Practical Uses of Machine Learning for Marketing

Fine Point Grey Michael King, iPullRank

Practical Uses of Machine Learning for Marketing

(40-minute podcast)


Everyone’s talking about machine learning, but how can it help you as a marketer? Mike King from iPullRank joins us to talk about his passion for blending AI with digital marketing. We also get an inside look at his transformation from hip hop artist and entrepreneur to performance marketer. Our host, Maureen Jann also covers the latest news, including YouTube’s advertising challenges, LinkedIn jumping into trending news headlines, & Facebook’s latest mobile ads venture, Collection.

Fine Point iTunes

Fine Point Stitcher

Fine Point Google Play

Featured Expert:

Mike King, Managing Director of iPullRank


Guests and Experts


Mike King, Managing Director of iPullRank

Bio: One of the hardest working, and most unapologetically bold, technical marketer’s in the game, Mike King is single-handedly changing the job description of marketers today as a groundbreaking force in the Digital Marketing world. An artist and a technologist, all rolled into one, King recently founded boutique digital marketing agency, iPullRank in order to allow his creative approach free reign. Self-taught, self-possessed, and famously self-assured, Mike’s potent levels of moxie drives the increasingly multifaceted modes he can operate in to deliver on a cross-discipline strategy.



Maureen Jann, Director of Marketing, Point It Digital Marketing

Bio: Maureen Jann is a veteran B2B marketer whose career in Digital Media has grown up with the Internet. A self-described jill-of-all-trades, Maureen has elevated creative problem solving to an art form and enjoys the daily challenges of driving business results in unexpected ways. Her skills as an entrepreneur, content marketer, creative director and passionate people manager set her apart from the pack. Maureen has worked in every corner of marketing making her a skilled tactical resource as well as a strategic partner.  Recently, she was the captain of the marketing ship for an award-winning professional services firm and is currently creating a content marketing strategy for Point It, a digital marketing agency.



Tim Mohler, Sr. Marketing Manager, Point It Digital Marketing

Bio: Tim Mohler is a multichannel marketer with experience building campaigns for travel, CPG, food, beverage, and technology companies reaching both B2B and B2C customers. He’s passionate about building marketing experiences & partnerships that are relevant to a brand’s message, exciting for the customer, and most importantly deliver measurable results. At Point It, he develops digital, social and content campaigns as well as managing on a day to day basis.


Maureen Jann: Welcome to Fine Point, a weekly digest of digital marketing updates. Each week we’ll feature industry experts to talk through what’s happening in digital marketing. I’m Maureen Jann, the director of marketing at Point It, a digital marketing agency here in Seattle, Washington, and I’ll be your hostess. I have Tim Mohler with me, our senior marketing manager, and he’ll be chiming in throughout our podcast in the Point It studios, also known the conference room.


I’m delighted to have Mike King here from I Pull Rank. Today we’re going to talk to him about his current passion: practical uses and machine learning for marketers. So you’re in town to speak at a conference tomorrow, Digital Marketing Summit.


Mike King: Yeah, Digital Summit.


Maureen Jann: There we go.


Mike King: Yeah. They actually have it all across the states, so it mostly like secondary markets and I speak at all of them. So as soon the year starts, it’s like, “Oh, my year is pretty much full with talks like twice a month.” So yeah, it’s pretty cool.


Maureen Jann: That is cool. I like that they’re here, it makes it very handy for us to have conferences in town that we can walk to because it’s literally like five blocks from here. We get really get excited about local conferences. The next one I think is, Mark SMX Advanced is here. And we’re doing a cocktail hour. So if you come, you should come to the cocktail hour, because we’re going to have a really good time.


Mike King: I mean am I going to say to free drinks?


Maureen Jann: No. Only a fool would say no to free drinks.


Mike King: Fair enough.


Maureen Jann: Well we’re get deeper into the machine learning a little later on. But right now we’re going to talk headlines. So right now my network is talking about Google, LinkedIn, and a Facebook, the big three. I like that we jump right to the deep end on this. No middle grounds for us. So this headlines is from Ad Age, and it’s J&J, Johnson & Johnson and Verizon join major marketers to spending YouTube advertising. Bad week for Google, lots of rough stuff going on for them this week. Advertisers have been jumping ship from YouTube due to ads and proximity to hate speech, terrorism, terrorist propaganda and other objectionable content. And considering our current political …


Mike King: Climate.


Maureen Jann: Yes, thank you, this is unfortunate. I’m looking forward to having this not be the top of our news, but in the instance of Johnson & Johnson, their digital video budget is a substantial part of their global advertising spend of a total of $2.4 billion. No big deal. That’s cool.


