How to Build a Paid Social Strategy Around Real Audience Behaviour | Mark Byrne, Brave Bison and Gareth Harrison, SocialChain

Welcome back, Gareth Mark, and everyone at home to the social minds last brave talk crossover episode. In this episode, we are going to be talking about how to activate the media lift effects. So if you didn't listen to the previous episode, we were addressing the reason that this exists, the problems facing marketers and brands right now in the paid and organic social landscape. And in this one, we'll be digging deep into how marketers can actually activate this to solve their many, many problems. So to kick things off, our question is, how do we take the media lift effect, activate it, and actually turn it into results? So the first step to help activate the media lift effect is to audit. So we need to audit to understand what creativity we have to work with is this creative goal for us to be able to move up the funnel. Do we need to make iterations similar to what Gareth was saying in episode one, we need to make sure creative can speak to each stage of the funnel, because we know there's different mindsets at play. We need to make sure we're catering to that as well. Secondly, we need to audit the ad accounts as well, because modern paid social when we're amplifying organic content, it is very much an automation based game these days. So we need to make sure that we are capturing conversion data as best as we possibly can. This conversion data then helps inform bidding strategies, bidding strategies then help us reach the right people with that creative. So we've audited that creative, audited accounts, then it's a bit amplify. So we also talked about previously, who is your audience, what they care about, where we can find them online. So we're taking this information that planning our media around this. So let's say we know 80% of that audience spends time on Instagram, for example. Let's plan that media here, and we'll allocate 60% of our budget to demand generation, 40% of demand capture. This is really starting point, just to make sure that we are feeding that funnel, reaching new audiences, helping us to grow. And our final step is measurement. We need to make sure that if we are moving further the funnel, if we are making more a point to amplify creative, we're proving the impact of this. It's definitely challenged, not easy, but there's three steps to this, which we'll happily talk you through later. So we've established that not all attention is the same. We have passive, active and customer. I'm keen to know if you can share an example of what happens when you're actually able to align media and creative and understand how people actually think and behave at each of those stages. Yeah, so I think some of the foundations for some of our thinking for the media lift effect came out of some of our organic and paid work with Shark Ninja, where we saw that once we started trapping high performing content into paid, the results just went through the roof. And when I say results, I mean, from both a social point of view, but also able to sell out of product as well. So thinking back to cryoglow as an example, or double stacked air frayer, you know, all of these different products in different categories, we're able to completely sell out by merging or combining work that's working well or performing well in organic and trafficking that into into paid. With the audience GPT, another tool I'm a little bit fond of, a little bit scared of after someone made it into EVE GPT for the last social lives. We know that you're simulating real audience behavior, which is very cool and is already proving super effective. But what is something surprising that you've seen, where I guess random marketers assumptions were one thing and then what a tool like audience GPT was able to provide was super different. I'm going to give a politicians answer. So one thing that surprised me when we introduced clients to audience GPT is just how surprised they are, because we generally will demo audience GPT in the room. There's part of a pitch process or as part of a onboarding, so we'll introduce the client to their audience. And for those of you who don't know audience GPT just very quickly, it's essentially an audience insights tool that we've built. It's built on an LLN. What it allows us to do, we can take persona data that a client might have. We can fill in gaps with third party research such as global web index. And then what this do is it brings what otherwise would have been a flat deck almost to life and becomes an audience we can talk to. We can talk to it as an entire hive mind or we can ask it to categorize itself into four separate audiences name itself, tell us what's important to it. So it's really really cool. But to go back to your question, that always surprises me is quite like watching a client's reaction, because they're surprised. Oh my god, this is. Do you think it's different to what they have sitting in decks? Because the amount of times I've seen a client deck and they're like, these are audience personas. I think that is not who I've seen in your establishment. Yeah, we did have a client this has gone back years ago, but they're one of their top audiences where it's a male brown. So their audiences were male 16 to 24. But their over indexing audiences were people who were in market for the products that were generally sociable women, so make up handbags that up. So if they were refused to listen to none of this is their audience, but over time, this is who's buying a product is generally gifted up stuff. But that's the name suffer of the house. Yeah, that's up things that sometimes they are these surprises that we're trying to feed back to clients