MAU [Talk]

Ep. 002 Ian Simons & Jack Dempsey Southerland III, Facebook eCommerce

November 10, 2020 Ian Simons & Jack Dempsey Southerland III Season 1 Episode 2
MAU [Talk]
Ep. 002 Ian Simons & Jack Dempsey Southerland III, Facebook eCommerce
Chapters
MAU [Talk]
Ep. 002 Ian Simons & Jack Dempsey Southerland III, Facebook eCommerce
Nov 10, 2020 Season 1 Episode 2
Ian Simons & Jack Dempsey Southerland III

In this episode, Ian Simons, Head of Industry for eCommerce at Facebook, and Jack Dempsey Southerland III, Manager of Client Measurement for eCommerce at Facebook discuss what advertisers – from the most sophisticated to those just getting started – need to know about value-based marketing and discovery commerce. Stay tuned for more information on Facebook’s value playbook debuting in the upcoming weeks! 

Show Notes Transcript

In this episode, Ian Simons, Head of Industry for eCommerce at Facebook, and Jack Dempsey Southerland III, Manager of Client Measurement for eCommerce at Facebook discuss what advertisers – from the most sophisticated to those just getting started – need to know about value-based marketing and discovery commerce. Stay tuned for more information on Facebook’s value playbook debuting in the upcoming weeks! 

MAU[Talk]:

Hey guys, welcome to, MAU [Talk], a new podcast from MAU Vegas, the premier mobile acquisition and retention summit. On today's episode, Adam speaks with Ian Simon's head of industry for e commerce at Facebook and Jack Dempsey Southerland the third, manager of client measurment for e commerce at Facebook, where they discuss what advertisers from the most sophisticated to those just getting started. need to know about value based marketing and discovery commerce. Take it away, Adam.

Adam Lovallo:

Okay, Ian, Jack, welcome to the podcast. Before we get into it, it would be awesome. If you guys could do little intros maybe Ian if you could get started.

Ian Simons:

Yeah, awesome. Adam, well, thank you for having us really, really appreciate it. So I lead the e commerce advertising business at Facebook, my background is really seeing this ecosystem from a lot of platforms and even the agency side. So I've worked at Google and their properties like YouTube, I was an early employee at Pinterest, as well as you know, Facebook properties. And then I then I spent some time at Digitas as well, way back in the day. So I've definitely seen this ecosystem and landscape from a different lot of different angles, which has helped me over time and, and really appreciate you having us. So thank you.

Adam Lovallo:

My pleasure. Okay, Jack, your turn.

Jack Dempsey Southerland III:

Alright. So Adam, thank you for having us here. Again, I'm Jack Dempsey Southerland the third, and I lead measurement for e commerce at Facebook. Prior to Facebook, I spent my time on the agency side, later moving over to digital, and ultimately working in programmatic and ad tech. Today, I lead the e commerce team, and we're working with our marketers on helping them to drive better measurement and better marketing. It's great to be here before Ian and I get into it, I just want to make sure we thank the cross functional teams around the globe who've contributed to this research in this work. And we hope it's incremental. And it more importantly adds value to the conversation of your listeners and audience.

Adam Lovallo:

Okay, awesome. So we're going to talk about this quote, unquote, value based marketing concept and this discovery marketing concept that you guys want to introduce. Before we get into that. I gotta ask, from your perspective, since you work across so many clients, so many campaigns, so many advertisers, what are best in class advertisers doing right now? And across the board, creative target, whatever, like, what does it mean to be best in class on the platform today?

Ian Simons:

Yeah, that's a great question. So I have to start off by talking about the recent events in COVID. So COVID has shocked the system in such a big way. We are looking, you know, when you look across all the industries here at Facebook, ecommerce has taken a massive step forward. And we're seeing first time buyers, we're seeing older buyers. And then we're seeing really two sets of kind of clients or companies, businesses that are really ready and have been investing in this D-to-C, mobile first e commerce space, and those that are just so far from being ready. And so there there is this quote from Winston Churchill that I think about often, which is "you never let a good crisis go to waste." And I really believe that whether you're ready, or you're not ready, now is the time to shake it up. So broadly speaking, we see or I see two, two kind of trends that are going on. And we're going to talk a little bit about both of these and go deeper in value based marketing. But the two trends are first, this notion of discovery, commerce, and discovery commerce, really is the idea that machine learning is so good right now that products are finding people versus the other way around of people finding products. In the olden days of Facebook, you'd set up campaigns that were so hyper targeted to specific audiences. And often these campaigns targeted audiences that were, quote, unquote, in market for your products. And this notion is flipped on its head and the best I'm going to unpack, you know, how do you do discovery commerce really well. But the best advertisers understand this notion of demand generation discovery, commerce, they're really leading into that. The second step, and this is where the more sophisticated advertisers are really, testing the waters right now, is this concept of value based marketing. And at its simplest form, value based marketing means that a small number of customers are driving a lot of value for your business. And the businesses that understand this concept that are executing against this concept are growing quickly and they're growing actually profitably. They're not letting marketing expenses get out of control. In fact, they're putting their marketing expenses to work in a really, really smart way. And so I'll give you a really simple example of this. A few years ago as on Instagram, I was not a skier, but I was served an ad for skis. And it was an awesome creative that stopped me in my tracks. And it made me think, you know, I'm spending so much time looking at screens and sitting inside, it's time to get outside, right. And so I actually, I clicked on my phone, and I bought these skis. And, and this particular brand did a really good job over time. And this is the value piece of outfitting me for this hobby. And then even providing me with things like ski tickets, etc, like getting me to the mountain, right the ski experiences. So it wasn't just about introducing me to the brand through the skis, this discovery commerce, it was also about nurturing the relationship with me, and turning me into a high value customer. And that's hard to do in practice, but this brand did very well.

