Arm Candy CEO & Founder, John Lods, sat down with Jermey Fain, CEO of Cognitive and founder of the Hidden Layers podcast, to discuss the future of advertising.
Welcome to another edition of the Hidden Layers podcast, where we talk about all the exciting ways marketing data and deep learning are colliding. Today we’re honored to have as our guest, John Lods, the CEO and founder of the media intelligence agency arm candy. Over his career, John has assisted the growth and development of some of the most well-known brands in the world today. Arm Candy specifically is focused on maximizing ROI for brands like Goodyear, Equifax, United dairy farmers, and many others. He’s also an avid Dave Matthews fan, and a graduate of Purdue University. Welcome to Hidden Layers, John.
John Lods 00:47
Thank you, Jeremy. Happy to be here.
So, Purdue did all right and the tournament not as far as I’m sure you wanted it to go.
John Lods 00:53
I would say it was a challenging season because we had arguably the most talented roster that we’ve had in my lifetime, you know, going back especially to like the Robbie Hummel era. So, a little disappointed, but happy to be back into the sweet 16. So, can’t complain.
Yeah, great, great. Well, this tournament was a little bit crazy all over the board. So, Alright, anyway, let’s get into it. So, talk a little bit about Arm Candy. What I’d love to know is why you started it and what is Arm Candy as an agency?
John Lods 01:27
Yeah, absolutely. So, we’re a full-service media intelligence agency and I’ll break that down a little bit on the full-service side. That means we execute both traditional and digital media that can be TV, radio out of home, print, programmatic, social and search and then on the intelligence side, we’ve been building our own data infrastructure, primarily aggregated campaign learnings into a data warehouse that our team analyzes, to help forecast and predict future outcomes for different plans that we’re putting together. So, we have a couple different service offerings one we service agency partners, so think big, Holdco, firms that specialize in creative or web dev, or video production, or anything that’s not media, we help lift up the services that they need on the media side, because it’s hard to be a top tier provider and all those different categories. So, we bolster those efforts, then obviously, we serve clients directly. So, when we were kind of going back, and looking back to 2019, when we started in early January, the opportunity we saw in the marketplace was one, it was really challenging for creative service firms to hire and retain top tier media talent, a lot of times they would try to bring in in teams and almost every director of media, for instance, it’s difficult to determine what their skill sets are, because they typically grew up in a specific type of media, whether it be traditional or digital and I think the demand nowadays was for more holistic planning. So, if we could come bring them, teams that have those skill sets to deliver the media side of the portion of their client needs, their overall revenue will be maximized and decrease their overhead. So, it was a bit of a staffing equation. To just in general, I think we’ve been dissatisfied with what the media agency offering has been bringing to market, it’s very templated, rinse and repeat type of executions and we thought we could build a different type of planning methodology that was more outcomes based versus audience based in simply impression delivery, for instance. So, we’ve kind of been building our own low offering internally and those were the two big advantages and two big opportunities that we had seen when we went to market, and it’s been a hell of a ride the last three years kind of getting this off the ground and seeing that concept succeed.
That’s amazing to hear, especially about the focus on outcomes. We get this feedback all the time at cognitive, we want to they ask us, we need to target this high value audience that we’ve decided that is our audience and we’re like, well, why don’t you just want to target people who are going to buy your product? Why does it have to be a specific cluster of people? I think that the industry is starting to change but we still have a lot of marketers that we talk to that are stuck in this old idea of me hit a demographic and that’s what we’re going to create all of our creative for. So, what are you seeing there, when you talk about outcomes versus audiences?
