In this session Kristy Barger, Sr. Programmatic Demand Lead, MoPub, and Jonathan Lau, Director of UA, PLAYSTUDIOS, talk shop about the what it takes to get the most out of programmatic—even if you’re new to it. Jonathan shares learnings and makes a strong case for companies to dedicate a resource to programmatic. He also tells us why marketers need to speak with other marketers and how sharing ensure everyone can improve the efficiency of the buy.
Kristy: So what I actually do on a day-to-day basis is work on strategic partnerships for the MoPub Marketer Program. And that program provides thought leadership and education around mobile programmatic to the marketer community. So today, we are going to talk about programmatic for app marketers. And then we'll have a fireside chat here with Jonathan in just a few minutes. So can programmatic help me grow my app business? Absolutely, it can help you grow your app business. And we see successful campaigns with over thousands of advertisers on MoPub today.
Today, we're gonna give you three ways in how it can help you grow your business. First, developing a performance strategy. Second, getting on board without sacrificing quality. And last, transparency. So the main takeaway with developing a strategy here is really having one that is rooted in data. And let me tell you why or how. First, we typically see performance marketers develop LTV models. And so these models are a combination of data points and prediction that are looking at which user would generate the longest term investment for that marketer. So once you have determined what these models are, you're gonna wanna identify the channels in which you are gonna generate those users in. Most of the marketers that I work with work from anywhere from 10 to 20 different acquisition channels across search, social and programmatic.
So now that you've identified what those channels are, you're gonna want to do analysis and optimization where you're going to rank against the LTV models that you have created. So it's very important that you're continuing to collect data through this process, not for just performance and ROI, but also to be able to optimize and iterate on your strategy to make it successful. So let me walk through the graphs here and then we'll talk this through. So the top bars are the number of installs that are happening per channel. And what the channels represent on the bottom with the four different sources there, search, browse, app referral, and web referral. So a search, which has the largest proportion here of generating new installs, this makes sense. So if you think about what search is, someone's going to the App Store, typing parameters into that search bar, looking for a very specific app to download on their device. So it totally makes sense that it's the highest number of new installs here.
Next, browser and app referral, these generate the second highest. And that makes sense too you if you think about how content is aggregated and how they're looking for information. So whether that be top apps in the store, top gaming apps, whatever it is, that's aggregated, and it makes it easy for the user to find that. And last web referral. So if you think about, like, desktop ads. These, I mean...the user might not necessarily be in an app mindset. So it makes sense that these are generating lower levels of installs. So it's important not to think just about the quantity but also the quality and where you're gonna be able to find those users and where you're spending your time, like, within apps. And if we look at media buying for performance needs, and chances are most of you are gonna use some media buying for your performance campaigns, this chart, so if you look at the top, the black lines, those represent the different areas and what you're going to want to measure or rank your different supply sources or channels. And the three different blue lines are the different examples of channels that you can buy from.
So if we take first and look at the first blue line that's direct buying, like through a publisher, a one-to-one relationship. So this definitely provides the quality that a marketer would be looking for, right, because it's a one-to-one relationship with that publisher. And it has...you know, the user experience is really good because of that one-to-one. But it doesn't necessarily offer the scale that a marketer might be looking for, which is why ad networks, the next one, provides the scale and efficiency for a marketer to buy from. And there are some levels of control like around contextual. But it doesn't necessarily offer, like, the data insights that you might need to be able to back out into those LTV models. And then lastly, on here is real-time bidding. So real-time bidding really emerged to satisfy both the scale and the quality issue with providing you kind of a one-to-many relationship here. It gives you the data insights that you're looking for because you're able to snipe out those impressions that are the most valuable to you that have those data insights in there. And then it also offers just a great way like efficiency and that user experience where you can whitelist, blacklist in real time and really be able to control your inventory sources that way.
So just to summarize the three areas that we just talked about, first and foremost, developing a performance strategy that is rooted in data, very important. Second, it's quantity does not equal quality, right? And so thinking about where to find those users and where they're spending their time, like in an app. And lastly, real-time bidding provides you both the quality and the scale that you're looking for with the efficiency to buy. So now we're gonna talk to Jonathan who has done all three of these things and has him talk us through a little bit about his programmatic strategy. Jonathan, do you wanna tell us a little bit about your role at PLAYSTUDIOS?
Jonathan: Sure. Thank you very much, Kristy, for the invite. My name is Jonathan Lau. I'm the director of user acquisition at PLAYSTUDIOS. We are a close to an eight-year-old social casino, app developer. based here in Burlingame, California. One thing I have to highlight is that I am not a subject matter expert on programmatic. I am one of many people that are really trying to learn and understand what is programmatic? How can we apply this to our overall strategy? And yeah, I'm just here to have a conversation.