Adam Lovallo:

Okay, um, so what does it mean to be doing that?

Ian Simons:

Totally. And I know, this is a sophisticated audience that's listening to this. And many will say, it'll, I'm totally doing this. And this sounds really basic this notion of discovery, commerce, but getting this right, it's harder, it's harder than you think. And, and I think everybody is on a spectrum of their journey to really mastering this, it breaks into three components. The first component is machine learning. So again, it's no longer restricting your campaigns to really specific demographics, or interest profiles. If not hyper segmenting your audience. Again, think about the skiing example. That's what we're talking about here on machine learning. And it's really leaning into this machine learning. The second is about creative. And you know, when I when I moved to, to Facebook, I really underestimated honestly, the the value of creative and mobile first creative contributes and all our studies 56% of sales lift, so not brand lift, but sales lifts is in this creative. And so what what we see folks doing in this space is doing few things, if they're not iterating, quickly enough, they're using Facebook marketing partners to supplement their own efforts. It's not about big fancy studios anymore. They're even leaning into dynamic ads. So even with smaller catalogs, they're thinking about ways to use dynamic ads effectively, that's a very effective format that we have to customize and personalize, creative. And then the last, and arguably maybe the most important part of this is really thinking about your user experience, and specifically your checkout flow. And when we look at all our research, it shows that right now, in today's day of a day in age, there's something like $235 billion of opportunity cost, which is actually growing year on year that is lost just to do to do bad checkout experiences, confusing, many, many steps. And what does good look like? Well, good looks like, you know, check out in five taps or less. And again, this sounds basic when we think about machine learning and creative and user experience. But those that are really nailing discovery, commerce, they're doing all of all three of these very, very well. And it's it's having a really meaningful impact in terms of their ability to grow their business.

Adam Lovallo:

You know, I'll give you a little anecdote, real real anecdote of getting my wife of President Shopify site, I won't name them, I use their integrated payment option, PayPal, I made a successful PayPal transaction. And then nothing happened. 80 to $89 went out of my account, I wrote them an email, like we have no record of your order. I was like, Okay, well, I think your PayPal integration is broken. And they're like, Okay, well, we'll look at that, I guess. So like, it's, it's um, that's a big one. Like, it's so easy, especially when you work in one of these companies to get kind of disconnected from the actual front end experience that people are seeing because so many devices you're in the Instagram current, like real rapper, browser thing, like this stuff breaks there re regularly See, I think that's an excellent, excellent. Okay. So we understand this discovery commerce premise, and I suspect that much of our audience is very familiar with this, this kind of new performance creative kind of agency articulation, which has popped up a lot in the last couple of years. So the second bit of your kind of macro hypothesis is around value based marketing. The lazy amongst us, myself included, would say, Well, yeah, don't optimize for purchase, optimize for value at the ad set level. Walk away. But I presume that you think there's more to it than that. So what like when you're saying, Yeah, like when you're saying value like, what are we? What? How does that how does that get involved with this discovery commerce notion?