John Lods 04:36
Yep, and this is exactly why we built the methodology that we do. So, when clients who have been in the marketing game for decades, typically they want to start that planning process with an audience analysis, or they already have preconceived notions on exactly everything rather than says 25 to 34 who do X Y and Z attributes and so when we’re kind of putting together our media brief on what is the goal of this campaign? Is it to drive revenue? Is it to maximize return on ad spend? Is it to drive specific purchases of skews or unique products that you’re selling that you have higher margins? Is it lead generation? What is the goal, we’re walking them through the pros and cons of different ways that we can deploy these advertisements. So, if we do want to be concentrated toward a specific audience segment, obviously, the pros are, you’re talking to the ideal consumer in your mind, and spreading awareness and building opportunities against that audience but you’re actually decreasing your algorithms or machine learning components of these different ad technologies that live in this space today, you’re limiting their ability to optimize, therefore, your cost per result is typically going to increase and so when we are a lot of different briefs that we get, every client wants awareness, but what is the intent of awareness is it to do something, and many times it is to drive sales, but sometimes it is to build more brand equity. So, on the awareness side of things, great, we can build story themed messaging towards that audience, and deploy it and the expected outcome, there obviously isn’t something immediately gratifying, it may not be an immediate purchase or website visit or anything tangible and then on the flip side of the equation, really our goal is to instruct our clients and help work with them. Like why it’s important to allow and lean into the machine learning programs today in more generalized audiences, because frankly, our goal is to drive that outcome and you’re right, Jeremy, we don’t care who purchases the product or takes that action, we just care that those actions are coming in and so it’s just a different shift in how to advertise and who to talk to and I do think advertising to date in general, still skews more towards what is the audience and what is the message that resonates? And that is the case in many of those upper funnels, you know, objectives. But where we really thrive is more, if we’re trying to drive a specific action or outcome, by recommendation typically counters specifically specifying what types of audiences to deploy against.
That’s really interesting. I think that depends on where the marketer comes from, right? If they come from linear, or traditional, it’s how they’ve been forced to choose and forced to plan and then if a person comes up from digital, they don’t have they have the opposite reaction? If it’s not trackable at all, it’s hard for them to figure out how they’re going to accomplish goals and I think what you’re saying is, you’re trying to bring them along, and put them together and say, hey, there’s reasons for both of these and let’s make sure that we’re flexible enough, based on what you’re trying to get out of it.
John Lods 07:45
100% and that’s kind of the fun part of the ecosystem that we live in today, especially those who are on the digital side, they’re like, why would anyone run traditional TV, we can’t target X, Y, and Z audience and there is no attribution that can really be at least one to one attribution that can be certainly measurable and there are so many opportunities at the traditional space in terms of total unique reach and reach and mass awareness, and especially on like direct response TV and kids, you think infomercial, like television, some of the inventory, you’re able to procure at the rates, justify those types of investments for specific types of product sets and so those types of campaigns, so to a digital person would go like, oh, my gosh, I cannot believe we’re running this, the actual outcomes and the overall sales results from those can be even better than those in like a Facebook, Instagram, Google ads and programmatic fashion. But they just hadn’t had the exposure to running those campaigns to really see the data that have much belief in them and the counter that similar to those who have grown up only executing traditional TV or, or managing radio campaigns, they still don’t totally understand programmatic and some of these more digital campaigns, and the data can kind of be daunting and so I think that is where it’s so important to simply educate both sides of those equations to see the pros and cons of these different campaign formats, because it’s really cool and now they can work together and nowadays with ACR and other types of ad technologies, you know, the digital and traditional worlds, especially in the world of attribution, and they’re blending so much. So, it’s just kind of a fun time to be in advertising, frankly.
Yeah, maybe to help our listeners understand exactly what you’re talking about. Do you have an example with one of your clients, you know, they can remain anonymous or not, but I’d love to hear a little bit of an example of how you’re bringing a client along or they’re experimenting with new trackable, like CTV versus linear or vice versa?