that, you know, this is who your audience is in. I think I think with the with the audience GPT work as well, I think there's been some really good use cases come out of it in terms of from what I've seen, helping build strategies or helping build like real time intelligence engines by taking mass amounts of data from social listening from creator or from trend tracking and then combining that with the power of AI or the power of audience GPT to give you again, this real time in intelligence tool or system, if you will. I think another really strong use case for it is when we've tested or AB tested creative. So putting side-by-side comparisons or even loading creative assets into the GPT itself and allowing it to say or tell you which the which creative the audience would be more receptive to. And we've tried this on various different levels when you put it when you put really terrible social creative in there versus really good creative, it always picks the good one. So it's almost like you've got an audience at your fingertips. It's a really strong tool. I also think it's like having a say on's and using Ouija board. So are you in the room on us? Yeah, that's how I feel. No, it is helpful. Does it validate you in the same way that chat GPT does even if you give it something rubbish or has it been modeled to actually be pretty brutal if something is crap? There's a huge amount of rigor that's gone into that because that's definitely the challenge and like, yeah, chat GPT can be your personal hype man. And wherever you say, yeah, that's great. I can definitely send to them that they're on parts of times. But we have gone through lots of rigor to say, yeah, be fully blown. So I think it is blown to Garret's point before is that it's nodding like he's been personally a type of guy. Yeah, not my tempo. But it will tell you if like this isn't for for me and I will say why as well. That's the go thing is that me on time poor I want X ones at my in my creative. So it's really interesting that way. And from us from a testing perspective, testing is always part and parcel of what we do is how it's how we unlock additional growth. But we can also take time because we're trying to do we might start with shotgun approach and then over time, okay, this is that the avenue we need to go down. But we can talk to rodents, what are some key themes of trends that are important to you already were three or four months ahead of what we would have been because we can zero in there too. But then, yeah, big fan of voices, the audience, GPT, if it didn't get already. And it has got, it's got a 90% confidence rate. So any question you ask it, it's well up to the standard of what you'd expect it to be. But in terms of the tool itself, like I just said, you've got a real-time focus group at your fingertips, are you necessarily going to get that 90% confidence from an in real life focus group? Because they, you know, there's there's probably some bias that comes into the group now to tell you what you want, they think you want to hear it. Yeah, no, that's so true. And I always think people end up saying things that like they think makes them look good or do they want to be rather than what they're actually doing. We're just untrustworthy. Now rate is not we untrustworthy sources. But Garrett, what is one way that social chain approaches audience segmentation? That I guess is, you know, slightly different. This is, pretend like I don't know the answer and I have your moment to plug. But for our listeners sake, how do we turn those insights into actual decisions and not just a nice slide? Yeah, no, it's a really good question. I think like any good strategy, it should start with an audience right? That should become, that should come as no surprise to anybody. I think the difference in what we do versus what I've seen from other agencies or even other clients is that we look at mindsets. So we start to take big audiences in the tens, in the millions or in the tens of millions. Sometimes if you're looking globally, it's in the hundreds of millions and starts to break those out or start to break those down into smaller segments through through through audience mindsets. That differs. from some approaches which have seen of agencies and clients taking a demographic or social graphic outlook. So we think that that mindset's work for a couple of reasons. First of all, it lets you or allows you to break those massive amounts of that mass audience down into more manageable segments, like I've just said. That then allows you to look at what makes those mindsets tick. So within those mindsets, you might have a cluster or a cluster of audience mindset that is into music or sports. You might have another one that is anti sports or anti music, unlikely, but but might be into TV and film, right? So it allows you to align what the edges take those mindsets, look at what they're interested in and align creative to that. Another reason why I think it's good is or where I think mindsets are more beneficial to brands is because it's the way that the algorithms act. So to my knowledge, algorithms don't push content or serve you content based on your age or gender, right? Or even your household income. Yeah. It's based off what you're interested in, whether you engage something, watch something like something, it will serve you more of that content. So you've got two, you've got two things they're playing, playing side by side. You've got breaking down mass audiences into manageable segments, but also feeding the algorithm, what, what it wants. And I'll just give you an example that I use all the time of what, of why we need to shift, change that mindset or shift that mindset from demographic and social graphic lead and more into