Ian Simons:

Yeah. So I think discovery commerce is it's the basics, it's where we start. The next is this notion of value based marketing. And what I what I would say is, it's more than just an optimization tactic and what we'll be coming out with, in in the weeks to come and then even more formally at Facebook, is this credit grid that's going to look across a number of dimensions, that shows how do you mature your company, but holistically from a marketing perspective, from when you're an earlier stage company and looking at things like traffic, right, or conversion volume, to being really tip of spear, where you're considering the differentials in lifetime value of your customers. And so there really is this arc of sophistication. And this arc of sophistication stretches across three different areas that that we're going to go into in this material, which is first, really stating your business need. And this, this evolves over time, and then ensuring you have the right foundational elements. So it's more than value optimization, it's what signals specifically Are you passing to us? Or what information? Are you passing to the algorithms? What ad tech and integrations do you have with the platform's to pass these signals? What analytical capabilities and insights do you have that then feeds into those signals and that ad tech, and then and then, and then how you operationalize your strategy with the targeting, optimization and bidding decisions that you're making to operationalize the strategy. But it starts with a stated business need. And again, this seems basic, but this evolves over time. And it starts with a startup saying, you know, I want as many conversions as possible. This is, where a lot of advertisers stay is just in this just volume of conversion lands. But then what we see when we look at, potentially the public companies that I work with, or even the most sophisticated public companies that have really driven outsized value, they start to pivot to their goals being you know, I want to maximize revenue, divided by cost or ro ass these, these, again, are the more sophisticated accounts. And then when you see the really sophisticated guys, base and gals, they start looking at how do we find the whales, the really high LTV customers, and then what are the impact of those whales over time on my business, and that is a real differentiator. And we see this transition. As these companies get bigger and bigger and bigger and more sophisticated. They actually have to get more specific about their marketing goals in much more specific and intentional about the signals and how they're structuring everything under the surface.

Adam Lovallo:

Great. All right, jack. Let's hear it. So we, we understand theory. So what do we actually do?

Jack Dempsey Southerland III:

So in terms of setting the foundation, I think he did a really great job of sort of laying out the actual vision of an end to end user experience and how marketers can add value and grow a customer at each stage of that customer lifecycle journey. But when we start with the foundation, there are some times we see marketers who shift a little too far to the right, and try to get to LTV and ROAS. And we need to recognize that obviously, we're operating in the real world, there are a lot of challenges. And there's the day to day business. And so it's more of an evolution versus a revolution. And it's about meeting marketers where they are and going at the speed of the business. So when I set the foundation, when we think about setting the foundation, we want to start with the technical prerequisites, what do we need to have in place in order to do proper measurement, and marketing, focusing on LTV and value? What are the signals? What is the ad tech? What data and analytics Do we need to help us understand, attribute and more for more importantly, assess the performance here. And so once we get the right data in place, it's about setting those realistic expectations. My team's client measurement, marketing science, and we work very closely with Ian and his team and helping clients set the right expectations at each step in the journey or each stage in the journey and identify existing gaps. Where might there be potential barriers and how can we sort of push through those to continue to test learn, deploy and evolve? So when I think about the signals Foundation, because it's so important I want to give you guys some examples of what we're seeing from different marketers who are utilizing signals towards focusing on value. First, when we look at ROAS, marketers that are using the value, assessing the value of sort of add to cart, or purchase, or their certain site behaviors that are indicative of a high or potential low value user, and how can I use that to help drive decisioning and more importantly, improve my performance and reach my target goals when we look at LTV, and obviously, we'll talk more about that later. But our more advanced, and I would say mid clients are what I would call the handoff. And it is about an ecosystem where they're passing predictable lifetime value signals to us via server to server integrations. And that data is being used to help inform targeting, bidding optimization and more importantly, measurement. So it is that what I would refer to as the handoff and the passing, that allows us to really drive that business and focus on the most important metric that the client knows very well. And the last piece would be subscription based companies, they are starting to break down signals and pass us signals across a number of different factors. Looking at seasonal subscribers looking at an annual subscription type of prescription, a segment of users, the segmentation of it allows for a lot of interesting research and analysis that can help us get smarter at detecting where we can drive the most value. So it can become as complex or simple as you would like. But it's about recognizing that you need to have the right foundation in place in order to properly build up this value approach. Without it, not tearing down. You know, at the end of the day, one bad test can set a business back. So we want to be very thoughtful in our approach

Adam Lovallo:

So one follow up question and one anecdote, because I suspect you run into this, I'd say in my day to day 50% of the people that I work with, do not know, the data points that are passed back with the events that have set in our set, like the parameters. So like, there's some value that's getting passed, but like, is it inclusive of shipping costs? I don't know. Is it inclusive? The discount? I don't know. Like, what is that event? I don't know? Like, what is initiate check out in our contact. So I think part of the reason is just like literally knowing what's implemented as a start, and even people that are super sophisticated that I see will be like, Oh, yeah, no, no, the add to cart values associated with add to cart, those are 100 x off. So just don't don't worry about that. I'm like, what? Like, why? What do you, What are you talking about? Okay, there's your anecdote, and then a question. So I mean, you start passing back value, right? Or maybe you start passing back, as you said, a project a PLTV, a projected lifetime value at the end of the day. That's some number, right? Is the implication there that I'm starting to use one value optimization and potentially even like, min ROAS or target ROAS some sort of row as getting like is that where people end up going? Once they have this data being passed back? Typically,