Mike King: Drop in the bucket for them though.


Maureen Jann: Right? It’s true. It feels really big to me though.


Tim Mohler: So there’s a prediction that somebody did over the weekend that’s it’s going to cost them $750 million dollars and they will actually have to have an adjusted guidance for this quarter.


Maureen Jann: What will cost them that?


Tim Mohler: The total lost ad revenue at this point.


Maureen Jann: Oh, YouTube?


Tim Mohler: Yeah.


Mike King: The things is though, it’s funny that this has just happened. Your ads were being shown against this content the whole time, so what is it that made somebody finally notice? Were people just asleep at the wheel the whole time? I don’t understand.


Maureen Jann: Now they’re all sensitive about that. They’re a little raw due to the unfortunate circumstances. Well I mean, happily it takes a trigger sometimes to get people on top of the game, right? And so now YouTube is taking steps to give advertisers ultimate control, and I’m saying that with air quotes because who doesn’t love the word ultimate control, and transparency over where ads run. So we’ll see what that looks like from a dashboard standpoint. I’m curious to see how that will hash out from the reality of ultimate control versus the perceived ultimate control.


Mike King: It just seems very difficult to actually implement. Who’s to say that a video might start with Care Bears and all of the sudden they jump into hate speech. There’s no way to really effectively do that at scale. So I find this to be more of them posturing and being like, “Hey, we fixed it for you,” to kind of stop bleeding money if anything.


Maureen Jann: Or maybe they’re trusting people to tag their videos correctly …


Mike King: Which isn’t going to happen.


Maureen Jann: .. This is about spaghetti and chicken and Care Bears, but no, it’s actually about terrorism. So yeah, well maybe there is, is there truth in advertising? Is there?


Mike King: Imagine that.


Tim Mohler: I think they’re walking a really high risk line especially if you think about this, this is Google. Google has been going through this with Google News now for many, many years. Everybody, as soon as Google News came out really learned how to game Google News, right? So as soon as we went to automated content it became something to game. So whether it’s report this content, they really have a fine line to walk. They’re already flagging some content that is … they were flagging an author who was covering abuse of women, so you have to be …


Maureen Jann: Wait, abusive women or abuse of women?


Tim Mohler: Abuse of women.


Maureen Jann: Different.


Tim Mohler: He logged in this weekend and he had a little yellow dollar sign on the platform telling him his content wasn’t advertiser friendly and risking him turning him off essentially. So they’re going to have to be really, really careful with where does editorial content and freedom of speech end and where does their obligation to advertisers and all that money that keeps them running go. So I’m very curious to see how this all pans out and how they navigate it.


Mike King: Yeah, but ultimately, even if there is a solid fix, I think at best it’s just going to be like SEO where it’s like something still will get through. It’s never going to be this ultimate control like you said because there’s no way to effectively segment the content 100 percent.


Maureen Jann: I want to say that ultimate control in like a booming scary voice. Ultimate control.


Mike King: That’s for post, we’ll do that in post.


Maureen Jann: I think we should just leave that in.


Tim Mohler: Now post, I’m going to have to edit it out.


Maureen Jann: So this is said. Leave it in. They like to know what’s going on behind the scenes. Well in that same vein, LinkedIn is trying to do trending in a, I actually love this headline, it’s from Wired, LinkedIn tries to trending in a non-disastrous way. They’re comparing it to Facebook, which Facebook has had so many challenges around their … so to avoid the fake news drama that Facebook ran into, LinkedIn has chosen to have business journalists from outlets from Reuters and the Wall Street Journal to handpick the first few stories in their trending feeds.


So no lie, I didn’t even know this was a release. And there’s like, “Oh, what’s that thing over there,” and I clicked on it and I was like, this is not awful. That’s how I felt about it. It was not awful. So I felt solidly about that. Have you guys seen them on LinkedIn? Have you seen this yet?


Mike King: Honestly the only thing I look at LinkedIn for is to see who looked at my profile.


Maureen Jann: I do too.


Tim Mohler: That makes it sticky.


Maureen Jann: I’m a total voyeur.


Tim Mohler: If you ever want someone to look at your profile first look at theirs and it’ll just keep ping ponging back and forth until you feel like you’ve met in person.


Mike King: Yeah I mean there’s apps for that, just to look at people’s profiles, and it ultimately yields people adding you because they’re like, “Well this person was interested in me, let’s be friends or whatever.” Its always been kind of interesting to me that the other social networks don’t have that feature, because people spend so much time on Facebook stalking each other. You might as well see who’s looking at you.


Maureen Jann: This is just something I’ve never really done. I’m a bigger stalked on LinkedIn then I ever am on Facebook. I don’t know, for me it’s different.


