EP113 – Unraveling the Evolution of Paid Media – Part 1

Author Sarah Hailwood 26 min read time

Welcome back to the Marketing Freaks Podcast. In this week’s episode Jon is joined by founder of ProfitSpring, Byron Marr. Byron got into Paid Media through promoting bands on Facebook and discovered a natural affinity for data driven creativity. Having worked agency side, client side, and now running his own consultancy business, Byron has a wealth of experience.

In this two part mini series, Jon and Byron delve into the changes that have taken place in the world of Paid Media over the years, discussing both the good and the bad. They explore how the industry has evolved and what marketers need to focus on today. During part one Byron shares his thoughts on the rise of automation, the importance of data management and strategic decision-making.

Subscribe links:

Apple Podcasts

The Impact of Machine Learning and Automation on Paid Media: Navigating New Challenges

It’s no secret that the rise of machine learning and automation has revolutionised the landscape of Paid Media, offering both opportunities and challenges for marketers. The fluidity between platforms allows for strategic budget allocation cross platform to achieve optimal outcomes. However, the abundance of platforms makes it challenging to align them with specific business objectives. Marketers must analyse campaign results to determine their effectiveness and decide whether to adjust, switch off, or reallocate budgets accordingly.

Machine learning has streamlined the creative and ad copy process, eliminating the need for numerous ad sets with different creatives. While this automation brings efficiency and performance benefits, it poses challenges in diagnosing and explaining campaign success or failure.

To navigate these challenges, experienced marketers who can analyse campaign performance in-depth are crucial. Applying these insights to the wider marketing strategy ensures a cohesive and effective approach to campaigns.

The impact of machine learning and automation on Paid Media is profound. Businesses must leverage experienced marketers who can interpret data and make informed decisions to drive successful marketing strategies. By harnessing the potential of machine learning while retaining strategic thinking, marketers can make the most of these advancements in the Paid Media landscape.

So if you’re questioning which platforms you should be prioritising, looking to understand how to apply learning from your campaigns into your wider marketing strategy, tune into this week’s podcast. 

If you’d like to discuss your Paid Media Strategy with our team, reach out and we will be happy to help.


Jon: Hello everybody and welcome back to the Marketing Freaks podcast. In this week’s episode, I’m joined by a very special guest, Byron Marr. I’ve known Byron for several years now. He’s got a wealth of experience and expertise in Paid Media agency side, client side, and now running his own consultancy business. So in the first of two parts, we’re gonna talk through how Paid Media’s changed since he got started, some of the good changes, some of the bad changes, and generally kind of the state of where the industry is at. So I really hope you find this one useful. Do make sure you subscribe and be in a good place to get the episode for the second part of this mini series. And let’s get started with the episode. 

Right. Well, Byron, we’ve been talking about doing this for what feels like forever, so it’s good to finally have you here. Thank you for coming down. 

Byron: Thanks for having me. Appreciate it.

Jon: Oh, absolute pleasure. So we’re gonna be talking a bit about, I guess just like generally talking about Facebook, better ads, Paid Social state of the industry where we think it’s going, what we like, what we don’t like, all of that type of stuff.

Byron: Sounds good.

Jon: Yeah. Perfect. Let’s kick off a little bit though with your background, because we met God, when was, it?

Byron: Feels like ages ago now. When was it?

Jon: It was pre Covid, wasn’t it? It was PC.

Byron: Probably, let’s think. 2017. Yeah,

Jon: Probably about that. Yeah, about that. So when did you get into Paid Media and how did you end up falling into this world?

Byron: Yeah, I think falling into it is the right way of describing it for myself and most people. So I’ve probably been advertising for about 10 years now. Early twenties I was kind of bouncing around different jobs, wasn’t sure what I wanted to do. And I’ve been promoting bands and stuff that I was in and friends bands with the early days of Facebook ads, which seemed to be quite effective. Even with a low budget of like five quid a day, you could reach the right people, get results better than going and standing in the rain and flyering at another gig, you know, for three hours kind of thing. So I felt like I had a natural ability with it or a natural enjoyment from it. And I wasn’t really sure what I wanted to do as a job, but I kind of knew the framework of what I wanted from a job.

