Lucas:
In today's episode I have an interesting guest here with me. I'm with Lewis Rothkopf who is hailing over as the president at Martin. Martin is an interesting demand side platform. Today, we're going to be learning about what this is all about. I had a quick chat with Lewis and he gave me a great example of figuring out-- With Martin, you're figuring out if people actually buy your products because of the ads you're running, or if there are any other drivers for that so that you really understand the efficiency of your ads. Today, we want to be learning about what that actually means, how it works, how folks are benefiting from it, and which types of results drive it. So I'm pleased to have Lewis, the President at Martin DSP here at the show. Welcome to Pathmonk Presents.
Lewis:
Thank you so much, Lucas. It's great to be here with you today.
Lucas:
Yeah. Awesome. Give us the 360 overview. What is Martin all about?
Lewis:
It's interesting. It occurs to me that Martin sort of gets to a point for marketers where they land on the landing page of the marketer site. Then you guys pick up and you take it from there. So there is a very natural and seamless transition, I think, from sort of pre landing on the site to then you managing the path and assisting in the path once they get on the site. Martin's a demand side platform; a DSP. That means we allow marketers of all kinds to come into our platform and buy media across the entire open web. So if you think about just about any open web property that has video, display, connected television, OTT, mobile native, we're able to help marketers get their ads running on those properties. We do things a bit differently than our much larger competitors in many cases.
We believe that the way that marketers have been measuring the success of their campaigns for most of the past, 10, 12, 13 years of programmatic, and even going all the way back to the beginning of digital advertising itself has just been wrong. So there are old ways of measuring the campaign's effectiveness. Things like click through rate, which I'm sure as your listeners know is a terrible metric to understand how effective a campaign was at converting users. Things like cost per action, which is better than click through rate, but still not great. The reason I say that is because none of these methodologies are able to point out the causal relationship between the advertising and the effect on the marketer's business. So really understanding how many people came to my site and bought the product because they saw the Ad, versus how many people came to my site, happened to see the Ad, but are going to come to the site and buy the thing and convert anyhow. So our focus is on measuring lift, otherwise known as incrementality. It's exactly that. What has been the incremental impact of a marketer's campaign on their business' bottom line?
Lucas:
Can you explain us a little bit more about that causality? What are you guys looking at? What is it that you are telling marketers? What can they take out from there? Some people start to get very interested knowing the concept but wonder what does it mean? What will I get out of Martin that will help me to exactly describe that causality or prove that this is causality for myself?
Lewis:
Great question. So you're Lucas and you're at home and you're like, "Hey, I would really love to get a new pair of sneakers." So you say, "I really don't know what sort of brand I'm going to get. So I'm just sort of open. I'm seeing all the different advertisements for sneaker brands. I'm seeing them mentioned in social, but I haven't really made a decision yet." It just so happened that you had an Ad that was served for sneakers and those Ads were paid for by the Lewis sneaker running company. You're like, "Well, this is really interesting. I think I'm going to go get some sneakers from the Lewis sneaker running company." So maybe they click on the banner or maybe they go to a physical location, or maybe they just not click on the banner but still go to the Lewis running company site and they purchase the sneakers. Now, that's case number one.
Case number two is you are Bob. You're like, “I'm going to go get me a pair of Lewis running company sneakers.” You happen to see the same Ad that Lucas saw for those sneakers but you had already made up your mind. You were going to buy those sneakers. You were going to buy them from the Lewis running company regardless. So now you have a situation as a marketer where you know how much you paid for your advertising, you know how many people converted, but what you don't know is how many people converted before they saw the advertising. So we do something around measuring lift. We can measure lift in site visits. We can measure lift in on site purchases, in store purchases. Really segregating the audience between those who saw the Ad and took the 'A' action because of it from those who saw the Ad and took the action they were going to take regardless.
