Erin Sparks: 00:00:00 Continue our conversation around machine learning and SEO this week, what do we need to know and how do we use that data? With machine learning, I hear the encroachment of Skynet, more and more, don’t you? Well, let’s talk about it today on the EDGE.
The EDGE: 00:00:16 Your weekly digital marketing trends, with industry trends setting guest, you’re listening and watching EDGE OF THE WEB, winners of best podcast from the content marketing institute for 2017. Hear and see more, at EDGEofthewebradio.com. Now alongside Tom Brodbeck. Here’s your host, Erin Sparks.
Erin Sparks: 00:00:39 Well, welcome back to the EDGE again, we’re always broadcasting from EDGE Media Studios in downtown Indianapolis, Indiana. Thanks for joining us on the livestream, and thanks for joining us on iTunes and all the podcast aggregators channels that we have out there. We certainly appreciate your fandom there, give us some feedback on how we’re doing in the show here. We are always talking to the latest influencers and marketing [inaudible 00:01:04] around the planet when it comes down to digital marketing. And we certainly appreciate your contribution as a listener base to be able to let us know who we should be talking to next. So, check out all the recent shows over the EDGEofthewebradio.com, it’s EDGEofthewebradio.com. We’re actually sponsored by Sight Strategics. They’re our title company, as well as a pioneer in the agile marketing space.
Erin Sparks: 00:01:30 So if you’re interested in what we do in that space, go over to sightstrategics.com, check us out. You can learn a little bit more about agile marketing methodology. And if you’d like to have a conversation about how we could possibly do that with your team or with your firm, just give us a Shall be happy to have a free hour consultation and talk about your digital online success. I’m your host, Erin Sparks and the owner of Sight Strategics and founder of EDGE Media Studios. We talk about this regularly, but we always want to ring that bell, why we do what we do? We do this for a couple of reasons. One, most importantly is to be able to educate the digital marketing ecosystem that’s out there to let people know about the latest trends, best techniques in digital marketing when it comes down to social media, as it comes down to search engine optimization, conversion rate optimization, just to name a few concepts.
Erin Sparks: 00:02:22 But on top of it all, we actually use this a bit selfishly because this is our R&D. We’re constantly learning from great talent around the world, of how to do digital marketing the right way. And with that, this is what we’ve been doing for the last eight years. In the studio [inaudible 00:02:40] production, actually today we’ve got Tom Brodbeck, he’s on audio, not video today. Tom, how you doing?
Tom Brodbeck: 00:02:44 Hey, how’s it going?
Erin Sparks: 00:02:45 It’s good. It’s good. We’ve got a camera down so we can’t actually show you, but you know what?
Tom Brodbeck: 00:02:50 It’s alright
Erin Sparks: 00:02:51 You’re over there.
Tom Brodbeck: 00:02:52 Yup, I’m good.
Erin Sparks: 00:02:54 A little bit of technical difficulty, but you know, hey, that’s-
Tom Brodbeck: 00:02:55 What happens.
Erin Sparks: 00:02:55 That’s what happens whenever you’re running a studio.
Tom Brodbeck: 00:02:58 The show must go on.
Erin Sparks: 00:02:59 The show must go on, and we want to go with a leading SEO scientist who’s lying in wait here. We got Britney Muller. How you doing ma’am?
Britney Muller: 00:03:13 Good, how are you guys doing?
Britney Muller: 00:03:23 Perfect.
Erin Sparks: 00:03:23 Excellent. Excellent. Well, we certainly want to unpack our interview with you, but always, each and every show we light up the recent news of the week. So you’re ready to deep dive into some digital marketing news?
Britney Muller: 00:03:37 Let’s do it.
Erin Sparks: 00:03:37 All Right, roll that beautiful bean footage.
The EDGE: 00:03:39 I’m going to do what God put Ron Burgundy on this earth to do. Have salon quality hair and read the news.
The EDGE: 00:03:47 THIS WEEK’S TRENDING TOPICS
Erin Sparks: 00:03:52 No, I don’t think I’ve got salon quality hair, Tom. I mean, do I? Everybody puts products, you just get your hair trimmed down to the nub. But I mean, yeah. All right. I got a little bit of product in there, it’s not shiny at all. Anyway, that’s some great hair over there, Brittany.
Britney Muller: 00:04:11 Thank you so much
Erin Sparks: 00:04:14 Absolutely. You got a good studio over there, is it over at Moz?
Britney Muller: 00:04:17 Yeah.
Erin Sparks: 00:04:18 Fantastic. Over in Seattle. Certainly appreciate you joining us on this show, and we’ve been fans from afar regarding SEO and everything that you’re doing out there and not only at Moz, but also just lighting up education regarding to the more recent changes and machine learning. But SEO is a whole, we need those evangelists out there, so we certainly appreciate you joining us today.
Britney Muller: 00:04:47 Awesome. I appreciate you having me on. This is going to be a blast, I’m so excited.
Erin Sparks: 00:04:47 So let’s talk about our first article from Roger Montii. A search engine journal. Ahrefs Announces Plan for a New Search Engine. Yes, indeed. A brand new search engine. This was some interesting news and a lot of people contributing into what this could mean and a lot of our own guests on the show here over the recent years, so here’s the deal. Ahrefs Announces the Plan for the New Search Engine. Okay. So Dmitry Geras menko, right. Close enough?
Tom Brodbeck: 00:05:21 close enough, Gerasimenko, I think.
Erin Sparks: 00:05:23 Gerasimenko. He announced a plan to create a search engine that supports content creators and protects users privacy. Dmitry laid out his proposal for a more free and open web, one that rewards content creators directly from search revenue with a 90/10 split in favor of the publishers.
Erin Sparks: 00:05:41 So he’s got a couple of goals here. Goal one. He believes Google is hoarding site visitors. Now we have never said that on our show. What’s-
Tom Brodbeck: 00:05:49 Never
Erin Sparks: 00:05:51 You know, he tweeted that Google is increasingly keeping site visitors to itself, resulting in less traffic to content creators. They’re constantly showing scraped content and search results pages more and more, so that you don’t even need to visit a website. And I think we said that once or two times here ourselves. Another point, he seeks to pry the web from privatized access and controls, the gatekeepers to web content such as Google and Facebook, exercise control over what kind of content is allowed to reach people. Along with that, the third point is he believes Google’s model is unfair.Dmitry noted that Google’s business model is unfair to content creators.
Erin Sparks: 00:06:36 By sharing search revenues, sites like Wikipedia wouldn’t have to go begging for money. And the fourth point, he says that the search engine should encourage publishers and innovation. He stated that a search engine’s job is of imposing structure to the chaos of the web should be one that encourages the growth of quality content that like a plant, to support, you know and be able to have a vine that holds up. That’s an interesting analogy. I’m just going to leave it there with that one. All right, so four points of why he wants to build a new search engine. I’m going to tee it up to Brittany cause a lot of your own cohorts at the MOZ side of things chimed in on this real quick. What say you?