So I wanted to have something that was data driven. You know, get to use my kind of analytical side of my brain, but also do something that’s creative as well at the same time. And then also have something where I could have a bit more flexibility, so work remotely, travel if I wanted to. And I just found that paid for me, ticked all those boxes. So I thought, yeah, I’ll double down on it. So I took a second job, started doing some training courses, going to networking events, and then eventually went agency side and that was where I met your good self. Yeah.

Jon:Excellent. It is. I do think, well particularly Paid Social has a really unique blend of creativity analytical, there’s the analysis side and the data side to it. There’s the technical side to it. It’s quite a unique combination of stuff, isn’t it?

Byron:Definitely. And it’s so different to search. Well certainly how search was back then as well. Because that was very much the data heavy platform where you just exported stuff into spreadsheets, doing calculations for bid adjustments. Paid Social was that, but then there’s also all this creative, imagery and videos and, you know, being creative with the way that you merge audiences and stack audiences together to actually target the campaign. There was a lot more involved in it which I found really interesting. And yeah, it just really appealed to me, so, yeah.

Jon: Yeah, absolutely. So just to clarify, you work on a blend of search and social, don’t you? But you specialize in e-comm now.

Byron: Yeah, so I’ve kind of taken the approach over the  last four or five years of a multi-channel approach. I don’t see any division really between, when I say search, I mean Google ads now as a channel. Which obviously lines are blurred there anyway. You’ve got campaigns like Performance Max, which has essentially what Paid Social had back then. It’s now got image keyword targeting, shopping involved in it… It’s a blend of everything together. So yeah, I’m like a 50-50 multi-channel now.  I don’t particularly deep dive into one more than the other. I kind of take the bits that work best for the client from either channel and sometimes it’s a bit more Paid Social if there’s more demand gen needed. Sometimes it’s more Google ads if it’s more demand capture. Whatever’s required to get the results really. 

Jon: I think being able to move budgets around between platforms is so important. And particularly now, like the levers you can pull on campaigns are totally different to the levers that you would pull on campaigns before. And we’ll get into the  “what was it like when you started versus what is it like now thing”, in a minute. But I think budget is such a big important control point or where you spend your budget more importantly. So having multiple tools at your disposal is, I think,  a massively important thing. And it’s, you know, right… if Facebook ads is going off the boil a bit. Meta Ads, I keep saying Facebook ads, I don’t like change!

Jon: I’ve only just got over not using Ask Jeeves. No. but yeah, so being able to go – right… well look. If that’s not performing quite as we need it, we’ve got search going that’s performing better, move the budget there or vice versa. And I think that’s a really important point.

Byron: Yeah, for sure. I reckon, I think for me it comes down to buying propensity. So wherever is best to target that is the best place to allocate the budget to. And you know, for sure, like, you reach different challenges with different businesses. But let’s say that you get going with search and that campaign’s working, you’re capturing people that are actively looking for the product or service. There’s a point where you’ve, you’ve done that demand capture piece and it’s like, okay, okay, what’s next? And then Paid Social supplements really well. In other cases it might be that, you know, Facebook is the right place to start and, and start there. And yeah, having that flexibility to be able to pick from different channels and also just to be able to go, do you know what, this is working really well in Google ads. Are we using that in Facebook ads? If you’re not in control of both channels, potentially from the client’s perspective, you’re missing out on that agility and that ability to just make quick changes and go, we’re seeing this work really well over here, let’s just get it implemented over here as well.

Jon: Totally. And there’s loads of cross learnings as well, aren’t there? And more exposure to more data is gonna help as well. We touched on it a second ago, but thinking back to 2016/17 kind of time first is where we are now… What would you say are the biggest things that have changed? Like, what’s the difference in your mind? Like how would you articulate the change between that era of Paid Media and where we’re at right now? Because it’s, and if you look at it through a particular lens, it’s just entirely, entirely different.

Byron: Different, yeah. Yeah. So the way I’d, the way that I would articulate it is I’d say it’s got simpler and more complex at the same time. So it’s the changes that you make on a day-to-day basis. There’s less of them required. Campaigns are less tactical. There’s, broadly speaking, there’s less to do just to be visible. So let’s just rewind to 2016/ 2017, every day in a Google Ads campaign, you’d be looking at keywords, broad search terms, bid adjustments on a multitude of different factors every single day. And all of that was to compete, just to be visible in the auction, not to be competitive and to get sales and revenue and all that kind of stuff as well. That stuff has gone to machine learning and automation. So now where you need to compete is the broader business stuff. It’s proposition on the website, pricing margins, creative, it’s all those different broader business elements, which are simpler to some extent cause you’re not exporting like a thousand keywords worth of bids into a sheet and then trying to do adjustments and then reimport stuff. So it’s less complex from that side of things, but it’s more complex because they’re harder problems to solve to some degree.