How do we do that? We use a variety of techniques to measure incrementality. One of them is called ghost bids. There is some really deep information on the web around ghost bids, including on our site, martin.ai. The long and the short of it is you have two groups of users segregated by AI in the platform. One group of users is considered the control group. They see the Ad. The other user is shown everything leading up to the Ad. So they're doing their internet session. They're browsing. They're learning more about all these different products. And now comes the time when the Lewis running company is going to show its Ad, but it doesn't. It stops just short of showing the Ad. All of the targeting is the same. All the sort of attributes about the user on an anonymous basis are the same except they just don't show the Ad at the final moment.
The reason this is preferred over older measurement techniques is that it removes much of the bias inherent to things like running PSA Ads. When you run a PSA Ad for your control group, you introduce the bias of, "Well now some other marketer can't run their Ad to the same users that the Lewis running company were going to display it to because now they're seeing a PSA." People who click on PSAs probably have a different mindset than people who click on paid Ads. So your two groups in the PSA context are really quite different when you get down to it. You're also not able to control things like other media exposures that the user had. So ghost bids allows you to draw that distinction between those who saw the Ad and acted based upon it. And then that's where you guys really pick up. So they act upon it, they go to the site, and then you make sure that their journey through the website and purchase process is a good one and one that's ultimately valuable to the marketer.
Lucas:
Super interesting. So I take from what you're saying that in the end the goal is also to have not to spend for people that would've bought otherwise?
Lewis:
You got it. One of the interesting things is if you're measuring based upon cost per action, you want to run as little advertising as possible until you get somebody to convert. So let's say you want to run with a frequency cap of three Ads per user over the life of the campaign. That seems to make sense on its surface because now you're paying less on a per user basis and you're seeing that your CPA, your cost per action, is within the range that you were targeting based upon all the CPA campaigns that you ran in the past. On the flip side, when you run a campaign that is optimized towards lift; towards incrementality, you actually find that the more times that you hit that user with that Ad, the more likely they are to convert. So in the lift universe, we find that anywhere between 10 and 13 exposures to the Ad is what helps drive those conversions among those who do.
Now you might be saying, “Well, that's crazy. Of course I want to spend less. So I want to show the Ad fewer times and doesn't it make more sense for me to pay less and get those action.” It doesn't, because remember in a CPA model you don't understand how many people saw the Ad and acted based upon it, versus those who saw it and happened to act regardless. And so now you're able to understand, “Well, I could pay a $10 CPM and generate five conversions. So yeah, that makes sense in the CPA model,” which has all the weaknesses we just described. “Or I could pay a $20 CPM and achieve a hundred conversions. What would you as a marketer like to do?” If you are sort of looking at the universe in a vacuum and you're saying, “My QBR, my goals are to achieve a CPA of ‘X’ or lower,” you can do that. And you wind up optimizing to weird things like session depth and other attributes of the Ad opportunity that don't really matter to your sales, or you could optimize to what actually drives your business forward as a marketer. This is not our opinion. If we have a marketer that comes to us and says, “We want to optimize on cost per action.” Sure, we're happy to do that. But we're going to spend the whole time trying to convince them that they really need to transform their business towards a lift based metric. Otherwise, they're doing themselves the disservice of not really knowing what worked, where, and why.
Lucas:
Super interesting. This sounds like it's also a big mind shift. A big shift in paradigm, a big shift in thinking for people to fully grasp and see the value that have been acting on Ads otherwise prior.
Lewis:
It has to be. That's right.
Lucas:
Very cool. So what types of companies benefit the most? Which industries, which types of business do you see sort of-- My first thought would be somewhere in the eCommerce space, but maybe tell us where does the tip?
Lewis:
Yeah, absolutely. So we have eCommerce clients, we have retail clients, and we have CPG; consumer package good clients. You can imagine that their KPIs and what they're looking to drive are different, right? So for our retail clients, they are generally looking to drive foot traffic to their location. For our CPG clients, in many cases they're looking to drive lift in their brand awareness, lift in brand consideration among consumers. For our eCommerce clients, you would imagine what the answer is there. It's folks that are looking to drive visits to their website; quality visits to their website. So not low quality clickers. And they're looking to drive onsite conversions. We do really well among eCommerce clients because we're able to measure lift on an always on basis, and then optimize while the campaign is still in flight to lift; to incrementality.