Mike King: So for LinkedIn, largely what I’m doing is research of some kind, whether it’s like who are some prospects for us to work with or who is this person that applied for a job. Things like that. So the newsfeed in LinkedIn it’s superfluous to me. Every once in a while I’ll post an update that, “Oh we’re hiring,” or whatever, and it will get a bunch of likes or whatever but it doesn’t really generate a conversation. So it’s like Twitter is way better for me for that or even Facebook in that regard. So yeah, I’m glad that they’re trying to do stuff and the UI looks a lot more like Facebook these days, but it’s not really doing anything for me.


Tim Mohler: I feel like it’s interesting that you mention Twitter, because Twitter’s been doing this as well, trying to cover especially timely news. So whether that’s news, covering what’s going on on the floor of Congress or live sports in particular, and I have liked it because Twitter’s content has been very, very tightly curated. They’ve got great content partners, and I do occasionally open it up and look at it because in part I’m on Twitter all day anyway for business reasons, let me be clear.


Maureen Jann: Oh sure. We know about you, Tim.


Tim Mohler: So I feel like this has opportunity for LinkedIn to be a source of business news, because everybody always has Bloomberg up on the screen at many companies, and I know they’ve been trying to become a content platform where it’s not just salespeople and recruiters that are on there all day long, but business people have an excuse to go on there without feeling bad or looking bad that they’re looking at LinkedIn in the middle of the work day and are you looking for a new job sort of thing.


Mike King: I think it’s going to have to start with me not getting a 500 intros or requests from people I’ve never met.


Maureen Jann: Right? What is that about?


Mike King: It’s just so strange. Like the culture of LinkedIn is just not conducive to what you’re describing from my perspective. And also, the posts on LinkedIn tend to be longer form which also doesn’t lend itself to that stock ticket environment that you’re kind of describing whereas Twitter does. It’s far easier for you to curate things, like for instance you can make lists and stuff like that for Twitter, so to your point, if I want to know the bite sized things that are going on throughout the day, Twitter’s just better.


Maureen Jann: I’m surprised that LinkedIn hasn’t picked up on that whole list concept, like organizing people in LinkedIn is a nightmare.


Mike King: They have the groups or whatever, they have lists if you’re on the sales side or on the recruiting side, but as a regular user, not really.


Maureen Jann: And it feels like that’s one of those things that I didn’t realize the power of on Twitter until I started really digging into it, and now I’ve got like 25 lists and they’re very specific, they let me filter and now if I have, like we have a Tweet chat coming up at the end of each month, and I want to reach out to people who are interested in Tweet chats and now I have a Tweet chat list and I use that. So it feels like you could take that idea and double click on it in LinkedIn and be able to actually use LinkedIn, but right now it’s like if I have to see one more overly beautiful human being touting their accomplishments I would be okay.


Mike King: So you’re saying you don’t like my profile?


Maureen Jann: I think you look extra gorgeous.


Mike King: Wow.


Maureen Jann: Because you’re my guest, and we still have an interview.


Mike King: I feel like the thing for LinkedIn, the obvious feature that they’ve just not done is become a CRM. Like every CRM is using your data, why don’t you …


Tim Mohler: I know that’s common. I worked at Microsoft, that’s common.


Mike King: Now through Microsoft I suspect it’s going to happen.


Maureen Jann: I think you should name it and give it a TM right now, since you have said you might as well just own that. No, it’s definitely … so the other stories they have, actually one more thing before we move on. The Twitter, I don’t know how you actually end up on Twitter. Like I always use Tweet Deck because it’s way more efficient and I never see that trending news. Never.


Mike King: I see it on my phone. I do use Tweet Deck on my machine, but I use the regular native app on my phone.


Tim Mohler: See, they should be more like LinkedIn and just not open up the API.


Mike King: Well Twitter’s had its own API issues.


Maureen Jann: It sure has. Isn’t that adorable? So the rest of the stories in their non-disastrous trending list are picked my algorithms showing individualized links based on connections, our connections, who I follow as a user and then the popularity of that particular article. So anyway, it’ll be interesting to see how they fine tune this algorithm. Hopefully, it will continue to be non-disastrous.


Mike King: That’s the KPI?


Maureen Jann: Good point. Yes. Okay, so our next article is Facebook debuts collection mobile ads aid shopping. So I had to read this article four times because I didn’t understand what it was saying.


Mike King: That’s a horrible headline.


Maureen Jann: It’s like scrambled eggs. So collection is a new mobile format designed to drive product discovery and sales in a more visual fashion. Basically it’s cards. It’s Twitter cards.


