Today, I interview Tyler Carson, a senior strategist here at WebMechanix. He’s a subject matter expert on any digital platform that you can advertise on. If you want to run ads on Facebook, if you want to create Google Ads campaigns, — this guy knows how to do it.
– It’s not just “Who are these people that we want to be in front of? And how much budget do I have and how many leads do I have to generate? How many prospects do I have to generate?? It’s now, “How can I get in front of machine learning? Use it to the best of my abilities?” And I think that’s gonna be more with targeting.- Welcome back to another episode of More Than Marketing. I am your host Arsham Mirshah. I have with me today the beautiful NASA bound Tyler Carson. Tyler is a senior strategist here at WebMechanix He’s a subject matter expert as well on paid media. Pretty much anything that you can buy media online. If you can place an ad. If you can buy an email newsletter. This guy knows how to do it. He knows the best practices. He does it all day every day. I think in particular I would say your strongest suits are probably Google Ads, Facebook, Bing Ads.- Yup.
– Do you agree?
– So Tyler Carson is with me today. And we want to talk about machine learning. Big buzzword out there. I think it’s akin to big data. Five years ago, It was like big data. Big data. Oh my god, big data.
– Everyone does it.
– Everyone does it. Everyone wants to do it. Everyone does it. People do it without knowing they’re doing it, right? So machine learning is kind of that hot topic of today. So why don’t you give me, just tell me a little about machine learning.
– So machine learning and big data they go hand in hand. Machine learning, the best way to think about it is the usage of big data in a lot of different aspects. I mean it affects our day to day too. Big data and machine learning as a consumer.
– Yeah, give me an example of that.
– Yeah, you’re using Spotify. If anyone here uses Spotify, then you’re gonna know that they’ll create a customized playlist for you based off of what you listen to in the past, what your friends are listening to, what they think you’re interested in, based off demographic data, and probably a whole bunch of stuff we don’t even get access to.
– Love that.
– There is that. I mean… Tweaking the SERP or the search engine result page for users that are using Google. Tweaking that using machine learning based off of what they think you’re gonna be most interested in. But also your past trends.
– Past trends. Your location.
– Your location, of course.
– What they know about you.
– Local. All of that fun stuff.
– And then, the other one, which is another easy one to kind of internalize is Amazon.
– You purchase something. What’s the next thing that you’re gonna see in your check out page?
– The customers who bought this, also bought this.
– Yup. Trying to get that upsell.
– So trying to upsell you right there on the check out or on the cart page or on the product page. Yeah, so. What they’re doing is they’re, Amazon is looking at their huge set of data and they’re saying look, show me all. They’re telling the machine to find patterns and then, use those patterns in marketing or advertising, right? So, and by the way. Spotify, great example. I love that because that’s how I find new music.
– Yeah. Discover weekly. The release radar is always good. And then, it gives you like the six day playlist
– I just tell Google Home. I’m like, “Hey Google, play Stromae.” Right? And then that’s a French artist, and so it’ll say okay, here is a radio mix of that. First song will be Stromae. And then, the next song will be these French artists I’ve never heard of. New songs. Very similar, I love it. That’s how I get new, new music.
– That’s amazing.
– So back to marketing. Cause we advertise marketing. You’re the paid media SME. Google Ads. Paid advertising on digital advertising. How does machine learning kind of relate to that?
– Yeah, it definitely affects our day to day. So machine learning in marketing is more prevalent than ever. It’s going to get more and more ingrained in our day to day.
– Every single day. It’s just gonna get more and more important. What we use it for now would be.
– Yeah, what do you use it for like today?
– Targeting. So who is our ad getting in front of? Targeting also in the sense of when are they gonna be able to see our ad? Or what is the best time for them to see our ad? Building different criteria or rule sets around certain metrics. So “if this, then that” statements. But using machine learning to improve those “if then, then that” statements. But most importantly, I think one of the coolest ones is definitely building the audiences and bid strategies.
– Yeah. I want to stay on audiences first. So, targeting. I think one of the coolest things for machine learning or ways that you can use machine learning today in advertising, particularly, is this idea of what Facebook calls lookalike audience. It’s super cool. I know we’ve seen a lot of success with it, and I’m sure people listening to this podcast have or hopefully will. Tell me about Facebook lookalike audiences. How that’s machine learning. How that works.
