Posted on

Marty Weintraub

Back to Episode page

Get into action! Download your FREE Checklist

Put the most important tips from this episode to work and take your marketing to another level! Get your free 10 Point Checklist for your next actionable steps.

S: Welcome to Marketing Speak. I’m your host, Stephan Spencer. Today, we have Marty Weintraub with us. Marty is the author of the books Killer Facebook Ads and The Complete Social Media Community Manager’s Guide. He’s a frequent speaker on the International Conference Circuit, speaking everywhere from Jerusalem to Sydney. He’s also founder of the marketing agency Aimclear which in 2016, won three US Search Awards including best large integrated agency. Pretty awesome! Welcome, Marty.

M: Hi, Stephan. Good to see you. Thanks for inviting me to be on your show.

S: Of course, you’re a rockstar so I wouldn’t have it any other way.

M: You were a rock star before anybody ever heard of me.

S: That’s very kind of you. I feel so old now.

M: Just because you feel old, doesn’t mean you’re young. Sorry.

S: Let’s start by talking about Facebook. You are a wizard at Facebook advertising and different kinds of targeting methodologies and things that are outside of the box. You’re an outside of the box thinker. If there’s any outside of the box thinkers in the industry, it’s you.

M: Thanks.

S: What are some of the outside of the box strategies for Facebook advertising that really deliver results?

M: Outside the box. I could not do the inside the box things. You just want to know the edgy stuff, right?

S: Let’s do some fundamentals too. We’ve had other episodes where we talked about lookalike audiences and retargeting.

M: Segmentation and retargeting. Have your folks go down the road of segmentation and retargeting and then coordinating creative or creative is the last mile of psychographic targeting.

S: Let’s talk about segmentation because a lot of the retargeting that people do is really just pretty basic. Focusing on website traffic and not figuring out, “Well, these guys can go into different buckets because they’ve come to different landing pages and ended up opting in for different offers. Because they’re interested in this lead magnet versus this other, I can put them into a different bucket.” We really haven’t gone there yet so we could start there.

M: Let’s spend a few minutes on that and how you could use social filters in retargeting search and search filters in retargeting social because usage of the channels in an integrated way. We don’t see a lot of that so let’s spend a few minutes on that then let’s talk about more radical audience analysis trends, what people are doing. First, most of the time when people retarget, it falls into the few different classes of retargeting. They either take everybody that goes to their whole site and retarget it or maybe they take somebody who gets the first three steps of a funnel, bails out and then retarget to those people or catalogue retargeting where you’re searching for battery on Sears. They’re chasing you around the internet, nipping at you everywhere you go. A four class unit that has the four things you looked on at that store or whatever. Catalogue retargeting, keep in mind that when you’re talking about catalogue targeting, you’re also talking about something like you’re searching for a hotel room in Hayward, Wisconsin because you want to go fish for muskies, an Airbnb hits you with a catalogue ad. Anytime you’re segmenting inventory, that’s catalogue targeting. There are other ways to think about those buckets. For instance, you could drive somebody into a website for the keyword Nikon binoculars sale. You could take the ones that don’t convert at scale, say you have quite a lot of them, or if you lookalike model them. And then you retarget them into Facebook but filtered by people that like to buy outdoor sporting equipment. Make $100,000 in like 300 geeky varieties of birds and then you serve them a retargeting ad in Facebook that has birds in the binoculars. You put the who they are behind the what they ask for. Reduce your retargeting bucket by removing generalizations, etc. Reciprocally, say you have a product that has a very long sales cycle or say it’s industrial, an industrial product that has a 12 to 24 month contract, it’s very, very, very expensive and there’s also really inexpensive versions of it out there. Say it’s like web hosting software and people searching for that are looking for $1.99 a month or $10 a month and you sell something that’s like $5,000 a month and they’re 12 months contracts or 24 months contract, you can’t really use that search inventory because it’s expensive. It’s one of the cases where search doesn’t work very well. You can’t get your return on ad spend. What you do is you identify hundreds or thousands of companies, people who have the exact geeky IT role, maybe filtered by income or other things that would denote seniority, like job title or function that they serve in their company, know it’s the long game, drive them into a webpage where you thought empowering content like a great asset of some sort, cookie them and run search campaigns to those people searching things. Scientific words that surround usage of it including the big query that you couldn’t buy before because it was too generalized for lower cost products. The first thing I said, I brought them in from a keyword then I filtered them into social. In the next one, we brought them in from social and then we filtered them by search with Google RLSA. Just keep in mind that no matter what we buy on the internet anywhere, we’re buying the audience for our ability to segment the cookie fills. Another thing, you have a beehive of keywords, the one where it’s really expensive to maintain impression share like really, really expensive, maybe it costs a lot per click. You know what, the first question I ask in any marketing system is what am I paying for? What am I saying to who for my money? What am I paying for here? A lot of case studies have keywords where a few keywords or a little cluster of them are a high percentage of the expense. Why would we ever take the shit we spend the money on and blend it in with all the cheap stuff that we bought for segmenting the audience? More thoughtful retargeting and using the features from filtered social retargeting of search and RLSA of social audiences. By the way, if they come in from Nikon binoculars and I filter the retargeting and I come back to the site, systems of the future will absolutely do that layer and keep track of the new cookie pool. Picture it as a series of buckets where that’s the first visit to the site then this one is the filtered retargeting then this one spills over and then this one spills over to the combined cookie pool. People already [00:07:28] this data on multiple hops. That’s actual persona building. That’s real persona, a layered cookie pool that keeps in mind what they ask for and who they are.