Way back when, you were really only able to see a lift report after the campaign had already ended, by which time it's too late to do much about it. And then maybe you can apply those learnings to your next campaign. It's too late. Your customers are operating on internet time. So in order to be able to achieve the benefits of measuring this, you have to be able to do it in near real time. That’s why we've baked incrementality and measurement right into the platform.
Lucas:
Cool. So how do these marketers, for example, and eCommerce companies hear about you guys? What is sort of a typical client acquisition channels that you guys leverage?
Lewis:
So we are a new company. We're about three years old. We compete with behemoths in the demand side platform space like Google. Driving awareness is the name of the game. We began our business several years ago focusing on one agency that has eCommerce, retail, and CPG clients. We learned a ton by sitting with them and really understanding on a day to day basis, what are the pain points that agency media buyers experience with no existing solutions? What are the things that their clients are really looking for? What are their eCommerce clients looking to do with a demand side platform and ultimately get the kind of results that they are looking to achieve? So all of our business has been outbound.
In the past year we have reached out to the market, we've brought on additional clients, we've activated additional campaigns, but we haven't really spent any money on marketing. And so you can imagine the impact of that is it's all been hand to hand combat, right? Reaching out to folks and saying, “You may not know you have this problem in distinguishing between conversions and causal conversions. But you do, and we'd love to help.” We also believe in the power of marketing. We sort of have to. So 2022 is a year in which we're going to ramp up our outbound efforts. We don't want to be a world in which the first time that somebody hears about Martin is when our salesperson picks up the phone and calls them. We want to begin to develop our brand and introduce our brand to the space so that when customers like yours think about, “How do I begin the purchase journey on the open web and get qualified leads to my site?” they'll think of us.
Lucas:
That makes sense. There's one thing that I wanted to-- You mentioned actually Google there. There's one thing that is coming up recently in the show quite a bit. It is the disappearance of third party cookies by end of the year, beginning of next year, roundabout. Is that something that you guys talk about already? Does that play a role for Martin? How does Martin think about that? Maybe give us an overview on how you think about it.
Lewis:
We can't wait for that to happen. We are literally management goals of being the first cookieless DSP this year. It's a stretch goal. I think we can do it. The reason we're wanting that to happen so urgently is that first and foremost it's just better for consumers. Cookies wound up playing out in the Ad driven internet in a way that was very far removed from how cookies were developed all those decades ago and what they were intended to do. The cookie was never meant to be a persistent advertising identifier. So from a consumer protection standpoint, from a standpoint of being able to reach users who don't have cookies and are therefore otherwise untargetable, well, that's not acceptable. And so that has to get fixed because you're just ignoring an entire customer segment. So we're very excited for it to happen.
We were very disappointed when Google, about five or six months ago announced they would not be making the Chrome changes as early as they had previously announced. Disappointed because it seemed as though advertisers who were very interested and concerned and making contingency plans for executing campaigns absent to the third party cookie, kind of let their foot off the gas a little bit pending Google reannouncing a new date. We think that's a mistake. And so we're working with our clients on the one hand, our suppliers on the other hand; the people who generate inventory for us to run Ads against, then also our identity resolution partners. In really trying to drive forward, we need to get there and we need to get there as an industry now.
That means publishers and supply side platforms passing unique identifiers that are privacy compliant and that are not cookie based, and that are truly anonymous or pseudonymous. It means advertisers coming to the microphone and saying, “We're doing this now. It is the right way to do it. We are able to reach users that we were not otherwise able to do in a truly anonymous and privacy centric way, or pseudonymous and privacy centric way. So we're ready, we're eager, and we're dragging people along. And we think that before long everybody will catch on.”
