searchmetrics email facebook github gplus instagram linkedin phone rss twitter whatsapp youtube arrow-right chevron-up chevron-down chevron-left chevron-right clock close menu search

TrendSpotting Episode 7: Scott McLin

Episode Overview

Join us for Modern Marketing Month as we discuss the opportunities presented by performance marketing with Scott McLin of Sojern. Scott McLin is a product leader well known for his ability to build cross-functional, scalable, machine learning-driven solutions for search performance marketing in the travel industry.

iTunes       Overcast      Spotify      Stitcher

Episode Transcript

Ben:                 Welcome to Modern Marketing Month on the TrendSpotting Podcast by Searchmetrics. In this podcast, we dive deep into the ways innovative marketers use experience and data to identify the macro trends that influence where you should be investing your marketing budget.

Ben:                 This podcast is brought to you by Searchmetrics. At our core, the Searchmetrics team is a collection of SEOs, content marketers, and data scientists who help sophisticated organizations leverage search data to improve their organic traffic volume, maximize the visibility of their content, and gain insights into their business, competition, and industry’s performance.

Ben:                 This week we’re going to continue our investigation into some of the overlapping trends in marketing like artificial intelligence, machine learning, marketing automation, and data science. Joining us today is Scott McLin who is a Senior Product Manager at Sojern, which is a data-driven performance marketing engine that services thousands of travel brands.

Ben:                 Prior to his role as Sojern, Scott has worked in various product and marketing roles for data-driven brands including Hotwire and StubHub. We’re excited to welcome Scott to the TrendSpotting Podcast, to help us understand the relationship between marketing automation and machine learning. Here’s our interview with Scott McLin, Senior Product Manager at Sojern.

Ben:                 Scott, welcome to the TrendSpotting Podcast.

Scott:               Thanks Ben. Great to be here. How are you doing?

Ben:                 I’m doing great. I am very appreciative of you being here, I know that this was a last minute request for you to join the podcast, and I appreciate you making time over the holiday season.

Scott:               My pleasure, happy to help out when you’re in a bind, so no, looking forward to this.

Ben:                 It’s the beauty of the internet when you want to find somebody who is an expert in marketing automation, and machine learning, turns out all you have to do is post on LinkedIn and people from your past just pop up and I’m excited to talk to you about not only what you do, but also what marketing executives need to know about understanding how marketing automation works and how they should think about it as it relates to machine learning and AI and some of the other topics we’ve covered.

Ben:                 But let’s start off with you telling us a little bit about not only what Sojern is, but what you do for the company.

Scott:               Perfect. Yeah, so Sojern, we’re a leader in the travel space. We basically think of ourselves as travel’s demand engine. For hotels, airlines, rental car companies, we key in on data partnerships that we have with other leaders in the space, and we get access to audience data and we can use that to really go out and target folks who based on our data and as we’ll get into more, our data science, target who we think is very likely to book.

Scott:               Basically, we want to put the right ad in front of the right person at the right time. That’s a big picture of what’s going on. I focus more specifically on our hotel side and more specifically with independent hotels, which we have several thousand hotels that we work with directly to drive direct bookings for their hotels, which they value tremendously, because it can be tricky working with the online travel agencies out there.

Scott:               These hotels would rather get direct bookings and that’s where we come in, and I specifically, my background is in search marketing and then I am now on the product side for Sojern, developing our SEM, our search engine marketing product. Basically, getting ads on Google and Bing, for hotels, in front of people who are looking for those properties, or looking for a hotel in that market.

Scott:               That’s what I key in on, and then also we’re working on a new metasearch pilot. Google Hotel ads, TripAdvisor, Kayak, basically anyway we can to get direct bookings for hotels is what we’re working to do and we also focus on more traditional display advertising, which was our flagship bread and butter offering for a long time, and then Facebook and Instagram. This year we launched, and has been a big success, as well.

Ben:                 Interesting. The things that stick out to me about the industry and the space that you’re working in, travel is hyper competitive and differentiation exists between all of the different hotels, in terms of feature set. Maybe not so much on the airlines. But when people are searching, it is very much about price. It seems like your role and the company’s focus is very much direct response related, and so you’re doing a fair amount of performance marketing. Is that right?