Tim Mohler: It’s Pinterest, is what it is.


Maureen Jann: It’s Pinterest. You’re right. It’s closer to Pinterest. Only with far more viable purchasing power, because Pinterest is brutal when it comes to their ads, I struggle. It’s always usually just a library of pictures which are not actionable. That’s how I find Pinterest.


Mike King: I struggle with Pinterest as a concept. I just can’t ever get into it. If I’m looking for like an outfit or something, I end up on somebody’s Pinterest board, I’m like okay, this is valuable, but any other time, I’m like why does this exist.


Maureen Jann: I use it for all kinds of things, so I can give you a list.


Mike King: Really? Hit me.


Maureen Jann: Okay.


Mike King: Give me a Pinterest board of Pinterest boards?


Maureen Jann: Sure. Sorry, but yes. Yes I can give you a list. So my list is, so when I work with clients on visual stuff like graphic design we collectively build inspiration in a board.


Mike King: So like a mood board on Pinterest.


Maureen Jann: Yup.


Mike King: That makes sense.


Maureen Jann: So also we put color palettes together there because they have this huge color palette thing which is awesome. So that’s one thing. I also do it for any home project because they have such a huge collection of home inspiration and then also colors go into that too, right, because you can kind of build it out. I have also outfits. Like I do wardrobes, like business wardrobes, casual wardrobes, all that good stuff, and it’s just inspiration because I can’t buy anything. I get so mad about that sometimes.


I also am an avid super crazy shoe lady, and so I collect weird shoes there to help make sure I’m buying things that I love. So I have one board that’s only for blue shoes. It’s real. I actually have that.


Mike King: As someone who is currently wearing blue shoes, I respect that.


Maureen Jann: Shut up. Oh excellent. I almost wore my blue suede hiking boots. No I’m kidding. They were fabulous. Anyway, moving on. Ad units take consumers to a purchase page on a business website or an app. It’s considered a display ad unit.


Tim Mohler: To me this feels like, I feel like a lot of this stuff LinkedIn as well, there’s a lot of me too going on in this industry. For LinkedIn it’s not having an overall strategy and just cramming in anything that has worked for anyone else to make this thing work monetarily.


Maureen Jann: Well, you have to test it.


Tim Mohler: Facebook, I understand this. I feel like this makes a whole lore more sense on Pinterest where people are gathering visual content and you want to see that visual content.


Maureen Jann: But it could happen. It could happen.


Tim Mohler: Maybe. If you linked it to Pinterest, because who’s going to gather that stuff on Facebook? I wouldn’t. I don’t want to see it. I would unfriend people instead of seeing all their product purchases.


Maureen Jann: And there’s something really fascinating about adding things to a Pinterest board. It feels like shopping but you don’t have to spend any money.


Mike King: So it’s like you’re making this persistent cart basically.


Maureen Jann: The world’s biggest cart. Yes. Categorized by recipes and shoe colors. But yeah, you’re right. It is probably me too-y.


Tim Mohler: They’re testing a whole bunch of different things, they’re also testing the messenger ads right now. There’s a lot of different … Facebook is just testing a whole bunch of different things.


Mike King: They’re like how do we keep squeezing blood from this stone, basically.


Tim Mohler: But as a brand it’s absolutely phenomenal. Everyone’s there.


Mike King: Sure.


Tim Mohler: This is one of those things where some things will stick and some things will probably morph into other opportunities.


Maureen Jann: Facebook messenger is making me want to kill myself right now. It’s so annoying.


Mike King: Why specifically?


Maureen Jann: Have you pulled it up lately? You’re like where’s the people? How do I write you?


Mike King: It gives you like your top five people you talk to, then the active conversations and it’s like the people they think you’re going to talk to next.


Maureen Jann: And some emotions.


Mike King: It’s weird.


Maureen Jann: It makes me crazy. I just had an exchange earlier today trying to book a Tweet chat for one of our experts, and I was just like, “Why won’t you work?”


Mike King: The whole messaging world is very me too-y, of course. But it’s also just getting really complicated for no reason. Like all I need is a list of names and to be able to talk to these people, that’s it.


Maureen Jann: It really is.


Tim Mohler: My iPhone, they introduced this whole graphics and balloons and you can color and write and fireworks. I can no longer talk to my wife who is also on an iPhone without turning off iMessage. So I’m not sure who’s using all these features, but it is absolutely, yeah … across all platforms I’m just like …


Mike King: Yeah but like how do we make messaging in general more like Snapchat?


Maureen Jann: Why don’t we just use Snapchat? Let messaging be messaging.


Tim Mohler: It’s instant messaging. It works. Let’s just keep using it like it is.


