probably not linked to your media exposure. And so next thing you’re feeding it piecemeal information and it doesn’t get a full picture. And so the insights that it’s delivering for you are incrementally interesting, probably still some nuggets in there, but maybe not the game-changing that you were hoping for to really prove in an early pilot that you should just, step on the gas. So, so one part for sure is what information are you feeding it? And I, I, I have found myself saying a bunch, like the AI is only as good as the data you feed it, and I do think this is as, as true as, as true can be. Like, that’s a key piece of this. But then even if you’ve got that right, and let’s not, let’s not discount the fact that for many companies, that still is a piece of the challenge. Organizations across marketing are still typically pretty
[00:11:17] Matt Spiegel: siloed in functional area.
[00:11:19] Greg Kihlström: Right.
[00:11:19] Matt Spiegel: Someone owns search, someone owns social, someone owns analytics, someone owns CRM, someone owns on-site, right? And the part of the issue then is the learnings that you’re gonna find from making things, a- agentified, if you will, is cross-functional, right? And so if you haven’t redone your org process and, or your org design and the processes around how you go to market, well, then you probably also have some friction that you can’t yet get there, right? You can’t unlock the goodness, if you will.
[00:11:51] Greg Kihlström: Yeah.
[00:11:51] Matt Spiegel: and then the third piece of that is just the change model that goes with it, which is we as people, it takes work to get us to change and do new things, right? Some of us more wired for change than others. I have always kind of enjoyed that process. But in general, people are comfortable with what’s working and what, they do today, and you’ve really gotta no- navigate a whole new set of behaviors. Which is why I think at the core of the research says, like- CMOs have a ton of confidence. Two thirds say they’re confident they’re gonna get there, but only a third say they’re ready on the data front and on the, and on the people and org front, right? And I, I think that means they’re slightly overconfident, is unfortunately where we are in that journey, right? but anyway, tho- those are the pieces that I wanna unpack, and I, I don’t think this is… This is not shocking to me. I think this is where we are with this journey. but like
[00:12:36] Matt Spiegel: any kind of big change, right, back, I’m old enough to remember when the internet was still potentially just a fad, and a lot of people were chasing a digital strategy versus a how do I use these technologies to make my business better. Think we’re a little bit in that same cycle now with AI.
[00:12:52] Greg Kihlström: Yeah, agreed. I mean, in a sense it’s, wh- I guess why do we keep coming back to the… ’cause every- everything sounds at first like a technology problem, let’s, let’s just use that word, but it ends up being, uh… it tends to be people and processes that are, to, to what you just said, it’s, it’s not even lack of thinking about it, it’s, things are very siloed. They, they were siloed. They were gonna be siloed for other reasons that we were already trying to solve, and now… is this just AI’s kind of accelerating us seeing the, the bad parts of this?
[00:13:30] Matt Spiegel: I think that’s actually a pretty good way to phrase it, right? Which is, I think each new technology innovation that has the game-changing power of which, again, AI… I, I kind of always find myself saying AI, LLMs, agents, right? ‘Cause depending on which phrase or thinking you wanna do, ’cause it is, sure, it is the fact that I have Copilot on my desktop, that’s useful. That changes, that changes things, but it is also the fact that we’re writing agents into our products, right? It’s also the fact that, an AI engine is gonna sit at the coding level of everything we do, right? So there’s all these different things. Anyway, but to your point, I think what happens on these major technology evolutions is that we stare at something and go, “Okay, we can see a future that is materially changed and materially better in some ways, faster,
[00:14:15] Matt Spiegel: more nimble, more, more intelligent.” And clearly AI is as, as, powerful on that potential journey as anything. and so rightly CEOs and boards are going, “What’s your strategy on that? You’ve gotta do it. We don’t wanna be left behind.” and I think we’ve all learned enough to know that the more that you work with LLMs, the smarter it becomes, and so the longer you wait, the more, quote, “behind” you could be.
[00:14:41] Greg Kihlström: Yeah.
[00:14:41] Matt Spiegel: And so we ch- we do chase it, and with some goodness there. There is good that it’s a forcing function for everyone to test and learn and really start to play, right? Listen, certainly 24 months ago, I didn’t use… I wasn’t, I wasn’t on Copilot daily like I am now so there is goodness, but the, the, the, to your point, what happens though is as you go through this journey, you then uncover the reality that there’s a certain level of you now have to change the organizations, and that’s not a given, and those, that’s the hard part which takes more time. [chime]
[00:15:15] Greg Kihlström: A new landing page turns into a pile of tickets and handoffs. Framer helps your team move faster. Teams can collaborate in real time, iterate on the same page, and publish instantly. Agents help you go beyond the vibe coded site and a first draft to bring you a production-ready site faster than ever. Framer is the AI website builder that helps creators, teams, and businesses ship production-ready sites faster than ever while getting every detail right. Prompt, inspect, edit, and publish in one place at a whole new pace. Agents and Framer work alongside teams to streamline collaboration on the same canvas, build custom code components, create and manage CMS content, optimize SEO settings, audit for issues, and ship everything all in one place.