mindsets. And, and it's a, it's a pretty old example, it's a, it's a good one of Prince Charles, Ozzy Osborne, both born in 1948, both live in a castle, both married twice, both mega wealthy, but come live on more opposite ends of the mindset spectrum, right? So imagine you graphed that massive demographic and that massive social, that, that short target audience, imagine targeting King Charles with heavy metal music. He just, he just wouldn't, maybe it wouldn't. But, you know, that's, you know, two different worlds completely. Coming back to the white paper, it says creative just 70% of the job. And I guess that's something we got really deep into in episode one. And, but there, there's still the case that not all brands treat it that way. So where do we think brands are falling into that trap, not prioritising creative or taking it seriously enough? And how I guess do you shift stakeholders to see it as the main road driver, especially on social? You think the, the challenge is that good creative is quite hard to do. So oftentimes, yeah. And so oftentimes we can build them, building out media plans that's hard to. But we can build up the media plans. And we can, okay, well, we can get you this, this reach, if we plug in X-Men to media, this is what the outcome will be. And then, okay, that's a lot more achievable. If you want, you can repurpose assets that I've been sitting on your website for a while. So it is a very low bar for entry to get live on them social from an amplification perspective. But for us to then showcase the clients, it kind of goes back to the point we had in episode one, is what's the missed revenue opportunity? Because we know if you've drawn creative, you'll get better cost CPMs, lower cost clicks, and just overall better performance. And then there are things, too, that we're trying to shift clients more towards as well, as how creative in itself is becoming targeting. Because how we're building our campaigns now, it's, as we said, maybe the top of this episode, but we need to start with collecting data, because we need to have conversion data that we can feed into building strategies. Historically, when we were building out, let's say, a meta campaign, you'd have 10 outsets, similar to a garage pointer and different mindset, someone who liked TV, someone who liked sports, gardening, etc. Or really granular, and then you'd just shift budgets between audiences. And that's how it was used for years. Like more of like an audience insights tool, really as opposed to a performance tool. But now it's a case of, let's run with large audiences. We'll take the data that we do have. If you have any existing audience data, let's take that gold standard. People who've bought twice plus people with a high average audience value will make local likes. I'll target a million people who look like that. And then we use creative, so people can essentially self-select, go back to one of the other points, dwell time and tomb stop rate. If they're engaging with the creative, they're essentially self-select themselves into an audience. Then we'll know, okay, they're engaging with creative, that we've built around this theme that we identify through audience GPD. So it's a way to filter through our audiences and allow them to self-select. There's also this idea of two for one creative. So one asset, sort of doing two jobs, which is an interesting point in the white paper. I'm not sure I agree with. So I'm keen to ask if you can explain that idea, or maybe give an example where a single piece of content was able to build the brand and drive performance effectively. Yes, I think where we were going with this two for one creative is that if you have, if you're short on time, or if you're struggling to know, okay, I don't know where to go with my creative, the best thing you could do is work with creators. They understand their audiences as we've discussed. They're really going to make them content because that's their living or at least their good chunk of their living. So that's essentially a two for one creative because they're allowing you to run creative that speaks to the audiences, works quite well, and it works. And what we could do from a paid perspective is amplify that. That's really just, if you have, you know, once you're put your money on creator content, we can amplify it from there. And I think, you know, to go back to another and stop with the, it's like 20% of our media is the recommendation, anyway, that matter or are suggesting these days is that we should be balancing our brand creative with creator content 80/20. Yeah, I don't disagree with that. I guess creator content is obviously effective, a powerful tool. I do think it can be a bit of a crutch or a lazy, a lazy solution because this idea of, you know, two for one creative existing is, you know, something that can do good brand work and build performance. The existence of that kind of also proves the existence of creative that doesn't perform. And what is the point of that? On social anyway, there's a crunching of the customer journey. So you can, if you think about brand being a professional performance being lower, there's a crunching of that, of that customer journey where you might discover and purchase all in one sitting. So I'll just, I'll go back to the sharp ninja examples in the case that he's there where we had a group of creators who were activating in that active phase. But the amount of sales or the sales volume that they got off the back of, of said creators was astronomical. And again, allowed, allowed them to sell out of the majority of the products that they went to market with. Yeah, whereas I guess in some cases it takes like a few steps to get to that point of performance further down. I know that