Jack Dempsey Southerland III:

it is, and I think of it from a short term to long term perspective. So obviously, in the short term, is looking more at value conversions. But as you move across the value equation, it's thinking about it from a long term perspective. And it's using more signals that attribute to long term performance, like your men row as like your LTV, and tying that into your optimization and targeting. So the operationalizing piece is probably very critical once the foundation's in place because we have to think about business logic targeting optimization and bidding as Leber's that can help you deliver your goals. Those targeting optimization choices can guide our systems and help you improve your outcomes and really tailor our buying approach based on where you see the most value for each particular users. And from abroad targeting Brock talkies and shipper Facebook, the companies in the future, as Ian said, are leaning on machine learning.

Adam Lovallo:

You know, and I want to add something here, this is going to be my editorial. One thing that I see a lot for our clients is if you pass back revenue, or you pass back a projected lifetime value or a one year value, that's all predicated typically in revenue terms, but not all products have the same cost structures behind them. The margins are different. And so I'm like, Hey, wait a minute, like yeah, you know, our row as is blah, right? But like this stuff, you sell it 90% margins and this stuff has 30% margins, maybe actually you in your particular context, like you want to be passing back both a top line revenue number and like contribution, dollars number. And we actually care more about contribution like, like, that's a nuance that I think is often lost in some categories. It's irrelevant, but in a lot like big e commerce companies, that's a huge, you know, that's a huge determinant when you're thinking in this like, ROAS/efficiency term.

Ian Simons:

I think that is, I think that's absolutely right, you absolutely want to be doing this. And as we break down our testing, and we look at those that are kind of splitting signals or identifying the most important signals, we see really meaningful performance. And so an example of that is a subscription company that has maybe an annual product versus more of a seasonal product, they're splitting that signal and pointing people towards the annual which is driving high LTV. You can think about that from a contribution margin perspective as well, maybe that is the singular focus of what you want optimize for, and you can point people to the highest contribution margin products. So I think that's really, really, really important is this idea of kind of signal segmentation, being really specific about what you're passing and how you're optimizing using those signals.

Adam Lovallo:

Yeah. Okay. So practically speaking, I'm an end customer, I'm on Instagram, I'm on Facebook, like, what does this look like, from a consumers perspective? Like, what how is this maybe different than, you know, conventional campaign approach?

Ian Simons:

Yeah, no, that's a that's a great question. And really important. So you can be the most analytical person, right? And look at all the math and all the signals and how we're splitting signals and how we're optimizing. But at the end of the day, if you are not thinking about your brand, while your customers shop with you, how do we find those high LTV users? What do they look like? What are the characteristics? What do they enjoy? And how do we keep them or the consumer experience? I think you're missing, you're missing out a lot. And I'll give you a really simple example. And this, this happened to me this year. So we have three kids. On Friday nights, our kids love to eat burgers, and milkshakes and watch, watch a movie, right? It's a weekly treat to cap the week. And you know, we used to go and we'll drive we used to pre COVID, drive down the hill and go to Shake Shack or whatever to to get our burger and milkshake. But then during COVID time, we were, I was acquired by two delivery apps, right. And, you know, I'm just a CPA to these guys. But what I cared about was the fact that I have this burger and milkshake night, every single Friday night. And I have three kids. So it's a big family. So when you start layering on all the additional fees, the delivery fees, everything that kind of goes into these delivery apps, it was becoming really expensive. There was one of those apps that offered me a loyalty program for it was very cheap, relatively cheap, and you paid it up front, you pay it every month, but it erases all the fees. And economically, they showed me the math of, you know, member non member, and I joined, I joined this program. And that sucks it like a Netflix subscription out of my account every single month. But I am now a high LTV user, or a high value user for this company. And this other company totally lost me because they didn't think about what was the offer that that I cared about, and that they the consumer experience what you're selling and your unique value or your superpower or whatever it is. It's so important. And it's so often ignored by by the mathematicians behind all these signals.

Adam Lovallo:

Now, that's that's an excellent point. It's, it's easy to lose touch, you stare at all this stuff through the Facebook ads dashboard, and you forget like, my mom sees the ads that we run. And she's like, Oh, I saw her. But I haven't really looked at them in the context of the actual user experience in like, a long time. Easy to get disconnected. At least I find it's very easy to get disconnected. Um, okay, jack. So so there's like from a consumer perspective. That's cool. So we set the foundation, we talk a little bit about how to operationalize. Like, how could you sort of maybe weave this into an acquisition strategy?