– Yeah. So lookalike audiences. They’re really powerful. Basically, you can take a subset of an audience or, for instance, a remarketing audience. You might have 100,000 clients. You can then splice up and take the top 10% based on revenue or based on whatever criteria you want to fill. Take that audience and you can now extrapolate that to an audience of say three million people.
– In the U.S.
– And now you can target this, what is now completely net new prospects that you’ve never engaged with before. Maybe you’ve never to your website. Never heard of your brand. You can take that information that you currently already have at your disposal and build a new prospecting audience to get in front of. And hence, the name “lookalike.”
– Then, we’ll use machine learning Facebook, Google does it as well. They’ll create these audiences that are similar to or “look alike” this one particular audience that you’re saying by whatever criteria you’re defining. They visited my contact page. But they never filled out my form. Build me an audience that look like these types of people.
– That’s so cool. So, what I’m hearing from Tyler. And I know Tyler has done this in particular. He’ll go to a client and he’ll say, “Okay, give me the email addresses for your best 100 clients.” or “Give me your email address for 10,000 of your customers,” right? Depending on the client. B2B, B2C, size in there. And then, we’ll take that, and we’ll upload it to Facebook, and we’ll say, “Hey Facebook, these are our current customers. We don’t want to target them. But what we do wanna do is tell you that this is what they look like. You, Facebook, create me a lookalike audience. And Google does this too. What does Google call it?
– Similar to audience.
– Similar to audience. So, that’s super cool. So you can say, “I already have these customers. Don’t target them with ads. But find people who look like them.” Now what Facebook and Google are doing is they’re using their massive, they’re using their big data.
– Big data. That’s the apex of big data and machine learning in marketing.
– In marketing. It all came together. They’re using their big data to find patterns. Say, these people live in this place. Or they like these things or they visit these web pages or whatever. So now, let me find people who are not in this list who also match these patterns and show the ad therein, right?
– That’s absolutely right.
– Super cool for finding new.
– Extremely powerful.
– Extremely powerful. We’ve seen the results so we know how powerful it is. Hopefully, someone listening will agree with that or say, “Hey, I need to leverage this today.”
– Cause lookalike audience has been around. I think recently, I don’t know. Two years. LinkedIn advertising also added “expand my audience” or.
– Little expansion button.
– Yeah, little expansion button. It’s just a button. It’s just a click, just a little button. Machine learning is just a button.
– Okay, let’s move on. There is this idea of the machines are taking over. Right? So, do you think machine learning is going to take your job away, so to speak?
– Replacement yeah, get a robot with a NASA shirt.
– Just set it in the dusk and let it do it’s thing.
– Yeah, exactly. As long as it has a NASA shirt, we’re good, right?
– Yeah. Is machine learning going to take my job? So, probably not.
– Right. I agree.
– I hope so. So, what I think it’s gonna do is it’s gonna evolve the type of work that I do.
– So whereas, and right now, that’s not changing. It’s a slow progression. And new people that come in and have similar job titles as mine or come into the industry fresh out of college, it’s going to be a seamless transition for them. People that are in this position need to be open to this type of transition and adopt these new techniques and methodologies and if they are, it’s going to be a seamless transition for them too. We are so, I’m not really worried or feel like anything is gonna change, but the way it is going to change is what I do in my day to day.
– So right now, if a brand new client or brand new campaign we’re building things out, I have to go and backtrack into who is the right type of person? How can I target them on all these different platforms?
– I will still do that. That still has to be done. But, I can do it even better using machine learning and enabling myself and these platforms to use the data and the criteria that I’m setting forth to better the campaign performance.
– So, I can still do all the things I’m currently doing but now, it just gives me another piece of ammo at my disposal.
– Another tool.
– It’s another tool, and it’s funny because we were talking on another episode to… I was taking to E.J.
– He was, we were talking about marketing automation, and I remember when marketing automation was kind of coming out and coming forth, and it was like people were like, “Oh my god, marketing automation. The marketer’s job is gonna disappear, right?” Like no. Like in fact, quite the opposite. Because it’s complicated. It’s powerful, yes. But it’s also complicated. You have to have a human come up with the strategy. Come up with the plan for how we’re gonna implement this. It’s not just put it on the site, flip a switch, and now, all of the sudden, you have ROI. No way. It’s actually hey.
– If only it were that easy.