Jon: Definitely. It’s also an entirely different mindset. You know? It’s less black and white and you have to think about the bigger business cycle. What creator’s gonna really hook my customer in? What do they really need to know? What do they need to see on the landing page? Are my margins competitive enough to get a profitable CPA on them? 

Byron: Can it even work at all? Yeah. And they are complex problems to solve, but yeah, cf It is definitely a more complete task, if that makes sense. You think back to like, you wind the clock back and remember stacking audiences in Facebook ads where you’d be like, right, I want to target people who live in these four post codes, drive BMWs, like these hobbies, shop in these shops, read these magazines and arrange between this and this and that would work.

Jon: The thought of doing that now is just, it just wouldn’t, wouldn’t go anywhere, would it?

Byron: No. No. It’s far too restrictive. And I’m sure you’ve seen the same thing yourself, but providing that conversion data’s coming back into the campaign in either platform. Generally what tends to work is broader audiences and broader targeting. Yeah. You know, and the conversion data is the magnet that finds within the audience those that have buying propensity. So having such restrictive targeting now is a problem because you’re starting off with a pool of users that’s only this big. And then the conversion optimization looking within that pool for like, you know, one occurrence perhaps of a sale and it’s just, yeah. It’s just far too restrictive.

Jon: Yeah. Definitely. I think that point, data is the task now is almost so where the task used to be, like you say, export the keywords, look at the bids, make sure you’re being competitive, re-import them, you know, all of that really mechanical stuff that is now in, in theory taken care of. With automation, the job now I think is about data management. So do you have the right data being fed back in to feed this beast of automation? Is it accurate? Is it the right data? Is it going in, in an updated quick enough manner? Is the pixel implemented correctly? Is that passing the right information? Are the shopping feeds correct? Have you optimized the shopping feeds as best you can? That’s a big part of the job now, isn’t it? It’s about “how do I influence the automation to get the best results?” rather than “how do I fight against it and still hold on to some degree of manual work?” Do you think?

Byron: Yeah. Yeah. So I’d say it  seems to have these types of dis like these types of topics seems to have come to a little bit of a close from what I’ve seen in this space now. But there was a lot of talk for a number of years about control, about the need for control, the need for control over automation control. But in reality it’s kind of, it’s like perceived control. It’s not actual control because if you’re feeding in the right data into the campaign and machine learning’s optimizing towards the right goal and essentially like you’re pointing the ship in the right direction. If you’re doing that, the output should follow and if the outputs are following, it’s what’s needed for control and visibility. You’re setting a target and you’re achieving the target and that is your measurement and that is your control. Is it, you know, you look at it from that perspective. Anyway, that’s the way I was trying to do it. 

Jon: Yeah. Do you think there are circumstances where a more manual approach is still valid? So lower volume, higher value that we, I guess we’re a bit different. So we do work with quite a lot of B2B clients. Often the search volume’s very low, the CPCs are very high for the amount. And therefore because the volumes a lot lower, the data that you get to feed back in is much smaller, therefore, inherently there’s less to feed the automation cycle. In those cases, sometimes I think there’s still a valid case for at least testing a manual approach. And that is not an anti automation thing at all or anything like that. But do you think in your experience I guess different kind of with e-comm, isn’t it, but in your experience, are there still cases for a manual approach?

Byron: Yeah, possibly. I mean, you have to do what works, right? And if it’s not working and you try something else and it does, then ultimately that’s the goal. I think what I would try and look at in those instances is I would look at, you know, let’s say it’s a high ticket item or if it’s lead gen for example, where there’s a period and a process that needs to happen after the lead, before the sale is closed, et cetera. Is you would look at what different data point could you push in to essentially achieve the same thing. So if you’ve got like a high ticket item in e-comm rather than optimizing for purchase where you might have a long consideration period between the first site visit and then the sale occurring, where then you’re not getting enough sales, as you say, to come back in to get out of the learning phase. That’s where you would start to potentially try and optimize for a previous action in the customer journey.

Jon: An add to cart or?