Jack Dempsey Southerland III:

Yeah, I love it. So the acquisition strategy for burgers and shakes night example is great, because it talks about the customer journey, it talks about the user experience. And it's important to recognize where your customers are in that journey. And how do you provide sort of the right marketing touch point? What is the best optimization through the lens of value? So one option is really more revenue per purchase, right? That's where we see the most growth, what are some targeting practices? What are some tests that we can run to really see if we can get more revenue per particular purchase, and how do we measure that on the back end? So it's really about that experimentation. The other piece is how do we get higher frequency purchasers? How do we get people to buy more, how do we cross sell and upsell. And I think as you start to think about that, for your particular business, you develop a matrix and you map that customer journey, and determine how your acquisition strategy can be deployed for said user based on where they are. And using that data that we talked about earlier to pass back an informed decision. So it can get really, really interesting, but I think targeted and really derive value for both the marketer and the user.

Adam Lovallo:

Okay, that's cool. And I mean, I think you see a lot of focus on upsell, increasing basket size, in your day to day, do you see a lot of regular old price testing? Do you see that as a common, like lever? And sort of second part question? Are you seeing strategies that are meaningful around increasing frequency? It seems like a really maybe a harder problem, but a higher level more leverage, like, more and more impactful thing to solve if you can solve it like what, in your day to day, what do you what do you encounter?

Jack Dempsey Southerland III:

Yeah, I think the first point you make, in terms of price point, I think that's an important piece, and really understanding user behavior. So oftentimes, our teams will work with clients to do sort of an LTV audit, or an analysis, and glean insights. This is where it gets interesting, because you can determine what price points and products are most popular with users, what indicators are indicative of value, who has the potential to turn just based on what we're seeing. And so I think it really lends itself to test the learn, and being able to quickly understand what's effective and what's not. When you look at the frequency, there's a lot of experimentation with frequency, we see that quite often pun intended, because we want to understand how to strike the right balance between one making sure Ian gets to the next level in terms of skis and his burger and shakes night loyalty, but not being to the point where it can lead to churn. So it's about striking the right balance and all of these levers that you talked about frequency and more purchase, can can be utilized for testing to understand which one is more incremental and can contribute to more long term LTV.

Adam Lovallo:

Okay. All right, cool. So we've, I would argue that we've covered more or less the basics, the basic premise. So now let's say you know what we'll talk about the more sophisticated stuff. So let's get into, like you've even touched on some of these things. But thinking in payback period terms, how to go about that, not what it what does a p LTV setup look like? What is that? How are people going about that? What what leverage does that give you and anything else that you'd, you know, you'd put in kind of the the advanced class, like what what jumps out to you, for a, quote unquote, sophisticated advertisers that are doing this stuff? Well, like tactically?

Jack Dempsey Southerland III:

Yeah. So obviously, there's so they're focusing on the long term. They're doing their in house models, and using that data to help inform our targeting, bidding and optimization. And it really comes down to recognizing their level of predictability can drive profit, profitability, right. So they're going a step further, they're combining their revenue per customer information with purchase, and to building different LTV models. These models can be across a cohort, a particular segment or at the user level. And that's being passed back to us via a signal. And going even further, it's assessing the different payback periods, they're really trying to determine at what point is this customer no longer profitable? And so is it over 90 days? Is it over a year? At what point do we recognize that we have met our return on adspend? And that we're hitting our goals and this user is becoming valuable to us? And so those are some of the things that our advanced clients are doing. They're also using LTV, you know, Ian talked about discovery. So it's not about looking at just your existing customers, but how do I grow my base? And so they're using LTV to seed audiences and find more consumers who are most profitable and high value and then bringing them into the funnel to can you continue to grow that business?

Adam Lovallo:

And practically speaking, we're saying is, where you've got a bunch of custom audiences, your seed audiences, you maybe segment them by value, you're building different look alikes. Maybe based on product category, maybe based on just value in totality, and then testing into them in the hopes that a look alike have a high value customer seed will translate to high value customers in the future is that do you see that actually play out? Like I mean, obviously, like intuitively, that would be a great outcome like is that is that an outcome that the machine or your system is often able to achieve?

Jack Dempsey Southerland III:

The important thing is what you're what you're proposing is a business question. And so from our perspective, we were focused on how do we test that via measurement? Does a seed audience using LTV drive value? And how can we compare the lift and incrementality to that? So it's really about setting up a test to learn and testing business as usual, and testing using seed audiences, using LTV signal to seed audiences. So I think that's how we can answer that question. And more often than not, we've seen clients see success there, because it's building upon a foundation that was already of high value users.

Adam Lovallo:

So you sort of touched on this briefly, but like this notion of predict predictability, like projected LTV's, as a far example, and how that ties into profitability, so could I mean, practically speaking do you have any examples like how that, you know, might look in practice, especially relative to a more kind of vanilla. Just broad targeting No, PLTV, no, nothing, just, you know, optimize for CPA call it a day.