S: Why don’t you define a cookie pool? I think we sufficiently defined retargeting at the beginning of this episode where you’re basically following the person around or they visit your website. They’re going into your retargeting audience and then can harass the heck out of them for your ad. But using the term cookie pool which is probably not a terminology that our listeners are used to hearing so can you define that?

M: First, no swimming in the cookie pool.

S: Okay.

M: A cookie pool is a fancy way of saying list building. Remember that when you drive somebody to a website from Google or from Facebook say, those two channels, there are two different ways you can keep track of them and build lists. One of them is inside of Google and their remarketing list for search. You can come to what that list is from various methods. Normally, associated with analytics like where they come from or what the tabbing is or something like that. That’s Google’s native cookie building and segmenting tool. Facebook has one of those too where Facebook will track of people that come to your website or parts of your website and then you could use that list inside of Facebook. You could the Google list inside of Google and the Facebook list inside of Facebook. What if it’s every visitor from search and every visitor from social, then we’ve got two really big cookie pools there. 180 days looking back on your website, maybe 20 million, 25 million people, whatever it is, in its most zoomed out sense. If you just retarget everyone that comes to your website and you use Facebook system for Facebook and Google system for Google. If you segment it, you go coming from Facebook, show me all the ladies, or show me all the men, or show me all the ski instructors or whatever. Then, you would simply create a new audience. There’s different ways to technically do that. Most commonly, with URL variables, most commonly it’s web app analytics kind of thing. Then I could say I’ve got my overall cookie pool, and I’ve got my male cookie pool, and I’ve got my female cookie pool. The males and females are subsets of the bigger one. For each way you choose to define your audience, and what I’m prescribing is some inbound keywords, keeping track of what it is, from inbound social psychographic segments, keeping track of what they are. A cookie pool, you see it in Google as an RSLA list that you name, that keeps track of those people over time. In Facebook, you’d see it as a custom audience that has a name, that’s gathering data. It’s really cool. The main point I’m saying here is that take the classic retargeting metaphors that everybody’s been doing forever: catalogue retargeting, funnel bail retargeting, general retargeting, etc. and extend that analogy to what you’re keeping track of as audiences. Cookie pool, it’s so ironic because here’s another one, this is a search only one. You drive them in for that big, hairy, amazing, expensive keyword. Say it’s Nikon binocular sale, we’ll go with that one. I retarget them into search which is RLSA, meaning I run search ads just to those people. Now, the original keyword of Nikon binocular sale is attached to every running future AdWords to those people and I target them for cost nothing keywords searching the Grand Canyon which nobody would ever buy to sell binoculars, ever. But since I have big, hairy keyword attached to it, in other words, I’m only running the search for Grand Canyon overlooks or whatever to people I know who already searched for Nikon binoculars sale. Then I can tailor the words in my search ad to match that as well. I’m keeping the expense of query, I’m spending $0.25 or whatever it costs to do the next one. I’m gaining creative insight which I’m using the best damn Grand Canyon binoculars which makes my quality score 2 for search. As a rule, we don’t see people tuning their systems for these new methods. It’s big. If say you run like $100 million rig or $40 million rig, you have your own [00:12:10] stuff. You use Google tools but you’ve got your own too. You can make a tweak in the system that includes some of these tactics and finding out they’re a couple of millions bucks sometimes. It just makes sense. At scale, it makes sense. At scale, it makes a lot of sense.

S: A similar analogy I would think is let’s say that you are collecting email addresses from different channels. Let’s say you’re advertising in newspapers, in different magazines, and on billboards and things like that, and if you were to keep those silos separated so that, “Well, I’m only going to collect an email list from this particular magazine and keep it with that magazine so I can only use that email list with that magazine and nowhere else. That would be a really dumb idea. Instead, you take all those email addresses and you touch them in different touch points and you collect data from multiple channels and continue to build your CRM with that. Similar sort of model, but now we’re talking about retargeting audiences and not keeping them within Facebook, or not keeping them within Google, or whatever other retargeting platform you’re using. Being able to keep that database as your own and then retarget them multiple places.

M: Right. Basically, you could drive people in so many channels and then you’ve got different capabilities to retarget them in other channels with and without filters. In Facebook, the filters are Facebook ad filters, like Facebook targeting filters so you don’t have to retarget everybody, just the ones who make above 50 and [00:13:58]. In Google, you could retarget everyone or you can retarget people filtered by custom affinity audiences. In Google, you could filter based on their display of targeting and search. You can apply filters. It’s really cool. It’s idiotic not to. What’s really interesting is Google can match up custom affinity audiences and search. They could say, “You could buy this query but it’s only the people who live in a trailer park and want to buy a used Chevy.” Whatever it is, whatever data they have, I don’t know exactly that data, Google could give us so why do you think they don’t give us that stack?

S: I don’t know. Because they want to make more money?