– If only, then We’re all consumers and that’s it right? And I think that’s what you’re talking about here. You still have to give the signals.
– Yeah, yeah. So signals. Everything that we’ll talk about in terms of machine learning that information that the machine has taken. That big data. Those are signals. You’re sending signals. Whether it’s conversion data. Visit data. Session data, any type of data. Email uploads. All of those signals. All comprise of what data you’re gonna share with these platforms and how you’re gonna be able to better your campaigns. The more data, the better, obviously. And they have a lot of data at their disposal that we don’t have and we have data at our disposal that they don’t have.
– Our jobs is to give them that data as much as we can. Label things properly. So that they understand what this data that we’re giving them is and then, such that they can then combine they being that advertising network in this case. Google or Facebook or Bing or whomever. So they can use that to the data that we’ve labeled for them to mix it with their data to then find prospects for us and get our ad in front of the right person. What do you say?
– Right person, right time at the right price.
– And the right price, there you go. Exactly. The right price, so bidding, right? It’s like, “Hey, if I know I sell more gutters when it’s raining.” I mean, that’s data, right? It’s like, Google can use and then. So, our job is to come up with that strategy and tell the machines how to use it.
– Yup. Google’s got that data. You can run scripts to help with that. There are a number of ways to go about that one very specific example you just mentioned. With rain and having a gutter client. Making sure that, right person, right time right price, using bid strategy, smart bid strategies is one that Google’s been pushing pretty heavily recently. But they’ve always tried to show your ad to the right person at the right time as best as they can. Now there is a little bit more power actually back in the advertisers hands to set parameters around what that means to us. So it’s interesting because machine learning has always been something that’s been in use if you’ve ever done anything in AdWords or Facebook. They’ve always used the tool to a degree Now, we actually have a little bit more control over what do we want to constitute as the right type of person. So now, we have actually a little bit more power than we did just a year and half ago.
– That’s good.
– Than we do today. Which is awesome. Google, I think has started to realize that advertisers know what they’re doing obviously. That they want to give us a little bit more power to say who is the right person to you Mr.and Mrs. advertiser to get it in front of all of these people because it’s not, it’s not a cookie cutter think where it’s “oh, perfect.” This is a client that Google can’t look at it and say, “This is the type of industry, the type of company that needs to go after these people.” We might be targeting a special type of insurance that’s only for a subset of humans on earth.
– That can actually purchase this.
– Yeah, right exactly.
– Might be hard to explicitly state those types of people. But if we have a list of people that we can upload or if we can over time optimize using machine learning towards those types of people, show them more frequently, it’s only going to help everybody.
– That’s how it works. That makes a lot of sense. Makes a lot of sense. So I guess the future. We talked about it a little bit but what’s the future of machine learning in paid ads or in your role? What do you see?
– Yeah. Someone’s gotta man the ship. Basically for us, it’s going to be identifying in my day to day, it’s gonna be identifying, “how can we use machine learning to the best degree?”
– What new things are happening on the machine learning side. Now, it’s a little bit more of a technological standpoint than it is just a day to day marketing standpoint. It’s not just “Who are these people that we want to be in front of? And how much budget do I have and how many leads do I have to generate? How many prospects do I have to generate, How many sits with a salesperson do I have to generate?” It’s now, “How can I get in front of machine learning, use it to best of my abilities?” And I think that’s gonna be more with targeting. I think that’s gonna get even more enhanced identifying who can we get in front of. And to what degree and splicing and dicing that data up in a bunch of different ways. Talked about smart bid strategies a little bit. I think that is just the first layer of bid strategies that we’ll be able to use. So things like “Maximize conversions” and “Target CPA.” Being able to let Google and kind of staying on that Google example of letting Google change your bids. Back in the old days, I had to set the bid. I dictated what would I pay for a single click. Letting Google decide what the right clicks are and bid higher or lower based off of that.
– We see them doubling your daily ad budget nowadays. They’ll stay within your monthly but they’ll double your daily just to learn. Because they want to get more data. You see that all the time.
– Or just to give me a heart attack.
– That’s so true. I love that.
– It’s one of those learning curves. You have to know, you have to know what to expect when Google is going to make changes like that or Facebook or LinkedIn or Bing or any platform that we work within. If there is a platform that’s not using machine learning right now, they will in the next five years. And if they’re not, then shame on them. They probably won’t be around, or their platforms not going to be used.