Byron: Yeah, exactly. Yeah. So you would still report and measure success based on the sales, but from a machine learning point of view you might wanna optimize for add to cart, checkout started, something like that where you would, where I would stop with that is I would then stop obviously like something like a product page view cause that’s too early into the customer journey. And there’s not enough degree of intent to buy there, but essentially like you’re still measuring on revenue, sales profitability, but you’re just giving the campaign something that’s, you know, if that add to cart occurs a hundred times more frequently than the purchase does, then you can again help the system to achieve the same goal without needing to go to the manual approach.

Jon: Yep. We also see I guess like conflicting results against best practice. So quite often, again, with B2B versus B2C I think this sort of thing is more prevalent with B2B, but you’ll be running a campaign where the conversion volume is far lower than the recommended amount, but automated biddings still work like it still works better. And then at other times manual works better. And I do think there is a, still a big thing about this where I think the big difference is in lets say the last six years is that the expertise needed to judge the best approach and test the best approach is you, you really need a deep understanding of how these campaign types, how the automation works or that type of stuff is that is so much more needed now. I think.

Byron: Yeah, I think you’re right. I think it’s you need somebody now who understands both approaches who can then strategically decide the right approach to use in that, in that given instance because Yeah, you know, it’s not as simple as just going say back seven years ago, the approach that worked was pretty much the approach that worked and it was kind of like there was our way of doing things that…

Jon: Was it black and white, wasn’t it?

Byron: Yeah. Whereas now, like in those nuanced instances like you’ve mentioned you might say, okay, like we do need to do a manual approach here because this quirk has happened. You know, the normal way of doing things that generally works in most cases isn’t working. The person that’s able to make that decision and to implement that is someone that has experienced the manual way of doing things and also the automated way of doing things as well. And that’s a more strategic individual that can look at it and go, in this one instance only I’m gonna change what is considered best practice cause it’s what’s gonna get the best performance. That’s not the same as someone who’s just implementing a process as it would’ve been seven years ago, like look at the search terms, export the bids, do the things, that’s a different type of individual as well. So yeah, there’s an even more complex challenge there in terms of resource as well to some extent.

Jon: Oh, big time. Yeah. That’s a different job, isn’t it? Yeah, it’s a different job. Do you miss anything about those? We’re gonna call them, I was gonna call them the old days,

Byron: The good old days.

Jon: It was like five years ago. Do you miss anything about that way of managing things, that way of managing campaigns or that way of working in Paid Media?

Byron: Not a huge amount personally. I think I came into it at the time where Paid Social was booming. Automation and machine learning was just starting to come into Google ads. And I was kind of, because I was trying to sort of build my career and my expertise, I saw that change as a way of the playing field being leveled. And for me to go, okay, if all this change is coming in, I wanna adopt that stuff because that’s gonna reset the best way of doing things. Which actually as someone who’s newer into the industry is quite helpful. The only things that I do miss, I do miss the general enthusiasm around Paid Social in particular. It was, so there was a period of time where everyone was just, everywhere was just like electric about Facebook ads and, you know?

Like conferences that you and I have both spoken at, there’s rooms filled. Like even the walkways are filled with people just desperate to hear anything about Facebook ads and stuff. I sort of miss the excitement and enthusiasm around it because it was new and exciting. But in terms of ways of doing things like I’d say like net outputs are on performance are much higher now when campaigns are set up in the correct manner. And, and I prefer how everything’s moved to a more commercial focused standpoint. Everything is much more now, even with lead gen is more about outputs and the channels trying to get you to think about close deals, revenue, profitability. Yeah. Big time as opposed to just impressions, clicks and you know, that sort of thing. 

Jon: As we were kind of preparing for this, I mean I do, I kind of miss the simplicity of it and I did quite enjoy that manual going through, looking at keywords, rooting out things that are working, stripping out stuff that’s not like that I think as well, and not everyone’s gonna look at that and go, “oh, that was fun!”. But it was a little bit enjoyable. I think there was a thing with that, but it was, the work was a lot more manual wasn’t it?

Byron: Yeah, and I think there’s something in that.

Jon: Some of that was just a pain.

Byron: I think there’s something in that mindset and that approach, which is useful because you are looking into everything with granularity and you’re starting from what gets sales. What keyword was it? What ad set was it? What piece of copy was it? That mindset and that troubleshooting way of looking at things is really helpful. So that is something I think that we’ve probably learned from that era and we’ve brought forward. If that makes sense?