M: Because we could do it now. We could send somebody to your content by buying the custom affinity, or affinity, or topics, or trends, or whatever targeting and drive them to your web page and then I can run an RLSA campaign to those people and it makes me do it in two hops. I’m pretty much confident that Google and it’s engineering freaking might ambition could build out a system. Different in their own tools than from the Google UI. Anyway, it’s not in Google’s best interest. Facebook doesn’t have enough search data across enough categories and subcategories. Google does and they do have the psychographic data. Not the same as Facebook but they’ve got really good data, like crazy data. It’s political decision. I have a book coming out about psychographics soon that I’m co writing with Mary [00:15:35]. We’ve worked together for years and she’s very involved in both of the books that I got to write and co write for [00:15:41]. This was a big conversation. There’s a section in the book about politics of dancing, the politics of psychographics. We could add so much more capability from Google because I’d rather they just charge me $25 for the freaking click or create a new business model that’s an elite subscription model to match up the data. It’s insane. I guess it’s an opportunity for Aimclear because one of the big things we do is go into companies and help them integrate. You noted that we won best large integrated agency in the 2016 US Search Awards and thank you for that. And the best way to be integrated is to integrate. These are integration topics. In the future, machines will just do this for us.

S: Another way to describe it is you should be platform agnostic.

M: Yes. Also, if I were most marketers, I had worked at becoming an ace at plugging in what we learned about creating invariable based snippets into AI content machines. When we get ads approved with our clients, we don’t do ads, we get things we’re allowed to say and snippets ranging from 25 characters to a tweet long, the images and the targeting that goes with that vertical and we can use any of it, anywhere. Ironically, that’s exactly what you need to plug into. A merge of AI tools as well. Let the machines prove the combinations, old school, what can I say to who is what matters the most. Ironic.

S: You have such mind blowing ideas. The concept of dynamic replacement and search ads is probably something that listeners are familiar with or for those who aren’t, it’s the ability to take the keyword that the user typed into Google and incorporate that into the ad so you have some resonance there and people say, “Oh wow, that’s exactly what I was looking for.” You don’t have to write all of these from scratch. But taking it to a whole other level is what you’re describing where you have this pre approved snippets of text instead of a whole ad itself.

M: You could verticalize them by theory too. You could go, “This is the one where we say it’s great. This is the one where we say you die and go to hell if you don’t use it. This is the one that’s the quotes from the tech people that use it. This is the one that demystifies it and this is the bucket that speaks to cost.” Inefficiency, you can get the things approved to say at Aimclear, when we create our ads, we create the top of the funnel stuff and the first landing experience like the key variables, headline image, main things on it because then, you could take the approved things to say. You go, “First, we told them what it does. Now, it’s time to brand call to action and tell them why they’re screwed if they don’t use it.” The first person who defines the top of funnel and the landing experience also does the retargeting background of the retargeting ad and the landing experience for that. We have a switch where we keep marketing to where we drop them into nurture. It’s really interesting. We shifted most of the ad text stuff that we do into people figuring out what to say to who with our clients in a collaborative way. And then, we use modern technology to organize that with the human entering first task halfway to purchase the cost into nurture.

S: Cool. Basically, what you’re saying is that understanding who you’re talking to is critically important and then having these pre approved snippets that have gotten the go ahead from Facebook or whoever and then you can build these on the fly.

M: From the client. The client says it’s okay. It’s okay, we could use any piece of it for anything. We don’t have to ask permission. If there’s an adjective, there’s an approved adjective set. They’ve been checked for colloquial relevance and short. We understand the intersection of few character count and colloquial relevance. We usually look at it from book data in the 1800’s through Google Trends. We want to know how to work. This is really cool because it takes the brand and it makes it an inverse pyramid where it’s balancing on anything you do at the complete control of the words that you are allowed to use wiggles out everywhere that the brand goes. It’s really cool. If you work off a creative matrix like that, a grid, everything’s approved, it’s all the right character, it’s everything that we want to say, and nobody’s going to make up anything. That’s not what we are. This is what we are. This is our brand freaking identity. Modular variable based, channel agnostic, most of the time, approve things to say. It’s so ironic, dude, because it’s like marketing comes down to list building and what you say. I’m judging the European’s Search Awards right now. The category that’s about best use of PR in a search campaign and it’s all about content completely. Like in Europe, PR is content. Creating that unbelievable asset and then major newspaper articles and then 100,000 visitors and 750,000 pounds of sales happened on the digital marketing side of that. It’s really controlling what the brand message is. This is what you are allowed to say. Our company hovers at about 30 people or so and 5 of them are 10 to 30 year PR professionals. It’s really cool. The difference between PR, SEO, modular creative, influencer distribution, nothing.

S: Yeah, that’s all. It’s all merging.

M: You want SEO results, do something notable.

S: Yup, or something remarkable as Seth Godin says in Purple Cow.

M: Right.

S: Worth remarking about.

M: Yeah. There’s a scandal in it. It’s on the front page. Google, you better think of what you’re going to do that’s reciprocally special.

S: Let’s talk about this idea of being platform agnostic or seeing what you can learn from one platform and applying it to another. One nugget I got from a recent presentation, you keynoted at Affiliate Summit recently and you shared this one about using your Google negative keywords on Facebook. Why don’t you go ahead and share that nugget with our listeners.