Jack Dempsey Southerland III:

Yeah, I would say it's being deliberate. So I'll give you an example. Right? It's 2020, we're seeing a lot of consumers moving from different places, sort of based on where they are. And so you're looking at a HomeGoods brand, they might identify someone as a mover, based on how often they've gone to the site, a particular purchase, how much are they spending in a transaction, if someone's moving, perhaps they're spending a high value trying to furnish an apartment. And so these types of signals and behaviors happening on site and through purchasing data might lead a marketer to place this consumer into a high priority group, and put them into sort of a mover specific category with a specific move or creative treatment. And so it becomes much more tailored based on the cohort that they're in. And so the PLTV model would understand current and future opportunities. So at the end of the day, we need to strike while the iron is hot, right? And so when the model picks up these signals, right, high number of purchase site visits, going to different pages, that customer is going to be moved into this cohort. And we're going to be deliberate with our communications and our creative and trying to get them to really what I would say convert and at the end of the day purchase, and at the end of the day meet that consumer need with the right user experience.

Adam Lovallo:

Okay, Okay, awesome. So, um, let's talk, organizational structures, as that is often a big determinant of whether or not this approach can be taken. I'll give you a real client example, talking to the CEO. He's like, well, we have to have a $40, CPA based on Google Analytics. And I was like, Yeah, but you care more about efficiency, right? Like, ROI, like, you know, the value PLTV, all this stuff. He's like, yeah, like, Okay, so then forget about the CPA goal. He was like, No, no, no, it has to be a $40 CPA, now's like you, what? Did you hear me? He was like, yeah, this is how we do it? So um, so my question is like, why is what you guys are saying, not more ubiquitous? What are the reasons? What are the challenges like? Like, what is the impediment that you see most often, to taking this more I guess you could call sophisticated or just even more nuanced approach, because that's, that's really what it is at the end of the day.

Ian Simons:

Yeah. So I think that that's a great example. And we see this a lot. And, and ultimately, it's very easy to do what you've always done as a company. But from an organizational perspective, what what at least I see is when they're earlier on as startup, sometimes sometimes these teams can move quickly in your meeting with the CEO and you can have these strategic dialogues. As as these companies grow up, you get a lot of fragmentation and the organizational structure which which you know, harkening back to discovery, commerce and the the customer experience, everything starts breaking down, and companies start moving really, really slowly. So what we see our most sophisticated LTV oriented customers doing is is they build what we call agile teams. And they're typically under one leader. And that means paid marketing by channel, organic marketing, creative media analytics, site analytics, UX design, you name it, every ad tech analytics, it's all rolling up to one leader. This is really important because then that leader has not only a view across the whole customer experience and can kind of optimize and move resources agilely around as the company gets bigger, but this leader can also align with the CEO, the CFO, the people who really have to understand how the metrics need to change As we start getting more sophisticated about ROAS and lCV

Adam Lovallo:

I think i think the listeners of this podcast and certainly I would swap agile, whatever, teams, which is great. We like that's what a growth person is really like, right? That's the point. I mean, it's meant to be cross functional. I think that things tricky about our industry, though. This is like, one of the first times I've heard anyone at Facebook or Google, like articulate what, like a growth person is I get that question all the time, like, Well, what do you mean? Like, we're not hiring record market, we're having record growth, why I'm like, Well, I'm like that, that's, that's nail on the head, it's like, it's, the description I give is like, it's a cross functional skill set, like your rifle stop, like, that's what it is like, cuz maybe, you know, maybe Jack's group is gonna come in, and we're gonna do a bunch of post purchase, surveying whatever to validate, blah, blah, blah, and like, that's in scope of our efforts, even if it wouldn't traditionally be like, the marketing team responsibility to go do that. So I think that's spot on. A lot of times, that's the last, like the last firewall, it's like, yeah, you know, what, I'm comfortable with the growth guiding measurement in front end development, and whatever, whatever. But I still need to have a shop around who's the head of brand or the CMO? Cuz like, I don't want to do anything dumb? And that's, um, yeah, I think that's a big hurdle for people to like, get over and or to find the right person that actually can span those areas. Like, I personally, we perfectly comfortable as a head of growth at any one of these companies. I'd be horrific if I was also in charge of brand marketing, cuz I don't know anything about it. So that's, you know, that's a that's a hard, that's a hard thing to solve. JACK, do you have any, any thoughts on any of this organizational business?