Kristy: Yeah, I think it would be helpful for the audience to hear how you define programmatic.
Jonathan: I guess the best way I can define it is the way it was kind of described to me when I first started. Imagine, you know, you're a parent, and you're trying to search for a prom dress for your high school daughter. And she has very, very specific colors she wants, vice versa, but the cut or size, and there are 1,000 different dress stores out there, right? You can either go to individual dress stores one by one to search for the dress you want or you can have a robot do it for you. You feed your KPIs into the robot, and the robot can canvass all 1,000 stores, look at their entire inventory, and pick out the exact dress that you want. That is the most basic in layman's terms what I would see described as programmatic.
Kristy: I love that. I spent a lot of time in dress stores looking for prom dresses. So I wish there was a robot that could have picked one out for me.
Jonathan: I haven't, so yeah.
Kristy: You might say no. So you've been doing programmatic for, what, about a year now?
Jonathan: Yeah, close to it. Yeah.
Kristy: What's that experience been like for you?
Jonathan: It's been interesting. When we first started playing around with budgets on programmatic, our initial thought was not do what everyone else is doing, right? So we would go to companies like Liftoff, Manage, Arche, these DSPs and programmatic networks to really engage and try to tap into whatever the programmatic is at the time. And at the time I admit, I wasn't quite sure. And, to me, over the course of the last year, the feeling is like getting on a cab, and you're riding a cab, but you don't know how to drive the car. And over time, we've come to realize as a team that we really want to understand how the car works and how we can drive the car ourselves, right? And that's kind of evolved over time.
Kristy: Can you tell me about, like, a campaign that you might have launched with programmatic or one that might have been successful for you?
Jonathan: Yeah, I can tell you that when you were originally started buying programmatic, it took a lot longer to ramp up than we initially expected, right? The way we originally evaluated is that here is a potential inventory source that could rival what we're doing on a Facebook or on a Google, for example. And there's this incredibly long testing process to find the right iteration of a buying model dependent on, a bid model, based on what you're trying to do and the right creatives and the right size. So there's definitely this initial ramp up and testing and cost associated to it. And for us specifically, I knew at the time...this actually...it's actually a little bit more than a year ago, but we realized we were one of the first social casinos to actively spend money in the programmatic space at the time. And it took a long time for the models and algorithms to really understand what we were looking for. And that made it suitable for us. And what we found over the past year is that as more and more of our competitors go into the space, we are actually able to find a lot more success based on cumulative data across all the different social casino advertisers.
Kristy: So I do work with some marketers that are new to programmatic as well. And it's definitely a longer ramp-up period, right? For that time, against like ad networks, that there might be performance within two to three days. How did you get your team on board to be patient for that period? Like, was it selling internally to those people? I mean, or how did you set expectations around that time frame?
Jonathan: It was kind of difficult. So technically, when we started, this was probably like very, very early for us. It was about two years ago, or two and a half years ago, when I was actively running the campaign's myself and really sitting there. And, you know, for you A managers, we sit there and we look at the metrics on a daily basis. And you're very much put into a position where you're motivated to shut it off. But I think what we understood was the potential behind such a demand source, right? You're tapping into multiple exchanges that is tapped into thousands, if not hundreds of thousands, of publishers in the space. So for us, we knew that if we can give it a much longer leash, we could potentially find something fruitful. And for us, that journey took about a year before we actually found something fruitful.
Kristy: That's great. I'm glad you stuck it out. Is there a certain format that performs better for you?
Jonathan: I would say that, from what I've seen and from the data that we've collected, even within the programmatic space, video still tends to do really, really well, which is interesting, right, because we have direct relationships with video ad networks. And oftentimes, there's always this internal debate like, "Are we cannibalizing our own efforts when we go direct versus going programmatic? And how can we reconcile these two efforts?" And the way we look at it is, why does it matter, right, because yes, you are bidding on different impressions. But there might be value. Like, you just adjust the dollar associated to that impression differently based on what you're seeing, in terms of KPIs. So we find that it's a little bit more of a canvas approach rather than a, you know, headhunting approach.
Kristy: Got it. So with the adoption of the new, like, a programmatic channel or buy in channel for you, what type of, like, internal resources or new things that you have to allocate, like, for this new buy in channel?
Jonathan: So in the very, very beginning, we treated our programmatic demand source, like, just like any other channel that we work with. So we would have one of our QA managers manage our programmatic partner, whether it be a programmatic network or DSP, along with all the other demand sources that they're working with. What we have come to realize over time is that we should really dedicate a dedicated resource to programmatic, right, given the breadth of the potential there.