Maureen Jann: It used to work. Let’s just be clear.


Mike King: And then it’s interesting to me when something like Slack comes about and it’s so popular when it’s not that different from anything else except for the fact that you can connect everything to it.


Maureen Jann: I love Slack.


Mike King: I love it too, but like why is it so popular?


Maureen Jann: I think for me, I like contextual conversations. It makes it very easy to collect all of that conversation, the files, everything in that one spot, whereas the rest of the world does not work like that, right? Like your email, you’re searching for crazy and I just want to find all the things that have to do with my upcoming piece of content, I can just go to that channel and that’s cool. But email is such a mess.


Mike King: Again, I love Slack too, but it’s not much different from what else is out there.


Maureen Jann: You’re probably right. It feels different.


Mike King: It looks great. The UX is fantastic. It connects to anything. But I don’t understand how it came in and just supplanted everything.


Maureen Jann: I don’t know. But if Slack was a person I’d probably hug it.


Mike King: Yeah. We’d probably have done it.


Maureen Jann: So as for collection, and we come full circle, e-marketers predicting success with a predicted jump of 32.1 percent or over 50 percent of the display market. What was that?


Mike King: Based on what?


Maureen Jann: Based on the … oh, that’s a good question. I think I may or may not have had this same thought when I was writing this but it was Friday and I don’t remember. So, oh wait, the next one. E-marketers also predicting a larger increase in digital ad spending of $83 billion in 2017.


Mike King: Just overall?


Maureen Jann: Overall.


Mike King: Okay.


Maureen Jann: So I think maybe they were doing some extrapolation. But you know what, I bet se-marketer will just give me a ring when we’re done with this episode and they’ll just …


Mike King: Hey guys, if you read it, it says it right here.


Tim Mohler: Well no ad dollars could go into Google, so suddenly its got to go somewhere and Facebook is the other option for marketers, so there we go.


Maureen Jann: Although they have their own as we have talked in prior podcasts, their own share of very unfortunate problems. Hello video advertising.


Tim Mohler: I think just that mobile space for Facebook, they’ve always had trouble monetizing mobile and yet most of their traffic growth is coming from mobile, so for them I think this is an absolute necessity to figure out how do we get things in front of people on mobile. Honestly, one of the reasons that I use Facebook on mobile is because there’s less advertising and less clutter. I almost never use the website anymore because it’s just too cluttered.


Maureen Jann: And they didn’t do a good job with their mobile ads. Maybe this is going to improve the Facebook mobile ads. That’s my hope. Hope springs eternal.


Okay, let’s talk to you now, Mike. Let’s talk about you.


Mike King: Let’s do it.


Maureen Jann: So, we have Mike King here today, and he’s the founder of I Pull Rank and regular industry speaker which we heard about apparently you do all the Digital Summits. So there you go. Now we know. Mike is the father of an eight month old, which god bless, hopefully you’re doing okay. As a mom of a four year old I understand it can be a little challenging that first year.


Mike King: I definitely sleep more on the road than I do at home.


Maureen Jann: I remember that. It was really exciting going to a hotel room.


Mike King: Right.


Maureen Jann: You’d be like, “This is so quiet.” So you also enjoy talking about cutting edge marketing practices and you’re the owner of


Mike King: I am.


Maureen Jann: Cool. And that’s an online store that sells hip hop related music, clothing and entertainment. I poked around there a little bit the other day. It was really fun.


Mike King: Thanks.


Maureen Jann: Tell us a little bit about that. How’d you end up in the business of hip hop?


Mike King: Sure. So before I got into marketing, actually me getting into marketing was a result of making music. I’ll tell you more in a second. So I went to school for computer science, didn’t really feel like I was learning anything, and I was like, “Hey, I can go back to school at any point. I can only be a rapper until I’m like 30, right?” So that’s what I did.


Maureen Jann: I think somebody should tell Flavor Flav that. Just saying.


Mike King: Good point.


Maureen Jann: Go on.


Mike King: So yeah I was putting out records independently. I was signed to a subsidiary of Sony, like the indie subsidiary, and basically I did music, played all over the world. I played a lot here, actually.


Maureen Jann: Oh really.


Mike King: Yeah, there was a lot of events I was involved in in Seattle. There was like this MC battle that was pretty big at the time that was … there used to be a place called, I think it’s called I Spy downtown.


Maureen Jann: Yes. The cyclops?


Mike King: I don’t know.


Maureen Jann: Okay. Go on.


Mike King: But it’s downtown. It’s not there anymore, and then like the finals for the battle was at UW because the folks that were putting it on had been students there. And so I used to come out here a lot to play shows after that as well because I had kind of a name from getting second place in that battle.