Britney Muller: 00:07:19 Yeah, I think you have to give him Kudos to the why, behind building something like this. It would be incredible if something like this could exist. However, I’d like to echo the thoughts in that article on the flip side as far as, has this really taken into account the users? Right, if you’re providing incentives for the publishers, it could get spamy fairly quickly. And what’s the incentive for the users? Right, in my opinion it should come first and foremost.
Erin Sparks: 00:07:50 Yeah, absolutely. John Henshaw jumped in there very quickly and literally was saying very close to what you’re saying is that it’s relevant for the producers of content, but where’s the advocacy for the users in that space, Right?
Britney Muller: 00:08:05 Right.
Erin Sparks: 00:08:06 There’s a number of some top line players in our industry that contributed. Ryan Jones also jumped in there. This sounds like an engine focused on websites, not users. So I can see the argument because we are pretty up in arms. You know, AMP came out, all the different properties that Google has, the local, the map, the Google guaranteed. Obviously ad space, everything pushing down organics, the algorithm changes, shoving a lot of us out of the woods, but at the end of the day, it is their sandbox. Right?
Britney Muller: 00:08:39 Exactly.
Erin Sparks: 00:08:40 Tom, what are your thoughts about a brand new search engine from Ahrefs?
Tom Brodbeck: 00:08:44 I’m right there with you guys. It’ll take a lot to change people’s user habits and I don’t think Ahrefs is coming out with a mobile device that could change a lot of those-
Erin Sparks: 00:08:56 Right.
Tom Brodbeck: 00:08:57 User habits and a Google owns a lot of the market share so, I’ll be mildly less successful than DuckDuckGo. It might be for SEOs only type of search engine, but I don’t know. Yeah.
Erin Sparks: 00:09:13 Are we in the days of, it’s a foregone conclusion that the dominating factor is Google and it’s going to be around forever. I hate to be that prophetic, but, the barrier of entry for a brand new search engine is so high, right? It’s going to be niche. I mean, we’ve been talking about the niche break off into these smaller communities of content and social, what have you, but are we in the days of Google just never going away?
Britney Muller: 00:09:47 I think you have to consider Google has such a strong foothold in the internet of things.
Erin Sparks: 00:09:52 Yup.
Britney Muller: 00:09:53 They’re already integrating themselves into all these different devices, into all these different, basically areas of our life. Like how do you compete with that starting now? I just think that’s going to be tough.
Erin Sparks: 00:10:05 Yeah. I mean, you just can’t catch up.
Britney Muller: 00:10:06 Yeah
Erin Sparks: 00:10:06 We’re going to have, there’s a constant vigil for fairness as Google makes their algorithm changes and they shift content in front of the users. They’re users. Right. It’s going to be more of a diplomatic plan to talk to Google as opposed to try to challenge in the marketplace. That’s such a depressing topic right there. Hey, check it out. Talking about crossover of what Google’s doing, from Search Engine Land from Greg Sterling. We always appreciated the content Google Guarantee local listings are now appearing in Google Home results. Speaking of the devil, Brittany, the Google Guarantee has been associated with the local services ads, which you may or may not know of users or listeners, but it’s a certainly a powerful tool.
Erin Sparks: 00:10:55 But the two programs have been technically distinct. So you’ve Google Home in the Google Assistant environment, right? But you also have this entire local services, where you can get the Google Guarantee badge as truly a distinct from the local services. This is what I want to say. The badge is different than the local services ad environment, right? So you can actually go through a process with Google to get that Google Guarantee badge. And what we’re seeing, what they’re talking about here is that badge and that authentication, that endorsement is making its way into Google Home results. Now, I think they’re picking a choice particular element out of the local services and you know, because local services and ad campaign that there’s going to be some monetization around that, down the road. But this what you’re talking about is that the internet of things, Google Home, Brittany, is in the space and now they’re crossing the streams from a paid environment on desktop into Google Home where there’s only going to be one or two recommended results by Google. Right?
Britney Muller: 00:12:00 Right. I think there’s a lot of confusion about that article though. They went back and revised some of what they had said previously. So to be clear, they’re not providing paid local businesses in voice results, but they are giving that guarantee or Google Guarantee stamp when they send it to your phone, or they email it to you.
Erin Sparks: 00:12:21 Right, and the Google Guarantee is backed up by Pinkerton and that’s the relationship that Google has for doing background checks on local servers providers like [inaudible 00:12:30] for example, and they go through background checks of all the installers that are out in the field, background checks of the owner as well as the history of the organization. And that creates the badge. It’s a whole different pocket of tool sets to be able to advertise that inside of the local services ads. So you earn that by that background check and that’s what Google is actually picking up and putting it into Google Home.
Britney Muller: 00:12:59 That’s interesting. I didn’t realize that was an actual background check.
Erin Sparks: 00:13:03 In fact, I’ve got direct knowlEDGE I’ve just been working on and one of those for our clients and it was an arduous process, especially when ever you’ve got some individuals that don’t want to have their background checked.
Britney Muller: 00:13:14 Right, oh my God!
Erin Sparks: 00:13:15 It holds up everything. So this guarantee factor is a valuable jam out there that is truly improving click through traffic who click through rates, should say. And it’s also right on top of, from a desktop standpoint, right on top of the ad side of things is right on top of the traditional search ads. So there’s some real estate there’s, some value there. And they’re kind of picking a choice piece of that to basically endorse listings on Google Home side of things.
Tom Brodbeck: 00:13:49 Yeah. It’ll be interesting to see how this ties into the Google Home, especially the Google Home Hub,
Erin Sparks: 00:13:52 Right
Tom Brodbeck: 00:13:53 Assuming there’s going to be another version of that Google Home with a screen on it as well. So I mean, the way this read to me is, with your Google Home Hub, you asked for a plumber and you can just click on it right there, straight from the Google Home Hub. It will call [inaudible 00:14:08] freight from the Google home to set up an appointment for you.
Erin Sparks: 00:14:11 And that’s all the local property. That’s the thing, is that Google is getting in the way of conversion, I’m sure they would say they’re helping their customers connect better to the companies and because that lead gen factor that we’ve been looking at on the desktop and mobile site, the click to call and the click to get a lead, that’s where Google is going to put their real estate again to be able to hold on to those, the form leads and inquiries. So it’s good for the consumer. But boy, that’s more and more real estate that’s being taken away from us. I don’t know, there’s pros and cons to everything. I’m just happy I was able to get our final background check for our client because that was a pain in the ass.
Britney Muller: 00:14:50 How long that take?
Erin Sparks: 00:14:52 There were some starts and stops with the previous vendor that was working with them. So it took about when all things considered, it took about two months to actually get through the entire process.