M: It’s so funny, most interesting things you could do, mostly just some freaked out brain thinking of it someday. That’s out of the box. There’s an excluded bucket in Facebook and you could literally go into your AdWords or Bing accounts. Wherever you have negative keywords, you could just paste them into that box and have Facebook parse them to see if there are any characteristics that would match that in the positive, that would be excluded in the negative. Not all of it matches but a lot of it does. The one that I showed at Affiliate Summit is the 115 million people who like free [censored]. Even when I have a premium product, I don’t like to market to people looking for free [censored]. I like to market to people who are willing to pay for it and would appreciate taking the software for a spin for free or whatever. Whatever your negative keywords are, see if they parse and exclude box in Facebook. It’s so ironic for all the developments in social psychographic world with Facebook, LinkedIn, YouTube, Twitter, Instagram, Pinterest, DMPs, GSPs, they still need words to describe the social segments. They need words which is why targeting that begins with going to, trying, eating out at, spending time with, studying, any of the ing words, or action words like remodelling, like that still can add a layer of intent in social that is advised by 20 years of semantic history SEO, PPC, etc. Search marketers are fantastically qualified to be social marketers because they know the difference. Search marketers know about intent and that gives us an advantage in social psychographic targeting. It still needs words.

S: It’s awesome. You can have I think up to 10,000 negative keywords in Google AdWords and you can use that further for your Facebook exclusions list and figure out a better group of people to exclude out of your Facebook targeting because they’re looky loos or tire kickers, or just poor quality prospects.

M: One thing you could do is you could drive people into your webpage from search and take the people who perform the poorest on the website and run a psychographic analysis on that in Facebook to see who those people are to make them aren’t. Here’s an out of the box thing for you. For years, since we’ve been able to make custom audiences inside of Facebook, we’ve tested theories. Here’s one for instance, you take everybody that went to your website and then you create a Facebook custom audience with it. Simple. Now, say I have 20 million people or 1 million people, whatever it is over 180 days and for years now, we’ve been able to make like we’re going to make a Facebook ad and start with the custom audience and then you can apply other filters. If I have those site touches embodied in the custom audience, then I could start with that as my targeting in Facebook, obviously. Most people are used to targeting that custom audience. That’s what they do. They target that custom audience. If you wanted to, you could add additional targeting and therefore subtract audience size from it. Say I have 10 million people in the custom audience, I could say, “Show me all the [00:26:17].” You have a research tool at no cost from Facebook to make like you’re going to make an ad and add and filter that and go, “Let’s check and see how many there are who are college professors who like astrophysics or whatever, any attribute.” For years, we’ve been able to test that to where we could use theories. In other words, I could take traffic to the thank you page or other success page and I could age bandit, just then make a chart, just by studying acting like I’m going to make a Facebook. What most people miss even in the last couple of years of generation, is understanding that you could take any email list or any cookie pool from site touches to any page or any cluster of pages and you could study it according to theories and make charts and actually see. The point is that in studying it, you have to start with a theory. What about if you took 25 amazing people in an ethical offshore country and have them test for every possible Facebook interest, behavior, and demographic, it wouldn’t be that hard to figure out exactly what is, what those are by various methods. You could just do a report that just listed every Facebook targeting variable for that audience and then you sorted that report by percentage of the original custom audience so 40% are men and 60% are women. I could just show the attributes that were most prevalent. The future of all this is to go past theories. If you think about Facebook, what they’re doing is they’re configuring content based on 100% awareness of all those variables. That’s like our experience is customized in Facebook. Smart marketers are going to figure out hacks to look at every Facebook variable and study traffic to say who they are.

S: This is traffic that is not all within Facebook but it’s traffic that’s come to your website maybe from Google Ads or from SEO and you want to figure out who these people are and how they tick.

M: Which is why the cookie pool concept is so important because if you’re keeping track, I literally could find out the psychographics of my big keyword cluster. I could turn my custom audience from that keyword cluster into a bigger custom audience with lookalike modelling and then compare Facebook attributes line by line, prove that it actually is a 3% lookalike audience and then take note of the lines that Facebook held onto and what they let go as they made up more people with the algorithm. The future of marketing, right now, to run a report like that, various methods, can take between four hours and three weeks depending on how many dimensions it is. Right now, the cloud on our side, Facebook can do that and Google can do that. They house the data so they could do it really super fast. For you and me to do it, it could take weeks to really understand after we’re along the tired, Stephan, the cloud will be fast enough that will buy products that configure content based on partnerships for the company like Facebook where they have an API for that or not there. For now though, I would encourage all marketers to make custom audiences of segmented site touches with the same clever savvy that we applied to retargeting, cross channel, and filtered retargeting and see who they are. Think about what they need. That’s crazy. Remember the days where we could look at organic keywords to a piece of content, look at behavior, and then create amazing content based on what we know works for searches that we know what they are. It’s like that now for social for the marketer that’s willing to get a little busy.