John Lods 09:50
Yeah, absolutely. So, much of what we do, especially when we are executing a campaign so one that I’m thinking of at the moment is in the mortgage space and lending space and historically, they had run primarily brand awareness campaigns in hopes that that would drive lead generation. So, the original brief that we got was primarily based on national TV exposure, executing expensive spots that were really well done by our agency partner in communicating that story and we see this all the time, no matter how many times we asked, like, do we want awareness or lead generation? They’re like, oh, no, we definitely want awareness, but then you execute that campaign, and about a month or six weeks in, but like, okay, where are the leads? Like was that the actual goal? You know, it’s not particularly easy to get your client to say exactly what the outcome or KPI they want is. So, when you’re running these TV programs, the types of methodologies you can use, and one of the attribution partners that we implemented is called TV squared. So, if you haven’t heard anyone who’s listening, TV squared is a fantastic overall modeling equation and it shows you top performing created by Daypart, top performing programs at a market level, and it actually allows you to optimize traditional and especially DRTV type inventory in a biddable way, you could argue direct response TV was the first biddable actual type of media, because you’re seeing the actual airings and ratings relatively in real time, and you’re optimizing toward different programs and day parts that are connecting with that audience. So, you fast forward to now the programmatic side and it’s almost like real time and you’re exchanging on the stock market practically, and you’re buying impressions and different audiences in the exact moment that’s happening. But going back to kind of the evolution of this campaign, it totally started from that awareness standpoint, that attribution cycle was helpful, but it wasn’t exactly what the client wanted. So, we pivoted mid campaign to 100% lead generation, set up a landing page with the partner got all the tags and attribution layered and when you’re executing in the digital ecosystem, the perks are, you can actually see from generally one to one environment, when someone is taking an action on your ads or not post view and actually seeing those results. The challenge that remains is based on your attribution model, you have linear first click or last click or a time decay. So, it changes the story or the overall effectiveness being reported by these web analytics platforms, and specific media platforms based on the attribution model and I think now there’s just so much optionality, it’s difficult to know which attribution model works for you and to really understand, you know, connected TV, for instance, never gets the light of day on last click, because guess what, you can’t really click on a connected TV ad. So, it’s never going to get the attribution that it wants. But there are ways to help bring that story together and give it some of the benefits. When you’re doing that. It’s just, there’s no one size fits all way to do it. Therefore, it can be a little bit challenging and working with clients. But that’s also the fun part of the opportunity.
Now let’s talk a little bit more about that attribution challenge, because we run into it a decent amount with our clients and we have to turn away business a lot of the time when they say that they’re last click, then we say, look, machine learning algorithms, they can play that game but that’s not really how you’re driving your conversion, right. To get to that click, you have to get a lot further, you have to do a lot more than talk about what ad was there that they actually ended up clicking on and we love multi touch attribution, of course, but very few marketers are using it. Most of our clients are using last touch attribution, which is okay, it’s still not great from how do you figure out your media mix but where are you seeing that trend with your clients?
John Lods 13:53
Yep, I would say definitely in the startup in tech side, the adoption and multi touch attribution, it’s much greater than that of any clients that we’re servicing that are fortune 500, or above or have been 50, the 100 year runs within their organization and it’s just because the machines that they’ve set up their overall infrastructure has been running in such a certain way, their analytics teams may be 20-30 deep and so making any type of change in the attribution system is very difficult. multi-touch attribution didn’t even exist, when many of these teams were fully built out and running and so I think that’s been the challenge. Any company like Tesla, for instance, that has been started relatively soon, and it has really started to enhance their marketing, I think they have the advantage because they’ve adopted attribution models from the get-go. Most of these attribution models the entire goal is to have a benchmark, you know, or a bench line and so at some point when they do flip from a last click attribute Should model to multi touch that year or two year period as they’re accruing new data in different ways is just going to be a nightmare for them and I think that panic can cause some nervousness on their side to commit to that with GA4 being mandated by summer of 2023. I think that’s probably the time, my guess, is because so many people use Google and Universal Analytics, when they have to adopt a new web analytics platform anyway, whether that’s Adobe, where a lot of people are moving to, or the new Google Analytics instance and GA4 I think that’s the moment that switch is going to occur and as an agency partner, and immediate partner, we are almost incentivized to be a part of that change for our clients, to make sure that all future efforts, whether it’s media run through us or someone else, but set up for success long term. So, it’ll be interesting to see how agencies take advantage of that over the next year while that change has to take place.
Yeah, let’s talk about that a little bit, too, because almost every RFP were, or every pitch we’re talking to agencies with, we’re now being asked, finally, what’s our cookie less plan? So, I’d love to hear from your perspective, especially as a media intelligence agency, what is your cookie less plan?
John Lods 16:19
Yep. That’s a great question and honestly, I don’t know if we have the solution figured out because we’re still waiting on these different partners, whether it’s the trade desk or an Adelphic or Google, who are all kind of sprinting in their own direction to adopt it.
What’s your gut, say? What’s your gut, you’ve been in the industry a long time, what’s your gut say is going to be the reality?