So, for example, if we were to dedicate a time to have a dedicated Facebook buyer to do nothing but Facebook, why not try programmatic the same way or why not treat Google the same way, right? So what we're kind of moving towards is moving our people from a generalist role to a more specialist role. And I think in the near future, I can see ourselves having dedicated personnel that is, A, we're buying and working with these programmatic networks in DSP. And in the future, potentially working on a self-serve DSP such as Appreciate or even working with companies like Beeswax so that we can better understand, okay, what are the sources were connected to? What are the ad formats we're working with truly? And be able to collect that data ourselves.
Our biggest gripe when we're working with a lot of these programmatic networks is that we don't get a lot of transparency in terms of the data. We don't really understand are we bidding for a first impression or a fifth impression by ad unit type? We don't really get a lot of visibility in terms of how we're capping the impressions on a daily or lifetime basis. So for us, as kind of UA professionals, we're naturally curious, right? We want to know these things. We want to understand, okay, how can we improve the efficiency of the buy? And can we do it ourselves, right, whether is it, us building technology around it, or hiring people, a dedicated BI analytics or a resource around it? And that's what we're currently exploring.
Kristy: So if you have this, a person that's dedicated on programmatic, and, you know, they're obviously looking at where you're plugged into today and absorbing all that data, I mean, what do you see as the best way for them to get educated or to ramp up on this special channel or maybe they might even be brand new to programmatic themselves?
Jonathan: There's a couple of ways. One of them, of course, is the mobile marketing program, was actually very beneficial for us to kind of open our eyes in terms of how can we get various data from the exchanges themselves, right? Oftentimes, we run into situations where we approach our partners and say, "Hey, can you give us this information?" And then there's always is like, "Why do you need it? You know, trust in our machine learning algorithm." And I hate that term, by the way. I've heard it, like, so many times, but in general, yes, approaching companies like MoPub and understanding you guys can provide a certain level of data, that is very useful in terms of how can I take this data and transform it into how we can do this ourselves.
Second to that is reaching out to our peers, right? I spend a lot of time going out there and meeting my peers in the industry, and learning from them, and engaging with them offline. And I find that to be very helpful, both in the programmatic space and an over user acquisition strategy. I've fallen in the trap where, you know, you don't know what you don't know. And ultimately, what ends up happening is I make decisions based on my own ignorance. And for me, going out there and hearing other people's perspectives is very helpful in educating my own kind of mindset. And I encourage people to go out there to speak with people even outside of your industry, right? So talking to companies. If you're in gaming. I speak with people in e-commerce. I speak to people in ride sharing or even like, kind of like meditation apps, right? It's interesting to understand how they approach the user journey, how they approach user acquisition. Even things that may not be suitable for you, it informs your kind of worldview around how you wanna develop your own strategy.
Now, of course, every now and then you're gonna get that asshole who's gonna be like, "Well, you're gonna have to pay me a consultant fee, if you wanna have a chat." But most of the marketers and most of the marketers I see in this room that I know aren't like that. They're very open. They're great individuals full of knowledge. I can see [inaudible 00:17:00] right there at Draft Kings. He knows a lot. You should call him. But yeah, it's very helpful. And even for my team, I see them go out there and engage with their peers and talk about best practices, right? Us, being a social casino, we have a fierce competition with, like, companies like Huuuge Games, Double Down, Playtika.
But I don't know if you guys notice, we talk a lot on the back channels to understand, you know, how would you approach this? What would you do here? And for us, it's a great way to do sanity checks. So when, in terms of programmatic, for example, we do talk about, like, what partners do you work with? How do you approach in terms of negotiating IOs, fraud, chargebacks or even, like, have you had experience working with self-serve? Are you considering building your own bidder? These are all great conversations to have to help inform whether are we at the place where we should dedicate resources to building our own technology around this? Maybe, maybe not, but it's a good kind of way to think about your roadmap and your future.
Kristy: Yeah, I mean, we definitely at MoPub have heard the need for marketers talk to marketers, which is why the community program was born, you know, this past year. We're just about out of time. So outside of that, you know, direct peer-to-peer connection, any last parting words of advice for someone starting programmatic tomorrow, you know, or next year?
Jonathan: I think a good place to start is to engage in the MoPub Marketing Program to really understand the space. I think everyone will start off working with a DSP or programmatic network as a... That's a great place to start. But over time, I would encourage everyone to strive towards understanding how to drive the car, right? And ultimately, it's not okay as a marketer to sit there and say, like, "Okay, there's a machine learning algorithm that's gonna do my job for me," but for me to understand how it's done so that we can potentially be more efficient with our spend.
Kristy: Great. Thank you so much.
Jonathan: Thank you.
Kristy: It's great.
Jonathan: Thank you.
Kristy: Thank you all.