Maureen Jann: Oh that’s fun.


Mike King: But played a lot of shows all over the world and then in 2006 I got into a bike accident and being that when you’re a rapper you don’t have health insurance, I had to get a job to pay my medical bills.


Maureen Jann: Oh, sad.


Mike King: So yeah, no Obamacare then. So the first place that hired me was an SEO agency, and so I would keep the job until somebody pissed me off and then I’d just go back on tour, and then after a while …


Maureen Jann: I wish I had this when I was young. That would’ve been so handy. I would’ve been like, “You, I am so tired of you, I’m going to play professional hockey.” I didn’t play professional hockey, just for the record. That didn’t happen.


Mike King: So after a while, I was like, hey if I stick with us maybe I can make some real money. So I got a job at Razorfish in Philadelphia and it just kind of snowballed from there. And now I own my own agency. Oh yeah, back to how I got to hip hop. So recently,, its been a store, its been around since like ’97. And they posted that they were about to go out of business. They weren’t do well, it’d been like a downward spiral. And I took a look at the site, I was poking around and I saw they got hit by the panda algorithm update in 2012 and never really recovered. I’m like, so this is an SEO problem, I can fix this. So I reached out to them, I was like, “Hey guys, is this for sale?” And there was a lot of back and forth, a lot of lawyers talking and all that stuff and we got it done. And now I own it.


Maureen Jann: Nice.


Mike King: So that happened in early January and we’ve just started rolling out a lot of the new things to it, and we’re rolling out an Amazon prime type program and that’s been picked up a lot by the subscriber base or the customer base pretty quickly. Like we rolled it out two weeks ago, we already got a lot of sign ups for it. We’re doing a twentieth anniversary concert pretty soon. So there’s a lot of cool things. And the same way we’re describing how Facebook is trying to figure out how to monetize more, well that’s exactly what we’re trying to do with this. We’re trying to figure out how do we turn this into a platform rather than just a record store.


Maureen Jann: Isn’t it always a marketing problem? It always kind of rolls back to a marketing problem.


Mike King: Basically. And it’s a lot of fun because it gets me materially involved in that world that I was very involved in before but have been struggling with how to get back to because the opportunity cost of making a song versus getting a new client, it’s pretty ridiculous. So now that it’s to a point where this is something that generates real revenue for me, it’s like, okay well maybe I will make a new record and do things like that.


Maureen Jann: Especially when I think you have kids it becomes challenging to really dig back in to what you loved doing and anytime you can find that opportunity and build it in is gold.


Mike King: Absolutely. My daughter, she’s awesome. But she’s eight months old and she’s crawling, when I say crawling it’s like essentially darting across the room. And it’s a lot to keep up with so it’s very difficult to have the head space to do either of these things, marketing or music. So it’s good that I now have an outlet for it as well.


Maureen Jann: Absolutely. Very cool. Well let’s dig into machine learning because that’s a pretty meaty subject and I’m pretty sure we’re going to fill up the rest of the time talking about that. So we talked a little bit about your interest in cutting edge marketing practices and that’s why we’re here talking about machine learning’s role in marketing. So tell me, this is kind of a challenging one for me to wrap my arms around because we don’t … we’re just small agile business, we don’t spend a lot of time in machine learning so much, so tell me what machine learning can mean to marketers.


Mike King: It really means the opportunity to scale things that otherwise can’t be scaled. There’s a lot of things that we do by hand as marketers, like in the case of digital advertising, you guys do a ton of that. There’s a lot of big management tools that can do that type of stuff, programmatically and such, and it’s really like how can we customize that to what it is we’re trying to do for the client so it’s less humans like clicking buttons and more humans making decisions at the higher level.


And so to that point, we work with a lot of clients to use it for like [inaudible 00:27:40] prediction, for predicative modeling as far as how much they want to bid on things, we have one client who does … they buy leads, so they get pinged that a lead is for sale and then they get all the features of the lead, some indication of that user’s behavior and then they run that through a model to determine how much they should spend on that lead. So they tie it all the way back to how likely is this person based on those features and their behaviors to actually convert for them. So not just like, is this going to be somebody who becomes a lead, is this somebody who’s actually going to buy the product.


And so those are things like, yeah you could do that, have a Riley sitting there doing the work for every person that comes through, but that’s just not realistic. There’s other things he can do, right. So having a mathematical model that’s built into your programming to say yes or no or $50 versus $10 is essentially the wave of the future. It’s the way that we can get the most out of everything we’re doing in marketing.


Maureen Jann: It feels like it’s pretty dependent on walled gardens being opened up in order to do really really well. Like for instance, all that data that Facebook is hoarding right now, it feels like you’re going to need that in order to make those automated decisions really impactful.