Lucas:
Frankly, yeah. We had a lot of marketers on the show and actually what stands out is a lot of them are sort of lightly touching on the topic. It's not fully sinking in, not necessarily-- obviously in the eCommerce space more than in others. What would you say are the key things that people should be starting to do to be ready, in your perspective?
Lewis:
Just get started. Go to your DSP, go to us, or go to one of our competitors and say, “I want to run a cookieless campaign. Are you able to do that?” The two key things in being able to do that are having a supply side that has these unique identifiers. One of them is the Lotame Panorama ID for instance. We're in a partnership with a company called Tapad which is owned by Experian. They have a product that we're piloting called switchboard. Switchboard helps make sure that regardless of which identity provider our customers want to use, the advertiser wants to use, we are able to deliver for them against that identifier across a pretty broad pool of supply.
The biggest impediment at this point is getting publishers and getting their supply side platforms to include their unique identifiers in what's called the bid opportunity, the Ad call, so that we can see it and then act upon it. This is a business that's driven by demand as our most. So the only way that the supply side is going to get on board is with the demand side mandating it. And so we want our customers to link arms with us and say, “Supply side, we know that you're excited about this. Now let's go do it. And let's run a campaign.” Critically, here are dollars that we're going to run against it that only want to spend against this type of supply.
Lucas:
Very last question on that before us sort of slowly coming to the end of the interview. Super exciting topic. What would you say, beccause marketers are probably also trying to figure out when is the right time for me to…? But you mentioned the demand side. When do you see the demand to be spiking? When do you see marketers needing to-- I know obviously right now, but a lot of them will also pull that off a little bit into the future. When do you see demand side spiking for that?
Lewis:
Hopefully soon. I was taken by surprise that Google pushed out the date. That just might be naiveté on my part. I didn't expect that. We were excited because this is something that should be happening right now. If I had to guess, I would think there's going to be a drum beat over the next, probably six to nine months of advertisers requiring that their campaigns be executed in this cookieless fashion. And then once the first one or two really big ones come to the table and the trade is right about how it's the first cookieless campaign and this big brand was the one to pioneer it, everyone else is going to follow. So what do they do? They just need to start. They just need to do it. They just need to go to their DSP and say, “Let's do this. We're ready.”
Lucas:
Cool. Awesome. I want to wrap it up with some rapid fire questions since we're slowly running out of time. Are you ready for those?
Lewis:
Yeah. Lightning round. Let's do it.
Lucas:
What's the last book you read or the last podcast you listened to?
Lewis:
The last podcast I listened to was yours. The most recent episode. The last book I read was “Project Hail Mary,” written by the same author as “The Martian.” It was so good. Really phenomenal for sci-fi fans.
Lucas:
What's the one thing your company is focused on the most right now?
Lewis:
Incrementality.
Lucas:
If there would be no boundaries in technology, everything is possible, what's the one thing you would fix for your role today?
Lewis:
I'd love for us to be present in everyone's mind on the buy side with a snap of my finger. Magic does not exist in the world so we're doing the work and we're competing with behemoths. But being present in mind would be a nice trick.
Lucas:
And the very last question would be, if today would be your first day joining Martin, what's the one thing you would give yourself as an advice?
Lewis:
Startups are a sign wave. You have really big ups and you have challenges. You sort of have to remind yourself that what's cool about a sign wave is that it is always going to go back up. So don't get discouraged. Remember that the lows are temporary and the really high highs are temporary too. So look to sort of flatten that sign curve and build towards stability in the ups and downs because that's what makes startups great. And it's what I'm sure makes your customers great.
Lucas:
Nice. I really appreciate you taking the time for the Pathmonk Presents podcast. I want to give you the very last word. If somebody would be forgetting all that we discussed about Martin today, what's the one thing that they should be remembering?
Lewis:
Email me, hello@martin.ai. Know how important it is to understand is your advertising working, where, and why.
Lucas:
Thanks a lot for being a guest on Pathmonk Presents.
Lewis:
Pleasure. Thank you so much for having me.