Britney Muller: 00:15:03 Okay.
Erin Sparks: 00:15:04 Which is also something interesting in its own right. It’s not a flick of the switch.
Britney Muller: 00:15:08 Right.
Erin Sparks: 00:15:08 And Pinkerton literally drags their heels, I think deliberately. I don’t know why. Alright. But we’re going to deliberately go to this next article here from Search Engine Land from Barry Schwartz, Google’s de-indexing issue still not fully resolved, but Google is working on it. Tom, what’s the de-indexing issue and what happened?
Tom Brodbeck: 00:15:31 We’re not sure what happened. But essentially what happened is Google accidentally de-index sites from appearing at all in searches. It was just a mistake that happened. They thought it was correct and Saturday, turns out even as of this morning, it’s Tuesday, 3:15 [inaudible 00:15:45] that we’re recording this. It’s still not completely resolved, although some websites are starting to pop back into the listing. So, they’re still working on it. I checked right before we went on air. Still hasn’t been the case that has been solved. But the [inaudible 00:16:01] a random bug you said,
Erin Sparks: 00:16:04 That’s a pretty heavy bug there
Tom Brodbeck: 00:16:06 It’s a heavy bug
Erin Sparks: 00:16:08 It’s a giant bug. And on top of that, seeing that they had a core algorithm change back on the 17th of March –
Erin Sparks: 00:16:15 I’m sure that a number of individuals are freaking out because it could very well be another shoe that fell. But Google says it was an error. What do you know about this de-indexing Brittany?
Britney Muller: 00:16:29 So I actually have some inside info.
Erin Sparks: 00:16:31 Okay. Lay It on us.
Britney Muller: 00:16:33 Doctor P is hooked me up with some data about this whole issue and he had said that there was a drop in stable rankings on Friday, April 7th there is a recovery Saturday and then another drop on Sunday. He says it seems relatively normal now, but the drop was about 4%. It’s noticeable, but not a catastrophe. He has a blog post coming out Thursday and he allowed me to print off some crafts for you guys. I know this is not to point out, but here’s like the really basic like, Can you see this? [crosstalk 00:17:11] He zoomed it in for me so you can see it’s still not totally recovered. This is today.
Erin Sparks: 00:17:21 This is on page one.
Britney Muller: 00:17:22 This is page one.
Tom Brodbeck: 00:17:23 Got It.
Britney Muller: 00:17:23 It’s all right here folks.
Erin Sparks: 00:17:30 So we got page one data. Right. But the de-indexing certainly applies to much more than the page one rankings.
Britney Muller: 00:17:38 That was checking the indexation of those page.
Tom Brodbeck: 00:17:43 I feel like we’ve seen this bug before cause I know Search Engine Land accidentally got de-indexed by Google a couple months ago. So there’s, I don’t know what this bug is.
Erin Sparks: 00:17:51 It seems more like a hammer than a bug.
Tom Brodbeck: 00:17:55 Yeah.
Erin Sparks: 00:17:56 They’re testing a brand new weapon of mass distraction. And [inaudible 00:18:02] letting out of the bag too quickly, I don’t know. Pretty darn scary if you ask me. But I mean, we’ve seen de-indexing happened before, but back in the day it was a DMCA issue or something like that would get you yanked as well as the link spam and everything from Penguin. But my God. Okay, well we’ll just hold on to everything and for everybody who’s listening, check out your own indexes and just watch carefully. I’m sure everything’s going to snap back into the shape. Right?
Erin Sparks: 00:18:34 Alright, so there’s the news if you want to snap back into shape each and every week we sent over to you our EDGE of the Web Newsletter, completely free of charge. We’re talking about everything that we could talk about on the show, who we’re talking to, who we did talk to, the news of the week and much more. If you want to sign up, just go over to the website right there. You can go and put your email in and we’ll use it for nothing but sending over digital nuggets of gold. And if you want to text to the word, to the number, 22828, the word EDGE Talk, you can sign up there right from your smartphone. Just don’t do it while you’re driving, please. There’s the public service announcement right there. So follow all the featured trending topics over EDGEofthewebradio.com. With all that aside, let’s deep dive with this week’s featured guest.
The EDGE: 00:19:20 Now it’s time for it to the web feature interview with Brittany Muller, senior SEO scientist at Moz.
Erin Sparks: 00:19:30 Alright, for our audience who are made up primarily of digital marketers and good, solid core of SEOs, if you don’t know Brittany, then shame on you, because she’s doing a bang up job regularly inside of the SEO circles and has been for a long time. We’ll introduce you and then take it away from there. Brittany is a senior SEO scientist, the senior SEO scientists over at Moz where she is helping in research on SEL concepts, creating educational content, helping to educate people about SEOs, which is constantly a vigil, internally as well as at conferences around the world and communicating SEO trends and desires to the product teams over at MOZ. How long have you been with MOZ Britney?
Britney Muller: 00:20:17 I’ve been with Moz for about two and a half years.
Erin Sparks: 00:20:20 Very cool. And you certainly have seen a number of changes over at MOZ and the culture shifts there. It’s a huge [inaudible 00:20:28] of top level professionals and we’ve been huge fans of MOZ for a long time in our SEO history, our SEO backstories there. But your backstory is really, really important to showcase. So tell us how you got into digital marketing and how you got into SEO?
Britney Muller: 00:20:47 Kind of a weird story. I couldn’t find a job after college and decided to move out to Breckenridge, Colorado to pursue my dream of being a snowboard bum and successfully did that for a season and just got really bored waiting tables and was really active on Twitter. And I connected with the local realtor who saw that I had a journalism degree and wanted me to write local listings. When I started doing that, he introduced me to SEO, Basic HTML and CSS. From there I just went crazy. You know, I thought it was so insane that you could figure out how many searches a month are being done for any given thing. From there I started building up these test sites to compete with major brands. One of my favorite, sort of, claim to fames as I’m learning this stuff is the day of Burton US Open.
Britney Muller: 00:21:50 When they first came to Vail, I ranked number one for Burton US Open above burton.com because I had prepared months and months and months in advance. I had created all this unique content. Had been working on these backlinks. That kind of got my foot in the door in the industry. And then Rich Stats of Secret Stash Media saw what I was doing, and was like, you know, you can do this stuff and get paid for it. Right? Like, why don’t you come work with me over in Vail and what we can do this as a career. And so, he taught me so much and then after a year I branched off and started Pride Marketing where we did strategic medical marketing for private practices in Denver.
Erin Sparks: 00:22:34 Very cool. So you went vertical there for a bit.
Britney Muller: 00:22:37 Yes. Which was fun. And then after about five years of doing that, was pretty burnt out. Had tried a couple of things to scale and those didn’t work and I made sure that all my employees had employment elsewhere and then was approached by MOZ. So was excited to apply and go through that whole process. And here I am.