– You make me laugh. I remember when we have alarms that we set for paid media and we’re kind looking at different things like, like, for instance, “Is our budget pacing on track?” And I remember when Google came out with this, “Hey, by the way we’re not gonna mess with your budget.” We got alarms like, “Whoa, your daily budget of 20 dollars but it spent 40 dollars.” And we were like, “What is happening?” And we were freaked out. But it was like, quickly realized, “Oh yeah this is because they rolled that thing out that we’ve been talking about for the last 20 months so.” That’s just what that is. Okay. Beginning to reset our scripts and all of that.
– Makes it a little bit harder in the day to day but at the end of the day, as long as it nets a positive.
– And that’s what it’s doing.
– When you tell Google to optimize for conversions, right? What it’s doing behind the scenes is using it’s pattern recognition aka machine learning to find you better prospects or prospects who are ready to buy at that moment thereby, getting you a higher click through rate, thereby, getting you a higher conversion rate, thereby, really ultimately getting you higher ROI.
– Yeah, and machine learning can be somewhat scary but as long as you know what’s actually happening on the back-end, it’s really not this big hairy beast that people kind of make it out to be but you have to know what you’re doing as well.
– Have to know how to set it up, you have to understand what it’s doing. And when you understand what it’s doing, then you can set it up properly and you can.
– First step is understand what it does. Second step is understanding how to make that even better and the third step is how to analyze it.
– That’s right.
– And if you can do all three of those things, machine learning is not that scary.
– It’s very valuable.
– It’s extremely valuable.
– Extremely valuable. I mean I can’t think about I can’t imagine a world without Facebook lookalike audiences, for instance. That would be, that’s just so commonplace now for us. I don’t even remember when it wasn’t around. Right?
– Yeah. It feels like it’s always been there.
– It feels like it’s always been there.
– Platforms have always been using it to a degree but now, we have a little bit more control over it, and I think they’ve kind of removed the veil a little bit, and we can kind of see into what’s actually happening and know what we’re doing.
– And thereby set it up.
– And more empowered.
– And more empowered. That’s right, empower the advertisers, baby.
– So we could talk about this all day. I think for the audience out there, hopefully, they take away that machine learning is already in use and it’s being used all day every day. Simple example, such as Spotify and Amazon, where they’re suggesting songs or suggesting products to you is being used in Google Ads and Facebook to target you at the right time but also, if you are similar to an audience or a brand or a business is trying to capture. It’s already happening. Your job is to then, make sure you’re using it in your advertising, especially if you’re high-volume I think.
– Oh yeah, even better,
– Even better.
– The more data, the better.
– The more volume, the more data the better. So already happening. We could talk about it forever. And I’m going to bring you back because I also want to talk about to changes to Google Ads. They keep making all these changes all the time so we’ll do another episode on that. Like DLI.
– That’s my favorite.
– That’s your favorite. See? I knew it’s good. So, I want to wrap this up but one last question for you. What’s your favorite hot sauce?
– Favorite? Ah, so depending on how we’re defining favorite, Mad Dog 357 is a good one.
– No idea.
– So that’s a good one.
– I’m not a hot sauce guy, like a lot of you. I know you are though.
– Anyone who watches the Hot Ones with Sean Evans from First We Feast, which is a YouTube channel, shout out to them.
– So much content out there.
– It’s all over there. They have good hot sauces too.
– On First We Feast. The Hot Ones crew. There is a lot of hot sauces, a lot of hot sauces.
– One day, these guys, bunch of guys here, and gals at WebMechanix, they did a shot of hot sauce just like the hottest sauce.
– It was Mad Dog 357. 357,000 Scovilles
– Scoville, yeah. My goodness.
– I think jalapenos are probably like
– I think it’s like 10,000.
– Something like that. Not 350
– Not even 100,000. It’s way off the charts. But it’s good stuff. It’s fun to do.
– I bet.
– Once in a while.
– Not every day.
– It’s not an every day thing, shouldn’t be.
– Alright, so more than just Google ads. He’s a hot sauce extraordinaire, as well, if you didn’t know.
– There you go.
– And signing off NASA.
– Yup. Space force.
– Space force.
– So thank you for listening, hopefully, this was helpful. Please subscribe. And we’ll bring Tyler back on to talk about Google Ads next time. Cheers.