S: Another analogy that comes to mind is that some people who are really in their little box, not thinking outside the box but they’re thinking, “Okay, I am doing AdWords campaigns and I need to do my keyword research to figure out which keywords I should be bidding on.” And they’re using your Google AdWords keyword planner for that because that’s the intended use case. Smart people who think outside the box realize that they could be using Google AdWords keyword planner for SEO purposes and not spend money on advertising with Google. Even though now that Google has hobbled the tool so that you only get these terrible, useless ranges between 100,000 and 1 million search volume per month is the range example that you get. Basically useless unless you’re spending money on AdWords.

M: Google hates SEO.

S: A really nice work around if you don’t want to spend money on AdWords and you’re a consultant like I am is just get clients to give you access to their AdWords account even though I’m not managing their Adwords, they give me access. I’m an SEO guy. They give me access to their AdWords, now I can use the AdWords keyword planner and get the real numbers which are arguably not that real. At least, they’re numbers and not massive ranges.

M: Real enough. For years, Google has been trying to turn off keywords. Now, when Google reports case studies about DSA on their website where they say the portfolios they examine had between like 16% and 20% of their case studies at scale with DSA where Google just makes that stuff up. They’re coming for us. I remember harassing Frederick Vallaeys when he still worked for Google from the audience at SMX West six years ago or something. They tried to search for a Google keyword tool, rest in peace, Marty. They tried to take it away years ago. They’re successfully making up keywords. I like the third party tools just fine. I think SpyFu and SEMrush are really cool now. I think there’s great data in those tools.

S: Yeah, I agree.

M: Most people don’t know that the performance marketing team that works for search could completely float socials tea cup or whatever the float boat. If you work for a search performance marketing, you want to make a really good friend, create a cookie pool for your most successful keywords in Facebook and then tell your social marketers to search that, to task it against every Facebook targeting variable and see what it is. In my boat, I spend a lot of time in the Canadian wilderness. I’m from Minnesota. On my boat, I’ve got really amazing data. It’s the 16.5 fishing boat because it’s a 25 horsepower edge of the wilderness lake. It has the electronics of an ocean going vessel or what it can do. It’s semi autonomous. It’s really, really smart. For one thing, I can either drive my boat where I know there is a lot of water under me or not. It could either have that data or not. How about this? I could head off into the wilderness with no concept of the weather or I can know. How about this one? I can fish where I know there’s fish or where there’s not.

S: Yup. These are all good analogies.

M: Stephan, have you ever seen a case study in your whole life, really think back where if we went, “I’ll tell you what, we’re going to tell you everything about this aggregated bucket sorted by how many of them makeup your audience percentage of custom audience for every Facebook targeting variable.” That’s just who they are. It’s obvious that this is a way bigger gun, there’s already audience profiling and tools that talk about demographic and psychographic profiling. One of the [00:34:38] does that. I would expect that the whole freaking world is going to expect that we could get output from just what that audience is, because you could do it in Facebook, Twitter already. If one were ever to create something that automated that, I would think it would be possible.

S: You’ve done stuff that other people think is impossible in terms of identifying these audience segments, percentages, the different attributes, and we’re talking about using for example search audiences and profiling them using Facebook and then providing these reports to your clients. How big of a deal is that expense wise and man hours? You have some super secret technology to help you accomplish that and you’ve got lots of minions doing probing inside of Facebook. What would be the investment?

M: People are coming to us right now to see who their audiences are. At this time, we’re not in it for the money.

S: What does that mean though? You’re still charging, right?

M: It means that if it’s an interesting case study, then we run the first level without charging to see what it is. Everyone will need more dimensions for it. It’ll cost in the thousands per dimension to layer on. Here’s the other thing, you could analyze by every zip code in America, every overlapped mandering district in America. This gives you empirical geographic data. Somebody should pay us $20 million to freaking do this for some people. Some people say that sort of thing could happen. How about this? Address clustering where when you walk to the door, you just know exactly who they are. Search for political psychographic segmentation. We’ve been thinking about this for years. This concept, like I said, we’ve been doing this for years. Just do the elbow grease to gather the data. Right now, what’s the date today? The day we’re recording this.

S: March 13th.

M: Thanks. That’s okay to say. People know that you record them ahead of time. You’ll share this with your audience sometime in the coming weeks, I would assume. By the time they see this, we may be charging to do that. I’m telling you about it really early. Right now, just assume that humans run it by hand. Just assume that and that they’ll be really well compensated to do research on Google to your website.

S: Okay. Sounds very intriguing and I bet you’re going to get some people contacting you. is your website.

M: My email is

S: Okay.

M: Anybody can email me anytime for anything. We have a super secret psychographic social distribution weekly tip. We’re the only one that never does sales pitches. We just don’t. Just whacked out a couple of tweet long from screen capture interesting stuff. 30% of my new clients in 2016 came from a list that we never market on.

S: That’s awesome.

M: I know. It’s really cool.

S: You add value first and then good things happen instead of trying to get, get, get stuff.

M: Switching to my second cup of coffee for this.

S: Let’s talk about Facebook retargeting based on Facebook video views. That’s something that you had mentioned.