Erin Sparks: 00:23:01 Here you are.
Britney Muller: 00:23:01 I know.
Erin Sparks: 00:23:03 You’re one of the premier spokespeople in the Moz community and you’re certainly lighting up the conferences. So where do you speak right now or have you spoken to in the last year?
Britney Muller: 00:23:14 I have been so honored to speak on stages like Search Love, MozCon, Learn Inbound, different international ones, even like Retail Global in Australia. Some local events, kind of covering all sorts of stuff there.
Erin Sparks: 00:23:34 That’s awesome.
Britney Muller: 00:23:35 Yeah, it’s actually crazy.
Erin Sparks: 00:23:39 So, who are you going, where are you going to be speaking here in the future in the next couple of months?
Britney Muller: 00:23:44 In the next couple months I will be at search leads in June and then MozCon and CTAConf by Unbounce.
Erin Sparks: 00:23:54 Very cool.
Britney Muller: 00:23:55 Really looking forward to those.
Erin Sparks: 00:23:55 Always wanting to go up to CTAConf.
Britney Muller: 00:23:55 You should,
Erin Sparks: 00:24:00 Got to talk to Yuri again and see if we can give [inaudible 00:24:02] do a little bit of broadcasting. Hey, we got you on board because, well, first and foremost, love talking to evangelists such as yourself. But there’s a particular topic that you’re pursuing right now. That’s machine learning.
Britney Muller: 00:24:17 Yeah.
Erin Sparks: 00:24:19 We’ve had another string of shows here recently regarding machine learning. And I want to give our audience a little bit more depth of knowlEDGE about what’s happening here and what it means to our marketing. So can you tell us a little bit about machine learning, what it means and why it’s so important to SEOs to learn?
Britney Muller: 00:24:38 Really, really good questions. Essentially what machine learning does is it takes a huge amount of input training data, and can output a model that automates things for you, so you can get it to write meta descriptions for your pages. You can get it to help optimize page titles. You can get it to identify major SEO issues or to use different models of groupings, like what JR Oakes talked about on your recent podcast with him, where he’s basically grouping key word topics. So you’re able to distill a huge amount of data and to very easy to consume insights. And it’s a beautiful thing when you think about it being applied to the world of SEO because for as long as I’ve been involved in SEO, which has been around 10 years now, it’s always been around these the same mundane, low level, title tags, meta descriptions, this things. Just imagine what we could do as an industry if we could evolve to focus on higher level strategy and allow basic models, machine learning models to take over some of that foundational stuff.
Erin Sparks: 00:26:01 Absolutely. I think that’s been the bane of every SEO existence, is having to rewrite the title tags or having to rewrite the meta content. I mean, we talk about putting chess pieces on the board regularly or this is the strategy, but when it gets down to it, there’s a lot of, it’s not difficult work. It’s just time consuming work to disco through each and every one of those pages and let alone have the hypothesis of, okay, if we’re tracking this, we’re going to change up, kind of the schema of how we’re going to, not schema in that way, but in actual organization of your titles and then you’re gonna change collectively the organization tiles, we’ll be able to improve CTR, all the while you have to measure everything and you have to have that data feeding back into the system. Right?
Britney Muller: 00:26:51 Exactly.
Erin Sparks: 00:26:52 So machine learning, we have a new horizon here where we can plug our own sites into machine learning algorithms to be able to set the test and be able to see what’s happening in glean a heck of a lot more insight then, we as marketers would ever be able to do on our own. Right?
Britney Muller: 00:27:11 Absolutely. And you have to think about it like there’s a lot of things machine learning won’t be so great for. And the way to identify those opportunities is to consider how much training data you have to feed it. So if you don’t have much input or much data to do a thing that you want to do, you’re going to have to find resources elsewhere or maybe pivot. But that’s where most of data cleaning, data munging comes into play as that’s 99% of this data science stuff and machine learning is having clean, high quality training data. Without that, you’re not going to get a good output. You’re not going to get a good result.
Erin Sparks: 00:27:54 Because garbage in garbage out basically.
Britney Muller: 00:27:56 Exactly. And that’s why the whole ethics thing, Google is kind of spinning out of control now too because if you have bias data or slanted data that you’re putting into a model, it’s obviously going to come out that way. So we as this new space, you have to be so careful about what you’re building and how you’re building it.
Erin Sparks: 00:28:16 That topic right there, let’s unpack that a little bit more for our listeners who are not aware of that is that Google has been on a path for normalizing data or as creating neutral, unbiased type of data and be able to learn from that. But by the very definition of removing bias from content, you’re actually removing that content from the users that actually consume that content as well. Based on regional bias. Is there a risk of normalizing it so much and homogenizing content that you’re losing connection to regional patterns of search?
Britney Muller: 00:29:04 Not in the ethical ways that,
Erin Sparks: 00:29:08 Sure, sure, sure.
Britney Muller: 00:29:09 Right. But as far as, yeah,
Britney Muller: 00:29:16 Words are hard.
Erin Sparks: 00:29:18 Words, people, words. But the effort of actually de-regionalizing content could actually separate you from your users, so to speak.
Britney Muller: 00:29:31 I don’t think you would want to de-regionalize content by any means or search data by any means. But if you’re, like for example, TSA has a Kaggle competition right now. Kaggle is the number one largest data science competition website and there’s a competition that they have on there for one and a half million dollars to build a model that identifies threats. And you can imagine that this training data isn’t evenly spread across demographics. Right,
Erin Sparks: 00:30:09 Right.
Erin Sparks: 00:30:14 No, I get it completely. So you’ve got to be able to have clean lines to be able to measure, what are some of the other problems with machine learning?
Britney Muller: 00:30:26 Oh boy. Computational power has gotten better. It was interesting. I was just at an AI conference here in Seattle a couple of months ago and they were saying how, there’s these three primary sectors of machine learning that makes it possible in the biggest bottle neck currently is people. There’s not enough people familiar and well versed in machine learning to execute on building effective models. And you also have to consider, you also need individuals with domain knowlEDGE. That’s where JR is talking about the medical implications have been incredible. But those have also happened with domain knowlEDGE experts in the field of image recognition for let’s say cancer detection.
Erin Sparks: 00:31:16 Right.
Britney Muller: 00:31:17 But what’s most exciting about this part of the bottleneck is that, Google is making it so much easier for the average person to do their stuff. Like I am not a formally trained developer. I never thought in my wildest dreams I would be creating models to output my tweets from here to there, and you wouldn’t believe what you can do if you know where to look. And so one of my favorite resources is Google Codelabs, and you filter it by TensorFlow and you can walk through step by step a machine learning process. And I don’t know about you or your listeners, but I’m sort of a monkey learner. Like I need to see-
Erin Sparks: 00:31:56 You got to see it.