[00:16:00] Greg Kihlström: Learn how you can get more out of your site from a Framer specialist, or get started building for free today at framer.com/agile for 30% off a Framer Pro annual plan. That’s framer.com/agile. Rules and restrictions may apply.
[00:16:56] Greg Kihlström: The other piece that you touched on was the, was the data part as well, and, I do a lot of work in the people and process and change management [laughs], so I, I know how difficult and, how challenging… Yeah, rewarding, but how challenging that can be. On the data side, that also, it’s a, it’s a challenge for so many orgs that I, that I talk with, and, being data ready, it’s, it’s the classic garbage in, garbage out thing, right? So, AI just needs good data. You mentioned, 36% of leaders feel, that their data is ready, meaning there’s, there’s a lot that don’t. what does good data readiness for AI really look like?
[00:17:37] Matt Spiegel: Yeah. Yeah, by the way, I’m not convinced that that 36% isn’t overstated-
[00:17:40] Greg Kihlström: [laughs]
[00:17:40] Matt Spiegel: … that just because it’s confident doesn’t mean it’s there well, listen, I, I believe… So what’s, what does good data readiness look like? So, so number one for me, I, I really do think that the core o- of, the future of this people-based marketing world we wanna live in does start with a common view of identity, a common key that allows you to link different s- different, touchpoints with, for me, in- into a common, view, meaning you don’t want Matt and Matthew to be two different people in your database. You also wanna be able to, as best you can, link my exposures in a digital world and a physical world. You’d like to connect what I’ve bought from you if I’ve pur- purchased from you in the past, right? So being able to have a world, a, a connected layer that actually helps you with that identity resolution is really key. And by the way, some of that means you could, uh,
[00:18:28] Matt Spiegel: y- you could bring that, quote unquote, “in-house,” meaning y- you typically, you’re not gonna do that on your own. Vendors like TransUnion are, I think, necessary partners, but we could give you data to help them do that. But increasingly, we believe strongly in use the, the cloud platform you’re, you’re already using. Most people these days are putting a lot of their data into a Snowflake or an AWS or a GCP or a Databricks or others, and we’ve actually spent a lot of time building our identity resolution products into those platforms so that you actually don’t have to move the data. Point being is I think that’s first and foremost, you gotta do that. Figure out how to get that common key across everything, and do the best you can to connect it to how you activate, whether that’s paid or earned or owned channels, to, through to your measurement, right? So when you, when you buy an activator
[00:19:13] Matt Spiegel: or, you, quote, “mark it” on one identity signal but then measure on another, you’re gonna spend a lot of time going back and forth between, well, which data is right and which data is wrong. So, so one layer, what I’m saying is I think that’s an important layer. Can you get that common key? And then it is, how do you work to actually figure out everything you can link to that common key, right? If you have a common key, maybe you start with two data sets. Maybe it’s your CRM loyalty database and your e-commerce transactional database. Maybe that’s first. But then over time, you’re gonna wanna add in, again, your media exposure. You’re gonna wanna add in your site-side personalization data. So if y- the more you link, the better. So that’s really what good looks like. And then, of course, it’s the… Now, I shouldn’t say of course. There’s a lot of things here which are not this is not the marketer’s bailiwick. This is not where most marketers have grown up, right? I mean, I, I have a weird reality that I’m a marketer by degree, but I’ve spent my career in this kind of data side of the journey. and I’ve learned the fact that, there’s, are these things about cleaning up data and about normalizing it. And like, those are just some block it and tackling things which don’t get a lot of attention in the marketing industry but really matter. And, and the one last thing I would say is I- I’ve, I’ve learned to very much appreciate that the marketers we deal with, they don’t work for marketing companies. They work for companies building product which we’re trying to sell and that means their technology teams have lots of other things to get done. They’re outside of marketing. And so you’ve really gotta find a way to partner with your technology organization and figure out the least friction way to work with them to create this infrastructure, which is why I think some of these cloud platforms are getting so much adoption, is ’cause it takes some of the heavy lifting and build outside of your walls, where you can much more quickly and nimbly, put things together so that’s a little bit of a long and wonky answer, but those are some of the things that we end up talking a lot about with our customers.
[00:20:56] Greg Kihlström: Yeah, well, and, and that idea of AI being a, a multiplier, is, it’s a good thing in that, it, it can take really good, good starting points and, and do a, do a lot of things with it and, more efficiently and, and effectively. But the downside is, when you have fragmented data, certainly it, it can magnify those kinds of things. what, what’s a… when a company is, is running on fragmented data, what, what are you seeing that might break first, and what does the customer see before the marketing team even does?