M: Engagement targeting. This is really cool. We’ve been talking about building lists. First party, Facebook list. I’m putting the Facebook pixel on your website where the data is kept track of in Facebook and we’re talking about building Google lists too. It’s first party because it’s on your website. We’re using Google to store the data, and Facebook to store the data using their tools. What’s really interesting is that Facebook has engagement retargeting which means that you don’t have to send them to your website and you could still keep track of the data in Facebook as a custom audience that could model. You could study them. You could study any custom audience, email list, page touches, engagement retargeting. That’s pretty cool because you could upload a video on Facebook, not send them to your website which means it’s cheap to do it because you pay more when you send people outside of Facebook most of the time. You could base segmentation of that Facebook cookie pool. What’s really interesting is that thir party data probably, it’s Facebook users on Facebook’s site. Anyway, you could build a Facebook retargeting list that could be used in Facebook, straight out filtered by how long somebody watches a video on your website or behavior inside a lead gen form. It’s pretty cool, different types of interactions inside of Facebook so that’s pretty cool too because if you want to prove that branding moves the needle, you could engage them in Facebook with a video. Keep track of them there. That would be as opposed to sending them to your website to see a YouTube embedded video and setting a retargeting cookie on them there. The reason to choose the Facebook one is if you want to make the first touch cheap.

S: Right.

M: Really cheap to do that in most countries. YouTube has something like that too. You can run engagement retargeting. It builds a Google remarketing list from YouTube video views without ever having to send them to your website. The catch is that you display retargeting to those people but not RLSA. The work around is to send them to your website and cookie them there and then you can run RLSA.

S: Okay. I’m just realizing we’ve mentioned RLSA a whole bunch of times, mentioned DSA. Some of the listeners don’t even know what those are so we should quickly give them what the acronym is for each of those.

M: DSA is really simple. You don’t give Google and keywords. You just play into the page. You don’t get many creative, and they just make it out.

S: It stands for Dynamic Search Ads, right?

M: Yup. It’s Google successfully getting rid of keywords. It works for very specific case studies. It doesn’t work for everything and anybody responsible tests it to see if it does because if it does work, then you’ve just put a search marketer out of work.

S: We’re all going to be replaced by robots.

M: No question. That’s why segmenting creative is so important. There’s not enough data on earth to make that input shit up right now. There will but there’s not. RLSA, that’s Retargeting Lists for Search Ads. It’s a freaking miracle. Almost every major winning case study I personally see on the paid side included are RLSA. Many of them from social audiences. RLSA, it’s really cool. This is Google’s retargeting pixel on your website or whatever the code is, JavaScript. And then Google cookies the users and then you can divide it by things like URL variables or behaviour, things like that. Basically, what it means is that when you retarget them, you’re not doing it with a banner ad, you’re doing it with a search ad. The additional filter is what someone has searched for. That’s complicated so let me explain. They come to my website and they’re just really busy little beavers on the website. What famous second tier Minnesota school has a beaver as the mascot? That would be the Bemidji beavers. Okay, basically, they’re busy little beavers on the website and so I run a search campaign just to those people. That means that if they search for important keyword, I know that they’ve been busy on my website because that’s the list I made. That’s the remarketing list that I made. By the way, the difference between remarketing and retargeting, Google calls it remarketing and everywhere else on the internet calls it retargeting. It’s the same thing. If you think about what RLSA is, it means you can run a search campaign just to those people, just to the people that came from Facebook, just to the people that almost purchased something but didn’t, just to the people who came from a LinkedIn organic post, just from whatever. If you think about it, it’s retargeting triggered by aka filtered by keywords remarketing lists for search ads. I’m going to run AdWords but not to everybody, just to the people in that remarketing list aka cookie pool.

S: Right. These are people who have a high inclination to buy from you because of the behaviors that you watched and pooled together.

M: Remarketing lists for search ads were revolutionary. Site retargeting was highly generalized before. It’s like by its very nature, remarketing list for search ads makes you identify a frontend audience and strategize for it with words. It’s revolutionary because it was the first technology that just pushed people to filter. There are two filters, double filter which is why it’s so effective, it’s filtered by the who you’re running the search ads to, the cookie pool, the segment. It’s also filtered by the keyword that somebody would select. It’s super cool. If the inbound query is really a money keyword, tons of it jump down ton the next keyword you run, it could be really informational and provide creative insight. That’s the binoculars one where you run them in for Nikon binocular sale and then you retarget them for searches about over Grand Canyon overlooks. Nobody buys Grand Canyon overlooks. Probably, it’s not as effective as Nikon binoculars are. The search ad which has room for text now with these expanded advertising units could save the best explorer, the blah, blah overlook with the best binoculars on earth. And then there could be side links that say Grand Canyon binoculars. The second query costs almost nothing because it’s not competitive. I have to check and see if these are actually contestant or not. The first one costs a boatload and because you’re running the campaign only to the ones who shopped the really big keyword, it’s always attached then you’re just gaining insight for cheap and continuing to market to that big keyword. It’s the big spoon, little spoon keyword theory.

S: That’s awesome. Let’s keep diving in here on retargeting and let’s do more on Facebook lookalikes and all that. Another thing that I found really powerful that you’d mentioned your keynote was that you could do Google Ads based on your YouTube retargeting list. Let’s talk more about that.

M: That’s a really interesting one. About once a year, some really sharp millennial clips my rings in a presentation so I’m right up to speed up this one because I was wrong about it recently, suddenly wrong. If you make a YouTube engagement remarketing list without going to your website, you can do the slay retargeting to those people but not RLSA.