Britney Muller: 00:31:58 Yeah. If I can see someone do it, I get really jazzed up and excited and I can do that too. Right? So, I learnt a lot from like YouTube videos, surprisingly. There’s tons of great resources on YouTube. But same with these different Google tools that we can link to maybe.
Erin Sparks: 00:32:16 There’s a whole another, There’s a new, I don’t wanna say a couple of industry because it kind of puts a diminishing factor on it, but there is a new level of language that we’re starting to find ourselves in. And it’s that modeling language that building of… It’s the Raspberry Pi of things is that you can literally start building this in the cloud space because now we have CPU, right? And you can start modeling this and no electrons get harmed in the effort, but it’s no longer a barrier of entry into understanding these type of data components where it used to be, if you.. You’d only get to play with it if you’re a developer, right? Because we all liked our SQL join statements right back in the day. And that was always fun to be able to see, oh wow, this is amazing what you can do. But now we’re in a space where this is all kind of bridging that gap and translating it for us.
Britney Muller: 00:33:14 Exactly. And to play off of what JR had said about, he mentioned Jupyter Notebooks, Google came out with their own version and it’s called Colab Notebooks. It does the exact same thing and you get free GPU. So, computationally it can go way faster and you can share it more accessibly with teams and with other people, which is fascinating. So it’s been really interesting and wild to see Google follow along the footsteps in providing this stuff, you know, open source, but yeah, it’s… Back when I got into the TensorFlow stuff, for example, a basic model was hundreds of lines of code and now it’s just a handful. It can be under 10. So from that perspective, they are working really hard to make it more accessible. And I just hope that we can sort of plant the seed of curiosity and you know, you and like your listeners and you wouldn’t believe what cool things you can do with it. It’s just really fun.
Erin Sparks: 00:34:18 You start looking at the Bitcoin evolution of how the community’s took over and started to develop this non nationalized currency, right? And being able to have open source at the backbone of this, you’re really looking at, I hate to get weirdly philosophical, but you’re getting into a space where there is new conversations happening that are non aboriginal know by us, they’re in this new machine realm and there’s now easily translatable objects to be able to work with and you’re going to have a community that’s going to pour into that. So you’re going to have different types of templates, different type of functional macros for lack of a better description, where even those are now going to be building blocks of how to really train machine learning. Right?
Britney Muller: 00:35:10 Exactly. Yeah, exactly.
Erin Sparks: 00:35:12 It’s exciting. I say it’s a brand new horizon for… Again, I think it completely provide a care for us as a human society to be able to start training the tools better and actually not have to do the grunt work that I actually get into much more cerebral mechanic side of things.
Britney Muller: 00:35:34 Exactly.
Erin Sparks: 00:35:35 Am getting excited. All right.
Britney Muller: 00:35:36 I know, it’s so fun. It’s funny, like this morning, thinking about this podcast, I was thinking about all that stuff and how exciting it is and how there’s already incredible people in our field doing the most mind blowing things. Like I would like to make a part time job of just following Paul Shapiro and JR Oakes around, and like translating some of the things they say, because I feel like I need to scream it to people for them to really get it or give them a little shake. Paul was recently at a conference Traffic Think Tank, and he was talking about how a couple of years ago he created a system to create automatic 301 redirects to relevant pages. And it just like, like think about that. That is insane. He created a crawler to go back to archive.org find all of your old URLs that currently 404, put it into this model that detects a relevant current 200 page that’s alive,
Erin Sparks: 00:36:41 Alright.
Britney Muller: 00:36:41 and talks directly to your htaccess file to do this automatically.
Erin Sparks: 00:36:46 Wow
Britney Muller: 00:36:46 This stuff is crazy, and I just… Yeah, I feel like we have to tune into them a little bit better because, they’ve been doing some of this stuff for years, and SEOs are just finding out about it now. So yeah, it’s so exciting and really fun to think about.
Erin Sparks: 00:37:02 Yeah, we had Paul on the show a couple of years back. Four or five years back. And it was amazing. We were in awe. I mean, he just sort of laid all. It was literally… We were aghast and just listening to what he was saying there. I got to go back to that show. I remember that conversation. We were all just kind of looking at each other like, oh my God, it’s like a fount of geekdom right here. No offense whatsoever. Paul. It’s amazing.
Britney Muller: 00:37:30 It’s incredible.
Erin Sparks: 00:37:32 To that point, there is going to be this new frontier of translation, of communication and sharing of methods that once the realm of stabilizing how you can use it, it gets embrace because you can hypothesize all the way by they’re feed. Don’t have the technology plugged into your system, and the training tools that are there, but also be able to receive that data, right.
Britney Muller: 00:38:00 Yes
Erin Sparks: 00:38:00 And be able to use that data. And once that’s in place, then Katy bar the door, we’re going to go hog wild boy. The euphemisms that I just used, my Texas roots are showing, I mean, it’s bad. Alright. There are two types of machine learning that you’ve discussed here recently. One is supervised, the other unsupervised. What are those and how do they differ in results?
Britney Muller: 00:38:25 Yeah, great question. Those are the two primary results. There’s also things like semi supervise, but for the most part it’s divided into those two categories. Supervised is labeled training input. So that is for example, let’s say you have, I don’t know, SERP results and the URLs for this one keyword that you want to optimize the title tag for, that maybe it could evaluate, that’s not a great example but, you could basically put in labeled training data like for example, my tweets that I did recently, I throw in a bunch of tweets and said, this is a tweet, this is a tweet, this is a tweet. I want you to use this model to produce something new, that’s known as a label or a basically supervised model. Unsupervised is let’s say, you have a geographical data and you want to lump things together.
Britney Muller: 00:39:22 So with unsupervised, it’s really exciting because you can just throw tons and tons of data at a model that’s unlabeled and kind of chaotic and, it allow different grouping models to take over and cluster things for you. So that’s where things like [inaudible 00:39:39] come in. All sorts of different clustering.
Erin Sparks: 00:39:42 That’s where the old mantra of the brain trust of getting a bunch of people in the room and all of a sudden something else that was not intended starts to come out of that. You have the ability to start reading these organic clusters of relationships as opposed to having more fenced in for a specific goal. Right?
Britney Muller: 00:40:04 Exactly. And the whole concept here is to allow the models to take over and to identify patterns that we as humans could never recognize. Right?
Erin Sparks: 00:40:15 Yeah
Britney Muller: 00:40:16 It’s incredible what things these models will come up with and output. It’s hysterical. But yeah, if you give it enough quality data, it should be able to identify different patterns depending on what you’re going to do. And then from there you leverage it to do whatever it is that you want to do.
Erin Sparks: 00:40:36 I dig it. So getting back to some practicality for the SEOs in the audience, right? One of the… we know the elements that help in enhance click through traffic, click through rates. I should say, and we know in the SERPs we’ve got the titles, we’ve got the meta content. How can we use machine learning to help us rewrite our meta content to improve a click through rate? Now, let’s get some practical science here.