Jon: Yeah, yeah, definitely. I think the thing that worries me a little bit about kind of where things are now versus where they were, is that lack of detail in the lack of detail in what’s running sometimes or people coming in and assuming, because I guess like the marketing around these types of campaigns are, it’s easy, set it up, give us your money, tell us what you want and just let it run and we’ll take care of the whole thing. And it’s not that at all, that you still need that mentality of why is this working, why isn’t this working? What can we do more of? Like the curiosity is still really important. And I think my worry is that people will just set ’em up and run ’em and forget about it and assume that it’s gonna be alright and not kind of go into that detail so much.

Byron: Yeah, for sure. And I think there’s something in that as well, like in the way you just described it is even when, so let’s say that you make account now in the, the modern way, should we call it? You know, a performance max campaign, smart bidding automation, whatever channel it is that you wanna use as your example, but let’s say that you’ve made that campaign and it’s working really, really well. The results are one part of the value and obviously it’s a core part of the value. That’s what the client wants, that’s what the business is trying to achieve. However, it’s gonna be difficult to then go to that business and go, this is why it’s working when it’s working and this is why it’s not working when it’s not working. So yeah, like I share in your, I guess the concern about that to some extent because the real value there for the business, like the results will be the results. If the campaign’s generating revenue, then great, it’s generating revenue. But the more important thing from my experience anyway with working with clients is them understanding why it’s working so that they can take that into the broader business outside of paid, is value in itself. And if you can’t provide that, cause you’re not able to diagnose why it’s working so well and what in particular is working about it. Then you’re missing a huge part of the consultancy element, which now is a huge part of our value, if that makes sense?

Jon: Yeah. Big time. Like the answer to that question being “it’s working because it’s working” isn’t good enough. You know, you need – to me and you – we need to know how to scale stuff so you need to know what’s working so you can do more of that and less of what’s not. But yeah, on the surface level it’s kind of, well, we’re putting this in and we’re getting this out and you still need more than that.

Byron: Yeah, for sure. For sure. And again, from the client’s perspective, there’s so much value in that outside of page. Like if you can say, right, search is working really well, it’s these keywords, these products, this piece creative, this copy, this message, you know, this is what’s working about that campaign. Well then they can take that into their email marketing, they can take that into their organic social, they can take that into their website proposition, they can take that into, you know, SEO, they can take it into new markets, new products. And it’s like all of a sudden it becomes this amazing, exciting thing that, not to underplay it, but the profitable sales from the campaign is, is almost like the, “oh, well that’s great, but look at all this over here”. Because when we understand that we can then replicate it in different countries, different products, absolutely. Different channels. So it’s, for me that’s the, that’s the new excitement that I get out of it is the insight finding really.

Jon: That’s a great way of looking at it because there’s so much more to it. Isn’t there so much more to it? Like testing ideas, validating the landing page template. Right. Let’s run some Paid Media to it and validate whether or not this is gonna work better, you know?

Byron: Yeah. Does this message convert more users and the current website message? Because if the answer’s yes, that’s incredibly useful. Because maybe that needs to come over to the broader website, which benefits all channels. If the answer’s no, well that’s still a useful test. Because then you know not to do that in the future and change it. So it’s, you know…

Jon: All the demographics that are actually buying the demographics that we think are going to buy, all of that stuff. All the locations, the locations that we thought would be the most popular or not. And there’s so much business intelligence that can come from a good Paid Media campaign that yeah, you’re spot on. Like it’s, yeah, there’s loads of pluses and it’s not just about sales is it?

Byron: No, and that’s where you’d have the time to spend the time now because you’re not having to export every keyword bit into Google Sheets every morning and change all that stuff. Now you can focus on that, on that bigger picture stuff. Yeah, exactly. Yeah. I mean it’s much more strategic than tactical.

Jon: Definitely. Well that’s, I think a really good point to end part one. In part two we’re gonna come back and we are gonna talk about just like the current state of Paid Media, where it’s going, where people are getting stuck and your kind of advice for people running Paid Media at the moment. So we’ll talk about those topics we think would be great. So thank you everyone for listening to part one. Make sure you subscribe and come back for part two next week. Cheers. Thank you so much for listening. If you enjoyed that episode, please do come and subscribe. Join us for future episodes where we talk about the ins and outs of running Paid Media and driving improved conversions of revenue for your business. See you next time.