John Lods 16:43
my gut tells me that there’s going to be a little bit of fragmentation on how this is going to get executed, implemented. I don’t know if we’re going to adopt a universal cookie less solution that’s effective and useful for all the different ad technologies that are there today, whether in different categories Social Search, what have you, I think the reliance on device graphs is going to increase meaning if a client can adopt a specific device graph provider that plugs into the different cookie less solutions that these different channels can execute out of, I think the device graph and the connectivity to how that’s being articulated in their attribution framework is going to be the most important. So, I think all of that gets executed easier as a client’s access to first party data is greater. So, device graphs, and first party data, obviously, hand in hand. So, anything CPG or consumer facing relatively seamless, I think any types of businesses where they’re more behind the scenes or don’t have direct access to a lot of first party data, I think they’re just going to be a bit of a challenge. So, I’m expecting turbulence, to be honest, you know, it’s going to be interesting and we’ve, even though this Google and everyone else has kind of been pushing back the date, because this is a harder problem to solve than we anticipated and then you sprinkle in Apple, definitely leading the charge on consumer privacy, and restricting access to certain types of data, which is not easy when devices start restricting access, that’s only going to add complications. So, I think we may actually, as an advertising community, take a few steps back into what we can attribute in how we can attribute it. We’re about to start modeling data for everything again and you’re kind of seeing that with Facebook’s initial reaction to Apple, and how they’re attributing conversions nowadays to
John Lods 19:30
Well, here’s a fun way to look at it that we’ve been seeing with much of our larger client partners that we work with. Legal teams are instructing the CMOS, many marketing departments to avoid touching data at all costs. We’ve seen so many data breaches over the course of the last 24 months with massive companies with a lot of data and security protection and hackers and other bodies are able to still kind of breakthrough to that and I think the concern of managing data is going to make any marketer leveraging device graphs really hard, especially with legal counsel continually to press on marketing teams to avoid touching it at all costs. So, it makes adoption or new age technology challenging when different teams within the sink organization has varying points of view on whether or not this is worth it and I think that’s a hard choice for a CEO, or anyone to make as is the liability and the security protocols required to manage this data worth the investment for marketing to be more successful or do we just not want to touch it at all? And so, I don’t know that’s going to be a different answer per company. But it’s certainly something that is going to be asked and discussed moving forward.
That’s a great way of looking at it. Let’s circle back to a comment you made earlier about machine learning and not focusing on a cluster of a demographic, etc. Tell us a little bit about how Arm Candy is looking at ml. I know you guys are growing really quickly and I just got back from a digit day conference and the main concern, which has been a concern for a while, but now it’s with a great resignation and all these things, it’s even more exacerbated, hanging on to talent churn, not having enough talent, not being able to hire enough people, the frustration of managing media campaigns, and how tactical it has become and how hard is become to do easily and quickly. Is automation on your mind? Is machine learning on your mind? How is it fitting into the Arm Candy strategy?
John Lods 21:46
Yeah, absolutely and that’s a big question. So, I’ll probably break it down in a couple of ways. So, let’s start with the great resignation because this has been certainly a big challenge for especially larger agencies, nationally and internationally and as an agency, that agency service provider, it’s I would say we’ve become more productive for our greater agency partners. As a result, because we as an agency, part of our origin story was all about adding value back to our employees. So, and how we do that we’ve fully paid for health care, we’ve set up 401, Ks within our first year of operation, we rolled out mandatory PTO where our team is literally in forcing people to take a minimum 15 days off and all studies suggests for instance, you know, the more time off more refreshed and more productive people are. So, it’s it’s in our best interest to make our team members lives as joyful, and enjoyable as an agency as possible and that was part of why we started arm candy to begin with advertising wasn’t fun anymore, because it was becoming so more, it was just too many hours in too much of a grind and clients weren’t particularly easy to work with. So, how could you change that dynamic. So, by focusing on our employees and our team members, I think we were kind of well positioned for that and we’ve gained a lot and kind of won the talent game. On a smaller level, obviously, we’re 30 people, but it’s helped us service more partners who aren’t as successful today with hiring and retaining their talent. So, that’s kind of one thing, too, a part of our kind of internal organizational structure, where we see larger agencies struggle, in our opinion is they break out their teams in the so many different silos, meaning they have a programmatic team, they have a social team, they have a search team, they have TV, radio, out of home print, they have all these different teams and I think people graduating nowadays have such good skill sets, that they want to diversify the information that they’re gathering on a daily basis, their curiosity is certainly at an all-time high and so by siloing people into specific channels, learning everything there is about social and programmatic and whatnot. I think they’re doing those team members a disservice, because they’re not able to see the full ecosystem in our beliefs. It’s important for people to know the values of TV but also know the values of programmatic and the way that larger media agencies are structured don’t really allow you to do that. So, we broke it out where we have ad ops teams that are fully fluent across those different ad technologies, so they can pop in and execute CTV. They’re obviously Google Ads certified, but they’re also executing and Pinterest and Twitter and Snapchat, and they’re diversifying that experience and I think it’s really helping us because our people are constantly being challenged to learn and apply new info and it’s just helped us so obviously, the agency benefits are good, but our ability to really leverage and develop our people, I think has really set us up for success and naturally allows us to service our clients with a more streamlined team. Instead of having seven different FTAs each with one skill, we can have one FTE that we’ve developed with all seven skills, and it allows us to be more cross functional and in service different media agencies. So, that was our thesis and really, we’ve put it into real life, and it succeeded. So, I’ll stop there and let you pop in before we jump into the machine learning side of things.