Britney Muller: 00:41:06 Yeah, that’s a really good question. So there’s a number of ways you could do this. One of which is allowing a model to basically [inaudible 00:41:16] the current page one Meta descriptions and come up with a competing meta description. I haven’t seen that to do so well as opposed to just taking your onpage text, running it through a summarizer model and limit lean it to the meta description link has proved to be incredibly useful for a couple of hundred mass URLs that I’ve been playing around with it on. But there’s different summarizer models and there’s tons of different ways you can do this. And to your point, the toughest part is connecting this into a workable solution, right? And that’s something that I talk with JR all the time about because he has that tactical mindset of, Oh, here’s what it takes to plug this into a Google sheet and allow you just to put in the URL and on the right column you could see the meta description. Or someone might come up with a wordpress plugin to connect these different fields.
Erin Sparks: 00:42:23 Right?
Britney Muller: 00:42:24 Things like that are moving forward. They’re not moving forward as fast as I wished that they were because it is just tough right now. But I think if you are really curious and driven to find these solutions, we have already been putting things out there for you to run with and to try these things with. JR is super approachable, so as Hamlet Battista and we’re all more than welcome to help people try to make these things happen on their live site.
Erin Sparks: 00:42:56 No, there’s an active community of development that want to, they actually want to test everything out. If you want to step up and be a guinea pig, I mean there’s ample opportunity out there. There’s these training elements that still have to exist when you’re writing a meta description, right?
Britney Muller: 00:43:19 Yeah
Erin Sparks: 00:43:20 You’ve got to fence it in.You have to have certain objectives that you put in place for any piece of content, any area in which you’re trying to effect a change, you still have to fence it in with grammar, for example, right? With different baseline structure of what the game rules are within that particular element, right?
Britney Muller: 00:43:45 Right. And I don’t think we’re anywhere near the level of trusting these models to go crazy on your primary pages. I think you still own that, right? And you protect those and you do that by hand. But when you’re looking at something like Moz’s Q$A forum that has tens of thousands of pages for these very unique random long tail questions. This works really great. And it’s also, you have to consider it’s better than nothing, right? Because a lot of people will just leave them blank when you have a huge website.
Erin Sparks: 00:44:18 Sure.
Britney Muller: 00:44:19 You can’t feasibly, write Meta descriptions for every single one. So if these can get you part of the way there, all the more power to you.
Erin Sparks: 00:44:26 Absolutely. And then that’s the thing is that, don’t wait until things are structured in such a finessed way. The summarizer models like you’re talking about, they’re built to logically pair content together. You’re not just going to have a random shuffling of terms based on a preset of… You don’t want to go back into the dark days of keyword density. Right. The goal here is to make a logical summary of content and these tools are already there for us.
Britney Muller: 00:44:59 Exactly. Exactly. Yeah.
Erin Sparks: 00:45:01 So, there is a cost to play in this space of machine learning, right? It’s not free. We can’t just pluck it off the tree. You’ve got GPU, you’ve got CPU, and you’ve got TPU right?
Britney Muller: 00:45:13 Yeah
Erin Sparks: 00:45:14 Now. Lay It on us. What are these differences and well how does this play into machine learning?
Britney Muller: 00:45:21 Got It. So I’m not super well versed in the technical aspects of it,
Erin Sparks: 00:45:24 That’s all right. It’s all right.
Britney Muller: 00:45:25 But I know that GPU is significantly more powerful than CPU and TPU is significantly more powerful than GPU. And TPU is basically created by Google to power up TensorFlow models.
Erin Sparks: 00:45:41 Right.
Britney Muller: 00:45:41 So it’s a specific hardware that allows you to computationally run these a lot faster, which is super exciting. But yeah, I mean you have to pay to play in some aspects in other aspects if you’re willing to sit on a GPU, like for a lot of my work I just do on my local computer, and it just takes hours and hours and hours. And in that time, I make sure that I’m not dropping out of the runtime or impeding on it in any way. Cause that’s a super pain if something interrupts your training process.
Erin Sparks: 00:46:16 Got It.
Britney Muller: 00:46:17 But yeah, there’s all sorts of ways to do it. But I do think it’s more approachable than people think. Like, we’re now talking about disposable AI. There are these TPUs that are coming out that are literally going to be a couple dollars that you can put a machine learning model on. And it was demoed recently where it was in the backseat of like a lift or an Uber and it could detect if an object was left behind when a passenger left. And the beautiful thing about things like disposable AI or AI that just lives on a device and it’s not communicated back to the server are logged on a server, is that there’s no privacy issue there, right?
Erin Sparks: 00:46:57 Right
Britney Muller: 00:46:57 It’s put in and put out, like it’s not being logged or transferred. It literally lives on a little device. And I think we’re gonna see more and more trends headed that direction or what they call the EDGE. So yeah, models on the EDGE.
Erin Sparks: 00:47:14 Exactly. And we’re talking about it on The EDGE. How about that?
Britney Muller: 00:47:14 Yeah,
Erin Sparks: 00:47:20 Gratuitous plug. So TPU, Google manufactured this chip and is their technology. I’m out of my league on the vernacular here, but they also have the cloud based TPU as well that you can also lease and use from Google. And so the difference there from the localized standpoint, right, you’ve got dedicated machine, dedicated logic and dedicated resources. You can also tap into a much larger power but so well I’m like referencing Google to God, that’s not the case. No, but it’s a larger power of processing and TPU itself. I mean at that point in time you are going to be relinquishing a little bit of that privacy. Right? For the trade out of speed and logic models.
Britney Muller: 00:48:09 Absolutely.
Erin Sparks: 00:48:11 It’s the coin of the realm and you better believe that Google is going to learn from all of these.
Britney Muller: 00:48:17 Exactly.
Erin Sparks: 00:48:18 Yup.
Britney Muller: 00:48:18 They do. Anyways. Like in our captures for the, you know, click the signs in this picture, 50% of those photos, they know where the signs are and 50% they don’t and they’re asking us to train it. So we’re literally training models every day and we have no idea. And I do often think about that when I see all these new products being put out, I get super excited about like Colab, right? And use different data sets that they’ll make public. It’s like, well crap. You know, they’re watching and learning from all of us as well.
Erin Sparks: 00:48:50 And we wondered what those are all about. Absolutely. And we’re allowing it… We’re training it because we want it to make our lives more convenient. So there’s an entire industry to tee it up. So, if you think about it, back in the day when Google was first opening up the Bayesean logic, right? Being able to do predictive analysis on what you’re interactive with to be able to give you your next set of SERPs that are going to be more relatable to your decision making. This is where we are. It’s learning from each and every private citizen. This on Google and they have these larger clouds of constructs, these larger concepts that are constantly being trained by the performance of behavior that we have on Google. Right?