Mike King: So not necessarily. You can essentially make these decisions on the outcomes, right? So let’s say, because you’re going to have inputs from your side as an advertiser. You know that you use this creative, you use this ad copy, you targeted these demographics and such. And then you have the outcome, like did this person convert? Did they ultimately buy whatever the KPI is? So you don’t necessarily need to know what happens in the middle. You just need to know what went in and what went out and then you can use the model to make the determination based on the inputs rather than the stuff in the middle that you don’t have access to.


Maureen Jann: Oh. Interesting. So I’m just noodling this trying to figure out, how do you figure out their behavior outside of just that initial interaction that you put out there that you have control over? How do you integrate that with more complex information?


Mike King: So basically what happens is like the vectors that you do have access to are essentially run through a series of mathematical models. And you pick which one actually gives you the most accurate response. So that’s the whole thing because the model will analyze all of the potential actions and say okay, this is the one that aligns with what you expect. So essentially, you’re putting all this stuff into a black box to figure out how it aligns with what you expect. And it’s not really a black box, but for marketers it will be a black box.


Maureen Jann: Feels a little black boxy to me. Definitely.


Tim Mohler: It sounds like a series of multiple regressions to me. Is that kind of how it works?


Mike King: It could be. So depending on which type of model you use, because there are regression models, there’s also other models like support vector machines. They’re all built on different flavors of how you could do it.


Tim Mohler: Cool.


Maureen Jann: I just have a glazed eyes at the moment right now. What? What are you talking about?


Tim Mohler: Any chance to bring out some of that education that I paid for and see it apply to anything in my real life is perfect.


Maureen Jann: More power to you, friend. more power to you. Hire ed degrees. Good for some people. Not for me. not for me.


Mike King: Sure, and my whole thing with exploring this and talking about it is that everything I just described the marketer doesn’t necessarily have to know. They just have to know okay, how can I use these tools so essentially guess and check. So you have your input, you have your outputs and you say okay, let’s try this model. Does this make sense. Is it giving me the accuracy I’m looking for? Okay if it is then let’s use it. And then you can continue to test the same way we do anything else. We don’t necessarily know what’s going to happen when we bid $50 on a click, but we just keep trying until we get what we want and we make those changes.


So I’m not saying that everybody has to become a statistician. I’m not. I know the basics of how this works. And just from my computer science background I understand it, but I don’t have the intimate understanding that other people on my team do, but I can still make strategic decisions around it.


Maureen Jann: That makes sense. Yikes. That’s awesome. This is why we’re talking about this because it’s fascinating, but for me just talking about this brings up all the reasons why I would never do this because it seems really … not I would never, let me rephrase that. This would be challenging for me to get into. So let’s talk about some of those barriers outside of my neuroses are practical uses of applying this.


Mike King: Well I mean, that right there is very much one of the ones. People see it as a very daunting thing, as did I when I first heard of machine learning because there have been programming projects that I could’ve solved better with machine learning, and I’m like, “Ah, it sounds like a lot of work. Like I got to read a lot to figure that out.” Whereas …


Maureen Jann: It’s like you’re in my head. Go on.


Mike King: But the more I’ve dug into it, the more I realize like, oh, people have pretty much done everything for me, I just have to put it together. So the biggest issue is that it’s daunting for a lot of people and then just kind of actioning it across an organization is going to be difficult because there may be people who either know more about it or feel like it should be in their wheelhouse or they should own it, so it’s very difficult for you to be like a like let’s say a marketing manager, saying like, “Oh yeah, we should start doing machine learning,” and then it’s like well you’ve got to get your analytics people involved. You got to get buy in from upstairs. You got to have all these meetings where you’re explaining it over and over and over which makes it difficult to then operationalize.


So I think those two things are the core of why it isn’t happening at a bigger scale, and at the same time we’re getting more and more platforms and tools that abstracted away from the technical ends of it where again you’re just putting in data, you’re labeling data and you’re getting outcomes and your able to use it that way. So I think the more and more tools that come out with this type of methodology behind them, the more it’s going to be adopted.


Maureen Jann: Yeah I think we saw that when Programmatic happened too. It was a little overwhelming and people didn’t understand it and it took a lot of change but now it’s like all the rage and everybody’s talking about it so I can see how that would be a transition that’s worth being patient for as the tools get easier to use.


Mike King: Right. And Programmatic is a great example because there’s a lot of people tying machine learning to Programmatic. So there’s this tool called beeswax where you can essentially have bidding as a service, so you can basically build your own machine learning algorithm for your bid management for Programmatic and then that way it’s not just kind of relying of whatever black box for whatever bid management tool you’re using. So I completely expect this to become more … it’s going to be a bigger deal for people once they figure out how they can harness it without needing to be a data scientist.