makes sense to me. There's also this, I guess, stigma against organic. And it used to be sort of an old wall between organic creators and performance marketers on whose job was the more serious. But the idea being, you know, if organic is so unpredictable and paid is so much more structured, how can we test ideas in this chaotic organic content world and then use what we're learning over here, having fun in our unpredictable landscape to actually be able to run smarter ads and more structured paid campaigns without crucially losing what made them so charming and so effective through organic in the first place. Yeah, so I think from this perspective, and building bridges between paid and organic, how we could work together quite well is that in paid, we recognize that we're sitting on a gold mine. If we have a brand that has a lot of organic content, that's essentially a lot of potential paid ads. If we add a logo, brand logo, a royalty-free music and maybe a slow edit, then we have a paid ad. So we have plenty of potential there to test. And that's the case of viewing it through the lens that we've talked about before. So traction rates that Gary talked about, that's like the canary in the mind shaft of, yeah, this could potentially be a good piece of content for us to promote. One challenge that we do have a paid is that algorithm-based. Algorithms like to, you know, they like a bit of consistency, so they like to run what's working well. So one challenge that I'm sure many advertisers or brands who are listening to this right now is like, well, we do want to test new creative, but when we put it live, puts it back into learning, and then Meta will just go back to favor the creative that's already been driving performance. So top tip is if you run other separate, separate outset or separate campaign, you can essentially run this new creative, get some conversion data behind us, sweet spot, 50 conversions, and then Meta has enough data for you to then bring this creative into the overall campaign, and then it has enough data behind us to go ahead to head with the previous winner, because if you don't do that, if it goes in cold, you'll never win. So top tip is just to preheat the oven for an organic testing. All of that. I will say this from an organic point of view that organics tend to generally seem as like always on right, we're paid not so much. So I think there is an element of trust that you have to put or more trust that you have to put in in organic social teams so that that then empowers them to almost live life in beat-em-ode, not life, but life on social or when they're creating live life in beat-em-ode. So that they have more opportunities to test what creators are working better or what platforms might be working better or what formats or a multitude of different things, right? Once you've been able to test all that, you've got a massive bank of evidence for when it does come to big paid campaigns and you're going to know what is what is working in organic and you can shift those learnings into paid to keep that warm oven running or make it even hotter if you will. Sometimes think testing is a rich man's game, though, right? That 60/40 brand to performance split is still the recommendation man, something that gets talked about a lot, but I'm keen to ask, and I guess just advocate for those who don't have really deep pockets, is that flexible, is that the only way to win? Because if you don't have the ability to create so much organic content that you have loads to test from or if you don't have loads to put behind paid spend, what position does that leave you in and are there any signals that we can learn from that or let a brand know if it's time to lean more one mile or the other if they only have so many eggs and so many baskets, you know? Exactly. So I think the 60/40 split, so that's the work, the long and the short of it, the less banana in Peterfield, identifying that's the generally the best budget allocation between demand generation, demand capture. However, we know there's always needs to be flexible, not every client can do that. So this is really for, if you're an established brand and you're launching a particular social campaign, 60/40 is probably the best starting point. However, we know with some brands, particularly startups or maybe a brand that's a bit more sensitive to performance, if they're spending 60% of their budget further than the funnel, they're going to be asked questions in two weeks, telling by the CFO, particularly if they are using a last click as well, so we need to have a bit of flex. So when we have clients in that position, how we tend to operate is, do we even go 60% to bottom funnel or even 80% and once we're capturing available demand, once we're at that point of dimension returns, then okay, because I'm making a case for, we're now talking to the same people, our cost per conversion is increasing, the only thing we can't do is go up. So then it's easier to make the case well, we've done all we can at bottom, so let's start in building on those foundations from there. And I think that's the beauty of this methodology, the media lift effect, is because you can start to use those different stages of the customer journey as levers. So you might not have an always on 60/40 principle, you might have a certain period in time where you just want to go really hard on the customer phase, or you might, you know what, we've kind of run out of that customer demand, let's build some future demand and go hard on the passive stage. Yeah, no for sure, that makes sense. I want to talk a bit more about discovery, we touched on in episode one, and again, just looking at platforms like TikTok and Pinterest that we haven't maybe spoken about as