Britney Muller: 00:49:36 Right. Yeah.
Erin Sparks: 00:49:37 So income Skynet, right?
Britney Muller: 00:49:40 I don’t think so, and that’s where like I think it’s so funny. Because, this is by no means going to take over or do anything wild. Like it is literally programming and statistics, living on servers. These things aren’t making their own decisions or turning things on. It’s not like a crazy thing at this current level. I don’t know.
Erin Sparks: 00:50:05 She qualified that.
Britney Muller: 00:50:07 I will qualify it, but I think you look at what Elon Musk was doing, OpenAI and he’s [inaudible 00:50:14] left, but OpenAI basically was able to produce, they said such a well trained natural language generation model that they wouldn’t release it publicly. But I looked at some of the closer information on it and that data set was still talking about underwater fires, and things that logically don’t make sense. And I think you have to think about that. We don’t… These computers that are trying to learn like humans learn, are going to have an impossible time on certain things because we’re not writing or speaking about how I am smaller than my house.
Erin Sparks: 00:50:57 Right.
Britney Muller: 00:50:57 The computer has no idea about these, like very basic logical things. That’s why the Allen Institute is so interesting because, they’ve been working years and years and years to try to feed that sort of information into this sort of be all end all of general knowlEDGE computing. Yeah. It’s so interesting.
Erin Sparks: 00:51:16 Trying to teach conceptual observation to a computer, You know, eventually it’s going to happen Britney, right? I mean, there’s going to be-
Erin Sparks: 00:51:27 Right, right. But what you’re saying is we shouldn’t be scared in the immediate future.
Britney Muller: 00:51:32 Yeah.
Erin Sparks: 00:51:33 So, though the Google assistants is not going to take over and kill my apple device just because. So, where do we go from here? How do we contribute as developers, as marketers? Because marketers really do have a unique play in the training of machine learning because marketers are by their very nature, trained to observe intent.
Britney Muller: 00:51:57 Yeah.
Erin Sparks: 00:51:57 And are continuing looking at ways to be able to produce value to that consumer as quickly as possible. So the nature of this type of training model that you’re talking about, right? That the computers need so badly, we’re marketers are primed to give that type of data back because of the goal oriented perspective that they’re in. Right?
Britney Muller: 00:52:24 Right, right. Absolutely. Yeah. That’s a good point. And I think marketers in general need to just stay really curious about this stuff and ask stupid questions. Like that is what has connected me with some of these other individuals is, I might come across one model but have no idea how to connect it to something in Google Sheets or my own website. And just asking, is this Google search console data available for me to do this? And then this way, you know, there’s so many applications we haven’t even thought about and I think you and your listeners are likely the ones that have the next great idea and you don’t even realize it. So we need that domain expertise. We need that conversation to be happening and to sort of fuel some of these advancements.
Erin Sparks: 00:53:18 Absolutely. It’s still very abstract. I’m sure it’s abstract to our listeners. Where are some places our listeners can go and find out more, kind of compartmentalize it, based on their own skill sets. You certainly talked about CodeLabs. Are there any other resources that one, we can learn about our role in training AI and machine learning, but also get our hands messy and start mucking about with stuff?
Britney Muller: 00:53:47 Yeah, that’s tricky because there’s so many online courses now that are selling machine learning courses, data science. Andrew Ning has a really great course. Google has their own data science course. Amazon has made their in house training open source. One of my favorites is from Harvard that you can find on GitHub and you just do a search for CS109 Harvard. It’s their data science one on one course, but that has incredible videos, insights that really help you wrap your head around this stuff.
Britney Muller: 00:54:20 But truly, you know, you don’t get a good grasp of it until you start playing around with it. It’s a lot like programming, right? You can’t read a book on programming. You have to break things and try to build something and, you just have to give it a go and be willing to fail, you know. It’s like, just try it. And I think that’s… I was listening to some podcasts about a data science teacher, and that was the one thing he said he wished his students would be less afraid of is like, if you haven’t tried that command in your terminal or whatnot, like just give it a try. Like, just go ahead and try to do this or try to break this. And-
Erin Sparks: 00:55:01 That’s when the light bulbs go off.
Erin Sparks: 00:55:11 Yeah, you don’t want to do that. But I liking it back to I’ve got a 11 year old boy who’s in sixth grade and he’s being trained with the robot Leech teaching, I ease, he’s got some competitive robot, a grouping. I mean it’s actually a national wide-
Britney Muller: 00:55:32 That is so cool. Wait, is that Amazon’s car?
Erin Sparks: 00:55:34 No, actually, this is an entire nationwide, worldwide actual competition of each year you are assembling a robot with teams and it’s high end competition was like six or seven teams per school and they are putting together their own assembly of a robot. Everybody across the nation has the exact same parts, and you actually have to take it on a course where you… This year it has certain grapples and you are actually remote controlling this thing with a pre programmed flight plan. So they have to learn their quadrants. They have to learn everything, but they also have to be competitive and pick up certain barrels. They have some hooks that they can hang from different bars and what have you. And it’s all scored on a judge’s point scale. So they’re all collectively across the world competing and learning these key principles. And I’ll tell you what, they’re learning it with visual clues, visual software representations. And you got that plus the Raspberry Pi Realm where you can program all this stuff. Now with this new level of visual management.
Britney Muller: 00:56:43 Yeah.
Erin Sparks: 00:56:43 I mean we’re getting a hold of all the stuff out of the way, and getting them into the pure mode of discovery, you know?
Britney Muller: 00:56:50 Oh my gosh. I think what’s so beautiful about that too is the fact that, they’re making it so fun and competitive. That is what is going to keep people sticky, right? To this idea of learning and continuing their education in the space. Like you have to enjoy it. You have to do it with things that make you laugh or that are fun. You know?
Erin Sparks: 00:57:09 You better believe, I mean, there was some kids walking around in caps because they were that bad ass. I kid you not. I mean, they were hitting like 26, 28 points can consistently out of a potential, you know, the average hitting around 12. And they were just, their team was Apex. Literally named Apex and they were just killing it. And I mean, it was just amazing seeing the pride and the competitiveness in such a specialized scientific concept. It fills you with hope that these are the next programmers and they’re, I mean, they’re going to talk to us about command line entry. Come on, come on. We got all these tools.
Britney Muller: 00:57:47 Yeah. Right. Oh, that’s incredible. I love hearing that.
Erin Sparks: 00:57:51 I appreciate it. Yeah. We certainly enjoy the time today talking with you about machine learning. Something bugs you about machine learning in your industry. Are they vernacular machine learning. What is it?
Britney Muller: 00:58:04 Everyone calls it AI. Everyone. And it’s not
Erin Sparks: 00:58:08 Alright.
Britney Muller: 00:58:08 Its a set of AI. And, you could argue the fact that we’ve not reached AI yet due to the touring test, right?