S: Okay.

M: The workaround is to drive them to your landing page where you embedded the YouTube video and set the Google thing there and then you can run a search campaign to those people. The engagement unit inside of YouTube where they don’t come to your website, you could display retarget them but not run search campaigns to them. Maybe that will change, maybe that will have changed by the time this comes out. For now, that’s the way it is. If you send them to your web page and embed the video, you get many of the benefits of YouTube and you have complete control and you could set a Facebook pixel too. I guess that my first choice would most often be to embed the YouTube video on my webpage. I always would rather have first party data than third party data because I control the segmentation there from. If you want to run a search campaign to a YouTube video view, embed the YouTube video in your landing page, run RLSA campaign to those people. Isn’t it amazing how many solo commutations there are? That would be opposed to uploading the video to Facebook, not letting them go anywhere, building an engagement list there. I guess if you ask me that, it’s the same limitation. I can’t run a search campaign to the people who stay inside of Facebook. I can’t run maybe a search campaign to Google engagement retargeting because Facebook doesn’t do a very good job of selling us search.

S: Yeah.

M: They don’t. That would be amazing if they ever. I just think the search is too quirky at this time. It’s too social. I put the video in Facebook and I can’t do search retargeting but I can do Facebook targeting with that. I guess that my first choice will always be to send them to a website that I own so that I can control the data. My second choice which would be based on cost or patience because I would nurture the audience for cheaper in channel because I could do a Facebook video only retargeting at Facebook with that thing that sends them to my website from there if I want to. In fact, Larry Kim. You know our friend Larry Kim?

S: Yup, from WordStream?

M: Yeah. As long as a year ago, at SMX West, Larry Kim was saying, “We know it takes three touches, let’s be looking for really cheap ass video units in channel and send them the next time.” It comes down to how much you’re willing to spend on the touch for first you know you have to make. It affects cost in some cases. One compromise would be to make it a Facebook video unit only but put a Bitly link in the caption so the most [00:49:27] customers would go to your website wouldn’t cost as much as making the whole unit send them to your website. Anyway, isn’t it amazing how just like nitty-gritty where somebody goes, “Oh, where should I post my video, Marty?” Okay. What do you want to do? You want to brand for cheap. keep track of them and drive them to your website later? Well, use Facebook for that or maybe YouTube engagement retargeting with display. Or do you want to spend more money? Get them out of social now. Set every kind of cookie and continue to operate on them with search ads just to those people, RLSA, that cost more money.

S: Yeah. This is just mind blowing stuff. A lot of the listeners, I’m guessing, don’t even know that you can do retargeting in Facebook just to people who have watched the Facebook video that you uploaded and say 50% or 75% of the video. I want to target those people. They don’t even know that. We’re just blowing their minds right now.

M: Where should I post my video? I don’t know. What do you want to do? How many steps? How much do you want to spend? What’s really interesting is that most folks also need street cred on their social profile. Street cred where it doesn’t look like something that got spun up in the basement of a kid in 1998 where there is 11 friends and nobody cares. It’s ugly and nobody comments. There a certain amount where people find things in Google search results see if the company they’re dealing with has any cred or not. It’s a perfectly reasonable thing to put into Facebook. You get internal engagement which is street cred. I recently saw a case study where what they did, they ran very focused social ads just to a really narrow geographic area in a Facebook page that didn’t have a lot of cred. They kicked off the ad by opening it up to national for the first few hundred dollars, got the engagement and then tapped it down so when they brought that thing into their local marketing, it was packed with cool [censored].

S: Lots of comments, lots of likes, shares, and all that sort of stuff.

M: Yeah, that is so brilliant. That’s really smart, man. [00:51:43] few hours to everybody. Also, it spends money in a way that Facebook likes. It helps you. It’s always good to come out spending money with efficiency no matter where you’re buying it. Anyway, so, “Where do I put my video, Marty?” Well, I don’t know, you need street cred. Can you take more than one touch? How much do you want to spend? Can you even believe the nuances of it? In 2008, I got to speak on a panel at Search Engine Strategies New York. I spoke in Portland that year which was important to me and then at SES New York. That infamous give it up session, to me it was infamous.

S: That was fun. I remember that.

M: Star [00:52:29]. It was like Chris German, and you, and Rand, and Michael Grey, and Todd Friesen, and Danny. It was amazing.

S: Yeah, that was a lot of fun. We were throwing out some secret sauce.

M: I was on that panel in 2008. It was one of the few social media panels at Search Engine Strategies New York, 2008. There were a few things going on. This was a Facebook panel. There are four people in the panel. It was 2008 and they all did radically different things. There are people there who are making a meal out of Facebook groups at that time. There’s somebody that was there for optimizing the organic newsfeed. It wasn’t even really a newsfeed then, kind of it was. I was like the Facebook ads guy. This is unbelievable. There are four different things in this channel and there was a room that held about 200 people and there were about 12 people in the room. All the old salty SEO dogs are out there. In five years, it’s going to be all their jobs. They’re going to say they were at the session, not outside making fun of us. You’re not many but maybe you’re the right people. You are the evangelical preconceived. Go for it and be badass. Now, it’s so nitty-gritty that if we were going to do a panel on video placement in the social world, we’re going to do like a 90 minute panel with 4 people on it arguing about where to place that, where to use it for, who does what topic, and it’s that granular.