Jack Dempsey Southerland III:

One thing I'd offer up is focusing on the learning agenda, right? And making sure learning does not live only within the analytics team or with the head of growth, but across the organization? And how do we bring different cross functional teams, or this product monetization, marketing growth together to align around a set learning agenda that moves our business from our current state, through the LTV spectrum. And so for me, it's about being in a constant state of learning. And it's about embracing testing and learning and experimentation. And I know that sounds very easy on paper, but you have to prepare your organization, because I mentioned earlier, just being in the measurement side, we all know, and we can think back to that one test that can sometimes derail an entire initiative. And so how do you prepare your leadership right? In your teams, that failure is an opportunity to learn? And how do you build that into your test plan in a safe way that your team is constantly learning and understanding what works and doesn't work? Because we've talked about pricing? We've talked about creative, we've talked about so many different variables, but how do we in a structured way, test and understand what's moving the needle? And so it's that mindset shift. And it's that learning agenda or that organized central document that helps everyone from the CEO, right to the actual on the ground teams understand how we're moving towards a more value based marketing approach.

Adam Lovallo:

Yeah, I think that's, I think that's exactly right. Spot on. Okay, so we have, we have maybe another 1015 minutes as long as we need, but we want to talk about measurement. So okay, we know, we want to be more nuanced in our approach to customers, some are higher value and some are lower. We know at a minimum, we probably want to be looking at efficiency in ROAS that's established. Okay. We probably want to give the machine more control. Okay, fair enough. But we have the measurement to know what is going on, right? To know if it's working. So same kind of question, Jack like foundation, right? Like what, like to do a good job to measure the performance of Facebook and Instagram campaigns, like, what what do you need?

Jack Dempsey Southerland III:

Well, I think the, at the end of the day, the most important question is does the shift to value and focusing on LTV, drive higher business outcomes? Is it more profitable to my business? Is it incremental? And so our biggest focus is incrementality. And how can we create a sustainable marketing approach that grows the business? And so that's the most important thing and I think prerequisites for measurement is understanding. We're having these conversations now. It's 2020. And there's a lot of opportunity in the marketplace, where e-com, Ian talked about we can never erase the good, a good crisis, but they have to balance the short term and the long term. So we're gonna we're going to prepare the organization to make sort of that mental shift from short term to long term but you need measurement and detailed test plans to guide everyone and to show them that we are making step change. And it's proving to be incremental. The two ways we do that incrementality testing, understanding our attribution systems and models, and making sure incrementality is the gold standard.

Adam Lovallo:

Okay, that's awesome. I'm gonna ask you, I put you on the spot. You may you may not like this question. So you may scrap it later. But like, one thing I see a lot is, you know, a client creates an understanding of user value a p LTV model, right. Okay, great. And they can feed that back to Facebook. And you know, you can use that to optimize as we've discussed, and that's awesome. But a lot of times, I find that the predictive LTV modeling business is predicated on relatively unsophisticated attribution. For example, people will say, Hey, you know, my search campaigns have a have a payback projected payback period of nine months. And my display campaigns, my Facebook campaigns have a projected period of three months. And it's like, well, but you know, all of this stuff is kind of working together. Like there really, there really is no prospecting, customer and retargeting customer, it's like, you know, they're the result of our marketing. So I'm just curious, like, depth Java perspective on that, do we run into that? Maybe that's just sort of solved by thinking about incrementality as that is, in effect channel, like a channel agnostic observation? Like? Do you see any conflict between those two approaches? I guess my question.

Jack Dempsey Southerland III:

Yeah. You know, at the end of the day, I think it differs by client and by industry. But I think what you get at is probably what's a big challenge in the industry, right? How do I utilize incrementality? in concert with my attribution, marketers have MMM, models, they have MTA. So just try to make sense of all of that. And I think foundationally, our best approach, in terms of just thinking about it, from a value perspective, is really understanding at a specific channel level. What is the influence of said media? In terms of outcomes?

Adam Lovallo:

Right, that's incrementality, basically, yeah.

Jack Dempsey Southerland III:

And our clients will use different models, everything is customized, you know, it's a great question. And hopefully, you'll have us back to sort of unpack that. And I know, we're going to sort of talk to that in our playbook. But that's on the far end of the journey. And I think our clients are actively trying to solve for that. Now, how would you want the systems

Adam Lovallo:

You trigger this result, like, people are at different stages, you guys see your credit have made Lift testing remarkably available. I think that is the number one most impressive, impressive thing to me in the Facebook ads platform like that sort of lift testing in any other context in any other channel is either A, literally impossible, or B, you basically have to beg an account manager to ever even consider configuring such a test. And you know, the cynic in me says, hmm, why would a source of traffic not want to enable incrementality? testing? I wonder if it's because they're worried about their results? You guys, on the other end of the spectrum, you've built this product that is self serve, it's incredible. So like when, like company size stage wise, when do you think that is appropriately? When would you start to introduce that? Is that is that now something that's so universal that from day one, you might as well run a quarterly incrementality test, because why not? Or do you view that as more of like, you know, a varsity thing, then you have to get off of junior varsity first before? It's Yeah, before you can do it.