Erin Sparks: 00:58:17 Yep
Britney Muller: 00:58:17 So, yeah, just the fact that everyone calls it AI or will use it with the wrong vernacular.
Erin Sparks: 00:58:26 Yeah.
Britney Muller: 00:58:26 Stuff like that.
Erin Sparks: 00:58:26 I think I actually did that during the show today.
Britney Muller: 00:58:28 That’s okay. It’s literally just become a norm and I feel like that, get off my lawn. Like I’m just the cranky, she learning old lady or something.
Erin Sparks: 00:58:40 That’s funny. Well, conversely, Britney, what excites you about your industry right now?
Britney Muller: 00:58:45 Oh my gosh. What really excites me, pertaining to machine learning is this idea that we are in a space where we can really improve the lives of others. So there’s no reason why you couldn’t create new medical models are models for detecting clean water or all sorts of things that could be done on a really basic level that could save people’s lives, right? Like they’re already doing that. Machine learning is already saving lives and the implications to do that are tenfold in the future. In addition to just automating the everyday task that maybe we don’t like to do.
Erin Sparks: 00:59:24 Absolutely. This is fantastic sediment and, we certainly want you to evangelize and be a sure pro for us to climb that machine learning mountain because there’s going to be a bunch of us, that are going to be scared of breaking a few things and you know, hopefully there’s a lot of sandboxes that are going to be created so we can play around with the stuff and really learn how we can use it for our own devices. Right? And not just the conceptual nature but making it a real tactic in our bandolier of digital marketing techniques.
Britney Muller: 00:59:56 Yeah, and that’s something that we’ve been talking about here at Moz is, you know, we have a data science club team and we have a data science course for anyone internally that wants to take it by our data scientist Neil. But we would like to provide this stuff to the General SEO digital marketer public, and see how it goes and hopefully try to distill it even further to really manageable education content.
Erin Sparks: 01:00:28 Nuggets of machine learning that we can use for practical use.
Britney Muller: 01:00:32 Yes.
Erin Sparks: 01:00:33 I dig it.Sign me up. All right. Hey, we want to know that fun fact that you gave us on our pre show, let our listeners know about that.
Britney Muller: 01:00:43 That I live on a boat.
Erin Sparks: 01:00:44 Oh No, no, no, no, no. Well, yeah, you do live on a boat.
Britney Muller: 01:00:47 Oh wait, what’s the-
Erin Sparks: 01:00:48 And on top of that, you’re living on a boat and you’re telling me not to be worried about AI, right? Or SkyNet, but you have an escape route right into the sea. Now, I’m getting mixed messages here.
Britney Muller: 01:01:01 That’s my backup.
Erin Sparks: 01:01:04 But you hacked your way into Harvard’s CS 109 Data Science class. Tell us about that real quick.
Britney Muller: 01:01:10 Yeah, so that was about five years ago and at the time I’m fairly certain they had no idea that there is a public access point to view these videos. And they would… I would be able to watch the videos as they were streaming it or an hour right after I forget. But I followed the entire semester’s course and started, emailing TAS because I had questions about these homework assignments that I couldn’t even turn in, but they would email me back and it was really lovely and I don’t think that any idea that I didn’t belong in there, but
Erin Sparks: 01:01:43 Oh my gosh,
Britney Muller: 01:01:44 Yes. So funny. So funny.
Erin Sparks: 01:01:47 Who the hell am I speaking?
Tom Brodbeck: 01:01:49 You didn’t have a harvard.edu email address? What are they emailing?
Britney Muller: 01:01:53 I know it was from a Gmail account and they were so nice.
Erin Sparks: 01:01:56 Wow. finally, Britney, is there anything that we can promote for you on the show today?
Britney Muller: 01:02:02 Maybe the new beginner’s guide to SEO that we published here on Moz. I worked really, really hard last couple of years making sure that we had a really high quality replacement and update for everyone. The old one had gotten over 10 million views, a unique [inaudible 01:02:19] which is wild. So I’m hoping that this can do some of that justice and be a little bit more updated for everyone.
Erin Sparks: 01:02:27 Yeah, you got a crush that. That’s going to be awesome. So we’re certainly going to add our few clicks over to the beginners out. There is the beginner’s guide to SEO. You certainly want to jump in there for newbies into SEO and yeah, you know, you’ve come across this podcast and we’re predominantly an SEO podcast from a focal point, but, this is great information and Moz is always a huge resource but I think I’ve gone back to this thing numerous times to seeing how deep it is. In the different angles and it’s great to be able to educate your clients as well, by the way.
Britney Muller: 01:03:00 Yeah, that’s a great point. That’s a really good point.
Erin Sparks: 01:03:02 Well, Britney it has been a pleasure speaking to you today and certainly appreciate your passion inside of machine learning. We want to recommend to our listeners, go track her down at Britney. That’s one. T. N. E.Y Muller, M. U. L. L. E. R. Also can track her down on Instagram as well. We certainly appreciate your time. Any final words for the futurist of digital marketing machine learning technician. How about that?
Britney Muller: 01:03:31 Just stay curious, you know, and play around and break. Don’t be afraid to break stuff and feel free to reach out to myself or any of those that are dabbling in this space on Twitter. We’re happy to connect and we’d love to have more of you in this little niche community we have.
Erin Sparks: 01:03:49 Very cool. Very cool. Well, thank you again for your time today, Britney, and we’ll certainly be lifting you up from afar.
Britney Muller: 01:03:55 Awesome. Thank you so much.
Erin Sparks: 01:03:56 You’re more than welcome. More than welcome. Thanks for listening to EDGE of the web radio. Special thank you to our colleagues here at site strategics, for putting on a great show again and again especially our guests, Britney Muller. Then makes sure to check out all the must see videos over www.EDGEofthewebradio.com. Tom, who are we talking to on Thursday?
Tom Brodbeck: 01:04:16 We’ve got Derrick Macoute, I learned how to say his name. Macoute.
Erin Sparks: 01:04:20 Macoute, Excellent. Yeah, he’s from director of consulting. That’s right. So we’re going to have our Thursday open as well with him and we certainly appreciate anybody jumping into the live stream and asking questions. You didn’t do it this time. Come on guys, jump in. And if you’re listening to our podcast, we certainly would appreciate our listeners to contribute into our conversation as well. So, always check us out Thursday. Usually Thursday at 3:00 PM eastern and you can then chime in right there. Make sure you check out all of our stuff over EDGE Of The Web. Listen to us on the podcast aggregators, iTunes, Google play, Stitcher, IHeartMedia, Spotify, all the places. Give some ratings and reviews because we’d really appreciate that as well. That’s how we are actually optimize in our space on the podcast platforms. So thanks for listening and we’ll talk to you Thursday. Do not be a piece of cyber driftwood. Bye. Bye.