S: Amazing. We’re running short on time here. I want to address one more topic that you had covered in your keynote. That’s lookalikes of thank you page visitors. I love that one.

M: Joke was on me. I’m a targeting creative director tech because I like to, when possible, check in on targeting when thinking of it before they send it to a client. Aimclear is completely self sufficient without me, I just like it and I sometimes bring some ideas that are useful. I’m a marketer that’s available to the mix. I’m going, “Maybe we should blah, blah, blah.” We’re going to model the thank you page before you do anything. I’m going, “What are you talking about?” They just explained that unless there’s a real shortage of qualified customers or sales in the marketing machine, it’s almost our first look to see if modelling success traffic. Facebook has at least thousands of psychographic attributes that you can target by, at the least. Some are obviously available to marketers and some you don’t know aren’t obviously available to marketers by API. Facebook doesn’t really release much information about it but if we had to guess how that worked, we would go and look at every attribute, weight it somehow and then find other people who are as close to a match as possible in as many areas as possible. It seems like a reasonable super computing thing to do. It makes a lot of sense line by line to every piece of data they have or a partner data that they have. It stands to reason that with 1 point whatever billion people in Facebook and huge numbers in America, that they could do a pretty damn good job of finding people that are alike. Now that we’ve seen visible lookalike modelling where we have analyzed it line by line, we could see how that works to some extent. That’s part of making audiences visible and see how lookalike audiences play as they extend. Like they play guitar, a lot of them play guitar and you go, “Whoa, a lot of them are into playing guitar.” Because if somebody says to you, “Oh, I’d like you to do a competitive analysis for SEO.” What percentage of the time do they come back to and the people that are worried about competing are not true algorithm competitors. You know what I mean?  They’re not really after the same keywords. They’re not winning. Like in algorithm, a competitor. Somebody buy what keywords they buy. How about that? Are they algorithm competitors or organic success using third party data? I think that we could pretty well do algorithm if we [00:56:49] lookalike. Oftentimes, we test it before we test top of the funnel targeting. It’s that powerful. It’s really powerful.

S: Basically, the bottom line here is the lookalike modelling that Facebook does that we can all do when we’re advertising on Facebook, we can target people who are similar to our custom audiences of our email list or buyers, people that touched our website and all that and we can target people who specifically gone to the thank you page, opting in for our lead magnet, or filling out an inquiry form, or buying the product.

M: Right. Yeah, for sure. It’s so powerful. Analysis of who came to your page using Facebook segments is so strong that you could drive someone in for bucketed targeting Facebook and then see which element for optimization, see which element resulted in the sale. Lookalike modelling, it’s kind of idiotic not to test it on your success page touches because here’s the thing. It’s like fishing. We could either test using all of Facebook’s data to make up people who are exactly like our customers and see if that works or not.

S: Go fishing where the fish are or we could just randomly throw our line out and hope something happens.

M: It might not work. It doesn’t work everywhere. You know where it’s kind of hard? Some things that have very short sales cycles like until next time. People who want tickets to go on sale where they’re gone in 10 minutes, or half hour, or an hour. It’s not necessary to model a thank you page. Things that have really short cycles don’t work well for this unless it says next time, Stevie Nicks comes to town solo and then you’ve got that audience to work with.

S: Got it. Well, we’re out of time. I wanted to send people to your website, and your email, and your Twitter, and all that. Let’s give them all your links and everything.

M: On, where you’ll find our blog. I’m I wonder where we got the dumbass email just like that. Email me if you’d like to be on the super secret psychographic social distribution weekly tip sheet or talk about modelling your audiences or talk about whatever. Please say that you saw me on Stephan’s show and what the date was that you saw it and keep in mind that things change in this industry a lot. If I’ve talked about a capability here or a methodology, then it will likely change in time somehow. You can see me speak upcoming at Napa Summit, at SMX West, at Ungagged in London where we’re doing a full day psychographic and social PPC workshop from an integrated perspective. Zenith Marketing Conference in Duluth, Minnesota where we’re also doing a full day workshop. I’m also keynoting Search Summit in Sydney and speaking a week later in Adelaide. I’m doing a workshop there. You can find the exact speaking schedule on our website. I’d love to meet you. Even if this interview was so technical that you could barely stand it, you’re in fine hands with my friend Stephan. He’s a really famous marketer. He thought of many things that are things that I do extrapolations. I’ve been following you for a lot longer than you’ve been following me.

S: Thank you. Thanks for the kindness.

M: That’s powerful when you know it’s true.

S: Thank you, Marty and thank you, listeners. I really hope that you take some of this seemingly esoteric knowledge about Facebook and Google and apply it in your business. There’s so much opportunity and this is ninja stuff. Re-listen to this episode if you need to. Go through the transcript which is on The show notes of course which will include links to different resources that we mentioned in this episode. All that is available at and go to to check out Marty’s awesome agency. Thanks again and we’ll catch you on the next episode of Marketing Speak. This is Stephan Spencer, signing off.