[00:21:34] Matt Spiegel: Yeah. Yeah. Well, the analogy I’d like to use is if you’re laying tile in a, let’s say, a kitchen floor, and you’re off, by a couple inches in the first corner, by the time you get to the other side of the room, you could be, feet off and I think that’s a kind of a good way to think about what can happen here. So, what, what c- what can go wrong? Well, what, what you think becomes segments which are highly predictive and highly likely to convert are actually not as highly likely to convert and so your expected increase is actually only nominal. by the way, let’s say you then model on top of a model. Now you’re, again, you’re exasperating that change. You’re again going from inches to feet. and so I think that’s where, in, in the media journey, in the marketing effect industry, I think you’re typically gonna find the problem right now around audience segmentation and audience discovery, right? because you will have fed it imperfect information, it will make assumptions and in- inferences which are not as strong as you might want. if the performance data is not complete and/or not properly linked, then it’s making improper inferences of what actually worked and didn’t work. And so, at the end of the day, your, lift that you want in conversion, be it again in paid media channels, where that’s actually got a cost to it, or be in
[00:22:46] Matt Spiegel: owned channels where you just expect a, a higher, rate of conversion on a certain landing page, you won’t see those uplifts as much as you want, and you’re gonna go, “Well, maybe this AI thing doesn’t work.” and in reality, it’s not the AI thing. It’s the data you’re feeding it and the test cycles that you’re running.
[00:23:05] Greg Kihlström: Yeah. Well, and I think that, that leads to, what I’ve certainly seen a lot this, this year, and, and I, I think the, the research points to this as well, is just this, this mandate from boards, from C-level leadership to prove ROI on AI investments. I, I feel like the last few years it was like, “Okay, we gotta do AI,” whatever that means, “but, we gotta, we gotta do this, get started, do pilots.” This year was kind of the, the year of, of reckoning, so to speak. to, to do this, I mean, certainly there’s, I’m, I’m sure there’s some vanity metrics. There’s certainly, I think we’ve been through the, the token maxing to tokenomics. [laughs] that wa- that was a very rapid cycle there from,
[00:23:50] Greg Kihlström: from one extreme to the other. But, to build a credible measurement framework, what, what’s a first step, and where, where do companies often go wrong trying to attribute this?
[00:24:02] Matt Spiegel: Yeah. By the way, I believe s- strongly in the theory of a pendulum swing of business, which is we typically swing from one end to the other, and we go back and forth, and over time, the pendulum swings just at least narrow a little bit and so we, we, we kinda find our, our path, so I, I think you’re right. We’ve, we’ve already done some swinging. but to answer your question, h- what, what does a good measurement framework l- look like? listen, I, I think this is the thing, that, the marketing industry writ large needs to reckon with and help Marketers who spend a lot of money, and again, most money is spent in paid media, to figure out a linked, way to think about measurement from a, a C-suite level all the way down to the practitioner level. And what I have found in, in, all the different clients I’ve worked with and what, what I, what I’ve come to, believe strongly is there isn’t an answer for the metrics, but there is an answer to the need to create a model.
[00:24:59] Matt Spiegel: And what I, I don’t mean in this case a data model, I mean an approach, right? which means y- the- who you are and what you’re selling and your, your position in the market will, will alter whether you care a ton about brand awareness, brand consideration, brand favorability, simply conversion, different other types of purchase intent signals. Again, are you a long considered purchase? Are you fast? Is a, is a car configuration a thing for you, or are you a gum purchase at an aisle, right? That means that measurement isn’t necessar- i- is not, not necessarily, it is not a one-size-fits-tall, fit-all things. But what I, I think the, the, the best place to start is actually from a CMO up with their peers at the C-suite, and then down, is to decide for that company what
[00:25:44] Matt Spiegel: are the, what’s the, what is that, their approach? What are the couple hero metrics they’re going to lean in on? And I do say a couple, typically not one, right? And what’s measured at the C-suite level, and then what’s measured at the granular performance level or, or tactical level, and some of those metrics will overlap, but they won’t always be the same. And to make sure that when you do those things… so I think number one, do that. Figure that out. And that’s, that is easy to say and not easy to do, and I think what happens for a lot of companies is they get stuck in the art of the possible there, and what I really encourage everyone is just find a path. Pick something. Stick with it for a while. If you r- you need to revisit it in six, 12, 18 months, okay, but don’t change every month or every quarter, that’s for sure. In fact, I don’t even know that I like that I said six months. Probably at least give it a year, right? So that’s one piece. And then, and then I think the other piece I would highlight is actually have the intelligence reside on your side of the aisle. So one of the other things that I see happens a lot, and I get why, it makes a ton of sense that this happens, is that marketers end up relying on the metrics from the media companies or their, or their-
[00:26:45] Greg Kihlström: Mm-hmm
[00:26:46] Matt Spiegel: … kinda technology vendor aggregators. and what happens is, is everyone then claims too much credit. And that y- you’re only getting a kind of a, a pinprick view, and you’re making decisions based on, well, what happened within one ecosystem, within one walled garden, within one technology aggregator. And the best case scenario is you actually are extracting it from one channel, and you’re actually measuring across channels. Again, easier said than done, very doable. We, we do it. There are other partners out there that do it as well, but you can do that. And I think the more that you do that, and the more that you begin to wanna internalize that intelligence in a cross-platform way, the more that you’re going to learn, and then the more that you can figure out what to do next.