Jack Dempsey Southerland III:

I wouls say we see Lift as foundation, right. And at the end of the day, we want to operationalize it, you can use it whether you're early stage or late stage, because at the end of the day, how much of my marketing is working? Right? How can we tie back marketing to actual business outcomes, and no matter the size of a business that is important? I think the important thing is that Lift has to be interpreted. And each business has to figure out how do I incorporate these live results into my business in a way that makes sense. Yeah, the day I believe in incrementality, and I strive for that with every dollar I spend.

Adam Lovallo:

Yeah, I think that's remarkable. Like that's I'm not to be, you know, I'm not worse at Facebook, but that is like in the industry that is a unique perspective, like the fact that it's possible is so unique today, and you have to I assume it's you're going to force everybody to be able to offer that because it's you know, it's so critical. Anyway, and you want to weigh in on this at all.

Ian Simons:

There's this notion of correlation or causation. So they break out testing control groups, and that's exactly what Lift is looking at. Not just correlation, but what is actually causing the business, the business results and that and that is Lift. That is the ultimate equaliser across all channels in this in this mini channel world that we live in. It's simply applying Lifts and doing it rigorously, often. And it's applying multipliers to equalize channels.

Jack Dempsey Southerland III:

And I would ask which one of us is on the measurement team? Because Ian, that was well said, and that's a great point, Adam, you know, when your organization has embraced incrementality Lift is when your non measurement and analytics person couldn't explain why, how, and more importantly, what it does, and I think Ian is a great testament to that.

Adam Lovallo:

Well, I've got I've got one more, one more random anecdote for you guys and for our listeners. So you know, if you're on the buy side, or even in an agency, you get a lot of inbound, right? You guys probably don't get this. But I do want to run this test this test this test. So the the troll move that I do to make those conversations very short, as I say, Yes, absolutely. I would love to run a test. I'll give you an uncapped budget. Can you describe your incrementality testing methodology, and I swear that you will get either no response, or people like, Oh, I'm not really sure there's a fit, like, maybe we'll circle back, like, hundred percent hit rate, it absolutely crushes. So there's a little troll tip for anybody out there.

Ian Simons:

Yeah, I love that. And it's so big. It is so basic, but so many startups don't do any incrementality. And then the biggest thing we work jack and I work with the most sophisticated e commerce companies, and it's really basic what they're doing, if you boil it down, maybe not the really crazy ones way to the spear. But what the what the majority are doing are just running lift tasks, and calibrating channels, and it's not sexy, it's it's

Adam Lovallo:

I think the always on lift stuff is cool

Ian Simons:

is really important.

Adam Lovallo:

I've only personally seen that in a couple instances. But like yeah, makes I mean, it makes a lot of sense. Like you have constant pace as a marketer growth guy, you have constant ammo to go back to the CFO and be like, hey, like, what's up? Like, no, this is, you know, this is real, like, this is not me, there's putting numbers in a spreadsheet. Um, okay, so we're gonna wrap up with a couple quick things. But this conversation is part of a bigger initiative of your guys, as I understand it. So what uh, What stuff? Are you publishing what you know, many case study, what do you have on this topic in the coming months?

Ian Simons:

Yeah. So later this month, your Facebook team will have more information about kind of this value journey. And next year, we're going to early on in the year go out, go out with something more formal. But really what this is, and it's really interesting is we've studied thousands of advertisers. And we've mapped those advertisers from the beginning of their journey to the most sophisticated, and we've mapped out buy as basic as business need, to measurement to the signals to every every aspect of their campaign, how have they gone from a very basic marketer to being the most sophisticated. And as Jack said, it doesn't start by being a basic marketer, and then rushing to LTV that that just blows up and doesn't work. But this grid will help us do is it'll help you incrementally test the different levers as you become more sophisticated and as your business grows.

Adam Lovallo:

Okay, gentlemen, this was awesome. Thank you. I we all look forward to whatever materials is for mention grid and all this initiative that you guys are pushing. And perhaps in a subsequent conversation, we'll go deeper on incrementality and lift as I think there's much more to say there, practically speaking. But once again, thank you both.

Jack Dempsey Southerland III:

Thank you.

Ian Simons:

Yeah, thank you for having us. Really appreciate it.

MAU[Talk]:

Thanks for joining us. Stay tuned for Facebook's value playbook debuting in the coming weeks. Make sure to subscribe wherever you get your podcasts and we'll catch you on the next episode of MAU[Talk].