much. I think naturally a lot of paid discourse comes back to matter, they're still absolutely monopolisers in the space, but more and more TikTok and Pinterest are where users are going to search and discover things and spend more time. So how do you guys brief creative teams differently for those channels, compare to your more traditional performance channels like Metta's own? And I guess what are some mistakes that brands are making in those spaces, perhaps trying to sell too soon? Yes, so I think it's a case of understanding why people are on those platforms, not to start high love, but on Pinterest, 98% of searches on Pinterest are not branded, so it really gives us an indication into that mindset of the consumer, they don't know what they want in terms of who they want to buy from, but they know they have an idea, they want to buy. Yeah, exactly, they're seeking inspiration, they're looking for ideas, it is that discovery engine. So for us, we know why they're on Pinterest, they're looking for ideas, we need to provide something that gives them that sense of cool, this is something that I need to and find it more better explore, but Pinterest might seem cliché, but there's definitely more of an editorial feel, more of like a slicker look that we do see that works well, there's always room for low five, but there is a difference there because generally, as much as there's lots of different verticals that do well on Pinterest, fashion, home, beauty, definitely them lead the way, so definitely understanding why people are on those platforms, TikTok, number one reason, entertainment, your content has to be entertaining, otherwise you're not getting, not getting that traction whatsoever, so we would generally work with clients, if you don't have the right creative, we'll push back and say really we need to do some iterations here, let's pull in Garth and the team to try and get something that will spark that interaction in. I mean, TikTok itself, we're on a whole campaign on it, and I still to this day see ads that don't feel native to the platform whatsoever. Do we think that TikTok and Pinterest are the only platforms that build discovery? Reddit. Reddit's a biggie. I would even say. We brands don't know what to do with Reddit, but more and more we're trying to educate them. We've had Reddit on a few times, and loads of stuff that you wouldn't think, but that is where people are searching for recommendations from real people about it. Absolutely, absolutely. It's a appear type of discovery, but I think that's in brands are definitely scared and hell you can tell because any ad you see, it's in 90% comments are turned off because brands want to be there, but don't really know fully how to engage because Reddit audience, you need to speak their language as well. Definitely. Yeah. It can be a hard one to put in. I almost think there's three platforms. You've got Reddit, TikTok and Pinterest. I like three points of the same journey, like Reddit's giving you those in-depth recommendations, but because it's already tax-led. Yeah. If I'm going to TikTok for the same thing, I'm like, okay, Reddit's told me that the best genes I can find when I'm having problems shopping for genes are Avocromby's curve fit. And then I go, okay, that's great. Thanks for time. It goes to TikTok. What do they look like on? Let me see someone wear them. Do a 12 form. So I can actually see. And then on Pinterest, I might continue that search and actually look at buying something or seeing what color I like or like, swatch the denim, that kind of thing. So we need a platform that does all three in there. Yeah. Yeah. Big time, yeah. So one last question. I know even though you guys work really close together, your teams have really different measurement models. I want to know where these models align, where they differ, and what you've both learned by actually combining them in real campaigns. Yeah. It's definitely true that we are seeing more disconnected teams, more disconnected creative, and that's leading to more disconnected measurement models. So paid and organic teams tend to be very, very disconnected. However, we actually work in a similar way. So we have got a similar measurement model or the same measurement framework. In the sense that we will take social performance, so social, social outputs. And we will connect those two audience outtakes, and then we'll connect those to your brand outcomes. So you've got this nice flow of what's happening immediately on the platform to some long tail effectiveness results. And that's why we believe a bit of a chest beat in moment that the media lift effect is the perfect methodology. Yeah. Because I think that these days there's no source of truth. There's no, like we used to talk about, what's the ultimate source of truth? Well, now we need to have like a suite of different levels of truth. So to go at this point, we have those three layers that give us that directional perspective from the platforms. Then we want to model data through like M&Ms, and then finally, it's those uplift studies that go standard of, is there activity having an impact? And then there's lots of stats in our white paper. I know we've talked about some stats today, but in terms of what that compound impact is. But essentially, it's lower cost reach, more efficient and effective activity, and we can get more from our budgets when we plan in this way. Once again, it's super interesting stuff. And now we've gone through so much that if you made it to the end of both episodes, thank you so much for listening, and there's even more information for you in that white paper. So the link is there for you in the show notes. But until then, we'll see you in a few weeks' time for the next Social and Sex.