Yeah, I know, that also seems to make it a more flexible staffing model to for you so that when you’re more TV heavy, you can handle it when you’re more digital, heavy, you can handle it. But when you’re light on one of those, those folks are still busy or have challenging days.
John Lods 25:48
100% and I will say one of the challenges. So, to counter that you obviously need a really strong training and development programs. So, we have what we call the ACI Academy, we’ve literally been building content, it doesn’t live anywhere yet, but I bet at the future state, there’s a login and a portal on and it goes through how to set up data layers, how the setup, Google Tag Manager, or Tealium are the different tag management solutions that are be using our best practices on running Facebook and Instagram. Facebook and Instagram has tried to advertise that it’s so easy, anyone can do it. But the reality is there’s a gazillion different ways to execute a Facebook and Instagram campaign. So, you need all those details to be right. So, we’ve kind of been building that content to just make our training and development that way. We’re setting up our teams for success when it comes time for them to execute those campaigns for the first time.
Yeah, that’s the cognitive University is launching this year to we’re, especially as we’re growing so quickly, we’ve just needed a better onboarding and training opportunity for our employees, for our team members, because they’re having the same meetings over and over and the folks that are training them have to keep moving fast. So, University is something we’re also investing in here.
John Lods 27:03
That’s awesome. Because if you open that up externally, I can tell you, our team would be interested to learn our ballot too.
That’s a great idea. We should look at that making the content internal and external at the same time. So, machine learning, AI, Arm Candy loves using that.
John Lods 27:24
Yeah, absolutely. So, I think on the automation side, this has been a big part of our just operating model and I will say the more that new platforms are rolling out, it is more challenging to automate your workflows, for instance. So, if you think of like the Instacart, it’s of the world and all the different retailers that have it, because we are also managing shopper campaigns, it’s not easy. So, simply, from a workflow automation standpoint, we built our entire agency around the sauna, and we are integrated with Slack. So, it’s it’s made it easier to execute in all these different platforms. One of the things that arm candy has been developing, and we’re starting to bring to the light and have been talking about externally and publicly is our data warehouse repository. So, all of the campaigns that we have managed historically, through our different API partners, we’ve been connected, and we’re pushing to different databases for us to harvest that data. Today, we are currently using that to forecast outcomes, meaning we can go back and look at the last 40 lead generation campaigns filter by sector industry seasonality close cycles. So, how long does it actually take to close the sale in different variables to help predict what our expected cost per result would be after a fixed budget and so I would say it’s not necessarily leading on machine learning to date. But we’re at least getting the data there, where this application can be applied in the future, our hopeful version, and this is a little bit more future forward is, once you have all those variables kind of managed, and you have enough data behind you, you can create a machine learning program to help you forecast exactly what is the perfect medium mix to maximize the budget that you’re working with. So, if a client comes in today, and says I have $3 million, that’s for annual, here’s my average order value, here’s some of my historical sales data that I’ve ingested into the platform and here’s some additional variables that we need to know like, what types of creative are we currently running? What are things we can and cannot do? We should be able to predict with significant degree of confidence, what is the best channel allocation and overall budget split based on all that historical data, and that’s where we really want to get to right now, we’ve been setting the pipes up to do that. We have not written the algorithms. So, that will be a fun moment for us as an agency and a technology company. But I hope that’s where we’re able to move forward to in next year too.