Maureen Jann: That’s reassuring. That helps me. I’m patting myself on the shoulder. It’s going to be okay, Maureen. It’s going to be okay. All right, well we talked a little bit about what machine learning means to marketers, but let’s dig into a little more about ways marketers like you see the future of machine learning for marketers. Let’s do some practical applications outside of what we were just talking about.


Mike King: Yeah. I would say some that’s more on the ground level that people can get value out of today is keyword research. So one of the things that’s become a bigger deal is segmenting your keywords in such a way that it allows you to capitalize on the overarching topic rather than just getting on a keyword by keyword basis. That’s difficult inherently because there’s so many different queries that people are typing in on a daily basis. So without going line by line it’s very difficult to do it.


AdWords itself gives you groupings and such based on standard taxonomy, but if that doesn’t align with your website then you’re stuck doing it by hand. So there’s a tool out there called Monkey Learn which allows you to basically set up a model for your keyword research, so then you can put in any new keyword and it will automatically segment it for you.


Maureen Jann: That’s pretty cool.


Mike King: Yeah. And it’s an API as well. It also works with Google Sheets or Google Docs or whatever so you don’t necessarily need to know how to code to do it. You just have to know how to make a spreadsheet which also allows you to label the data. So the labeling part can be time consuming, but there’s tools for like, like you can use Mechanical Turk or something like that and say, “How would you label this keyword,” whatever whatever and then you put it in the Monkey Learn, Monkey Learn learns and then you can just keep adding keywords and then you never really have to think about that process again.


Maureen Jann: Right on. Learn how to make a spreadsheet. I kid. I kid. Awesome. That’s cool.


Mike King: There’s always been a lot of things that we’ve leveraged as marketers that have been built on machine learning. So anything that’s like a recommendation system. For instance Amazon always, like the one you highlight or Netflix …


Maureen Jann: Oh Amazon.


Mike King: And the way they build those is a concept called collaborative filtering which basically says if a person that’s like you liked thing a, b, and c, and it’s likely that they’ll also thing d, se, and f, well you will also like it as well. So it’s the same type of concept that goes into lookalike audiences and things like that. So we’re surrounded by machine learning it’s just how can we then take more control over it and use it to our own benefit.


Maureen Jann: I always boil things down to application and context because those are the two things I think we struggle with with all of this new technology, VR’s the same way, it gets all like, ah VR, I don’t want to go on a rollercoaster while I’m at my desk. That’s not interesting to me, right? It’s all for novelty. But when it gets context and we can apply it in a way that’s meaningful to us as marketers that’s the point where those technologies turn from theoretical to practical and we can really dive in and we can utilize them and help us be better marketers. And that’s the exciting point for me. But I struggle with the theoretical. But it’s cool to hear ways that you’re seeing it used and just actually pointing it out too, because you’re right. Algorithms are everywhere. Machine learning is everywhere. And it’s just a good reminder that’s big brother is already here, it’s just a matter of how we’re using it.


Mike King: Yeah. Absolutely.


Maureen Jann: Cool. Well hey, thank you for being here. We really enjoyed learning about machine learning and I should have thought of this earlier, but I should’ve said are you willing to rap on the show? But I didn’t know you were a rapper and now I’m disappointed. By that’s okay.


Mike King: In me or in you?


Maureen Jann: Me.


Mike King: Okay. Just making sure.


Maureen Jann: Me. Me for not knowing. I Googled you but I didn’t get any rap references and I’m disappointed.


Tim Mohler: Machine learning rap. This is got to be a thing.


Maureen Jann: This is so nerdy, I love it.


Mike King: Well someone did write a machine learning model that can write Kanye West lyrics.


Maureen Jann: This does not surprise me for some reason. Why does this not surprise me. Well fabulous. All right. Well if you get a chance stop by I Pull Rank’s blog, they have some helpful articles that can help you get better results from your campaigns. And thank you guys all for joining us out in podcast land. We’ll be including any links to the show in our show notes, clearly for your reference. If you have any suggestions on topics or guests that you’d like to see us cover, let us know by emailing us at If you like our podcast, we encourage you to rate us on your favorite blog platform. We like ratings. Sadly, it’s time to say goodbye. I’m Maureen Jann, signing off from Fine Point Digital Marketing Updates. Find us on Twitter for our latest content, podcasts, and more. Subscribe to our podcast via your favorite podcast distribution source, including the iTunes store. Looking forward to next week, and for now stay on point.




Additional Resources