[00:27:26] Greg Kihlström: Yeah. Yeah. And so, maybe going back a little bit to the, the people and process part of the conversation, I, a- and maybe this is just a, a sign of the maturity of, of the industry. Let’s, like to look at the h- that glass half full [laughs] part of this is, the, the conversation in buying is, is shifting from, which AI tool are we gonna need? I think this sounds familiar from MarTech and AdTech purchases as well, it, it’s shifting from that to how should we organize this team to do the best work and get the best results? So again, I think that’s a sign of progress here. In organizations that are getting this right, what’s actually changed from, from what you’re seeing? You know,
[00:28:11] Greg Kihlström: who owns decisions, where budget sits, all, all of those kinds of things. Like, how, how are organizations thinking about this that are, that are doing it well?
[00:28:20] Matt Spiegel: Yeah. I, I, I think we should acknowledge that mostly organizations are changing versus changed. I, I think you ha- you’ve really gotta dig to find ones that you can look at and go, “Oh my, you have really, truly made game-changing org design change or process changes.” so most are still in the, stairstep, side of this. But what some of the case studies that I’ve seen, case studies, examples, some of the, some of the things I’ve, I’ve heard directly from marketers that really excite me is when you go, “We n- we have rethought our process,” right? “It used to be that when we did a creative brief and went from here to here to here to here to here, and now it’s from here to here, and the first thing that goes is an LLM,” right, “does dot, dot, dot, dot, dot.” Okay? “And we’ve taken these people out of
[00:29:05] Matt Spiegel: the process, or we’ve connected all of these people in the process, but it’s at the same time.” Right? That, that, those to me are, are, are fascinating. Or other companies where I’ve heard where I say, “We have take- We have spent the time to take all these different data sets, and we now have it in one common platform.”
[00:29:19] Greg Kihlström: Hmm. Yeah.
[00:29:20] Matt Spiegel: Okay? Um-
[00:29:22] Greg Kihlström: Yeah
[00:29:22] Matt Spiegel: … and again, just in doing that means you are changing how you operate. So I think the core of your question is, is right, and, and I’m, I’m, I’m where you are now. I wasn’t always, but I am now a believer of every change is positive. I try not to be negative about what hasn’t happened, and instead be positive about the evolution that is happening. And it is happening, but it’s not gonna happen overnight. So I am glass half full, and again, what has worked well is when really in many ways the stuff we’ve talked about is being done, and you don’t feel you have to do it all. but you create a learning agenda, you create some milestones and goals, and you do chase solutions to business problems, not the technology. You figure out how to leverage the technology around it, and you’re willing to say, “You know what? I need less people doing this,
[00:30:07] Matt Spiegel: but I need more people doing that maybe. I need new processes that are streamlined and different, and I’m willing to re-architect.” That’s, that’s where the most interesting change happens and, again, let’s not, let’s not kid ourselves. That isn’t al- often easy to do in big, large companies.
[00:30:23] Greg Kihlström: Yeah. Yeah. Definitely. Well, Matt, thanks so much for joining today. Thanks for sharing all your insights and, and ideas here. two last questions as we wrap up. if we were having this interview one year from today, what is one thing that we would definitely be talking about?
[00:30:39] Matt Spiegel: Man, I’m a, I, I believe we’ll be talking a lot about the same, for better or for worse. I don’t think this is a year journey. But I think we’ll have more case studies, and we’ll, and we’ll start to point to some marketers and some companies that we’re like, “Man, they really changed things,” and it starts to be a, a blueprint for others.
[00:30:57] Greg Kihlström: Yeah. And last question for you, what do you do to stay agile in your role, and how do you find a way to do it consistently?
[00:31:04] Matt Spiegel: Yeah. Well, that second part of the question, I wish I did every- anything consistently, or I guess I do some things consistently. for me, I… Listen, actually conversations like this are relevant and interesting. Conversations with, our com- our customers is always valuable. I do find content at industry events often useful. I try to read a bunch of newsletters. Like I said, try is the key word there. I would say at, at the core, I spend as much time as I can listening, and listening both to answers and questions, I think both give you a lot of insight to what the industry’s thinking about.