Great. Well, we’re almost out of time. So, we’d love to get to know John Lods, a little bit more outside of the media agency world. What happens to John when he takes a CEO hat off for the day? I know that only happens like five minutes a day but during those five minutes, who is John Lods besides watching Purdue basketball?
John Lods 30:28
That’s a great question and I’ll be honest, I don’t know if you ever take the hat off. So, some of my favorite hobbies every Sunday morning, for instance, I play golf at one of the local golf courses here. I’ve been a big player, pretty much my whole life. I think my handicaps may be down to a five.
So, Masters is on your mind right now?
John Lods 30:50
The Masters is literally right in front of me. I’m watching the show and Collin Morikawa right now on the TV here in the back left corner. But yes, so definitely playing golf. As you mentioned, I’m a big DNB fan. So, I don’t judge me all those who are listening. I’ve been to 53 shows to date and so I do every now and again. I got about seven to 10 shows here and we just picked different weekends. So, my fiancee sister lives out in Denver. So, we’ll go see her Friday and Saturday and her aunt and uncle are out there and I have family in Charlotte and indeed, so I travel back and I take the family to go see some Dave Matthews. So, between Dave and golf, those are easily the places and spending the most time outside of the office.
Alright, so with Dave, your fiancée, was she a Dave Matthews Fan before you made her be a Dave Matthews Fan?
John Lods 31:43
I love that you asked this, as Dave would have it, her first show ever, her first concert ever just happened to be a Dave Matthews show. So, now if you ask her, I think she’s definitely coming along for the ride because she knows I love it, especially the more that we have shows. So, each year when it’s your schedule comes out, I let her kind of pick the shows she would prefer to join, so it’s hard to push too hard sometimes.
yeah, friend of mine is a big fish fan, and they always play or they have in the past always played New Year’s Eve, like a week of New Year’s Eve concerts in the city in Madison Square Garden and he lives in Maryland and he’ll take the train up after work, go to the show, take the train back in the middle of the night, go to work, and then do it again. Like two or three times.
John Lods 32:36
I understand it. This opening show this year is in Austin. So, on a Wednesday, so naturally, I’m setting up meetings in Austin that day just so I can drive down, go see it and then we’re taking our full agency, and plus ones to the Dave Show 514 got the bus by the bunch of tickets, and we’re just all going together. So, that is about to be an event.
Alright, last question with two parts one can Tiger Woods win? It doesn’t look like it based on first day, but does it matter because his first day was amazing, based on what he just went through?
John Lods 33:16
Another good one. So, Tiger got asked this question in a press conference. I want to sound Wednesday and he was like, yes, I can absolutely win this week. If you look back at historical performances in round one, he was actually this is way above average. So, finishing under par In the I want to say top 20 years. So, he’s actually in the hunt, just for him to make the cut this year, or even play is extraordinary and what he’s done for golf is just one of the most amazing things in all of sports. So, I was getting chills just seeing the amount of people watching him on his practice rounds. So, I’ve been eating it up. Absolutely. I think he’s already done more than we ever imagined by even playing and I’m absolutely wearing my Sunday Tiger read here in a few days.
We’ve talked about ratings boost for the tournament this year. I mean.
John Lods 34:11
Unbelievable. All right. Well, thank you very much. That does it for another edition of Hidden Layers. I’m your host Jeremy Fain, CEO, co-founder of Cognitive. It’s been a pleasure to have John loads with us CEO and founder of Arm Candy. Please listen in to our next episode as it comes out and thank you for listening to this one.
About the podcast: Hidden Layers
Hidden Layers, the new podcast from ad industry veteran Jeremy Fain, connects with some of the world’s leading experts in and around the disciplines of deep learning, neural networks, machine learning and data-backed insights to understand the implications of each, how they’re evolving, and intersect with the art and science of our everyday lives. Each episode is intended to gives listeners access to short, informative breakthrough stories and updates on the latest industry developments from the best and brightest experts in their respective fields.