Scott:               Yes, correct. Exactly.

Ben:                 Okay, great. It makes sense that somebody that is an expert in, your background being search engine marketing, but marketing automation and understanding and working with engineering teams to implement some machine learning techniques would work in the travel space, in something that is performance marketing related. Talk to me a little bit about how you think about the definition of marketing automation. What do marketing execs need to know when they hear that term?

Scott:               That’s a great question. I think for me in particular, I think about how do we create and manage campaigns as efficiently and effectively as possible? Which basically, how it translated for us is we have almost 1000 hotels now, running SEM with us, through us. And to create hundreds of thousands of keywords and write ad copy, and manage bids on those hundreds of thousands of keywords, you’d need an army of folks doing that if you did not have automation.

Scott:               I think a lot of times that’s maybe the traditional agency approach, right? Just throw a bunch of people at it, and hope they can pull the right levers, which I think can work to some extent, but our goal is to provide great value to our customers. What we try and do is leverage our engineering and data science teams, and we have a great analyst team too, to basically automate as much as possible of the creation of those campaigns, creating those keywords and ad copy, and then also I think the real bread and butter and secret sauce we developed is a bidding and budgeting optimization logic.

Scott:               So, using data from across all of our customers to inform what’s the right bid for each keyword. You can use spreadsheets, you can have someone get in their manually, but to do that effectively across so many keywords is impossible for even a small army of people. You really need to leverage the data science, machine learning, automation, and the Google and Bing APIs to push all of those updates as we do daily.

Ben:                 A couple of things that stick out to me about what you’ve said is that the underlying I guess definition of marketing automation or your way of approaching is you’re using data to understand a customer’s propensity to buy. You’re aggregating multiple different data sources. You mentioned that Google and Bing have their APIs, and other data sources related to travel, to understand how valuable someone is, and then you’re building technology around the use of that data, to insert different variables.

Ben:                 Whether it be keyword targeting, whether it be copy, whether it be your bidding strategy on the search side. Is that a fair representation of how you’re thinking about machine learning?

Scott:               Yeah, totally. That’s a great summary and the only thing I’d add on top of that, you addressed a lot of great points, is our audience data. So basically getting information about what folks on the internet do, right? So, did they book a flight from SFO to JFK? And we get access to that data through our partner data, so if that happens, seconds later we can start putting an ad for one of our New York hotels in Times Square in front of that user, since figure they’ve got a pretty good chance of wanting to look for a hotel, at least compared to your average person, if they just booked a flight to New York.

Scott:               Then, on the search side, we’ll utilize things like RLSAs, which are remarketing lists for search ads. Basically if someone’s been to your website before, they’re a hot lead, right? You want to get in front of them as much as you can, so they don’t go to a competitor, or end up booking through an online travel agency instead of direct.

Ben:                 There’s a couple different components there where you’re saying that you’re getting transactional data to understand how when a customer is in a state of I guess the buying process, right? They’re past the evaluation, and the example that you used is they’ve already booked the flight to go to New York, they’re going to be in buy mode to book a hotel as well. You’re remarketing to your existing customers, to make sure that you’re staying in front of them.

Ben:                 I understand this from I guess more of the display and social campaigns, where you’re able to go chase people down across the internet once they exhibit a behavior, but how are you thinking about using search once someone has had a transactional moment that you’re interested in following up on, retargeting?

Scott:               Yeah, the good thing is that beyond just pure transactional behavior, like a purchase or something, we have a lot of amazing data which we have folks here at Sojern that could speak to it much better than I could, but just things, demographic data, crazy sophisticated information that’s at our fingertips.

Scott:               The good thing is yeah, it’s not just like, “Hey, they bought an airline ticket from X airline. Now we know to advertise on hotels.” There’s a lot of cool stuff there. From a search side, I think you’ve got a great point. It’s different, right? You really do need intent from the user to tap into them on search, right?

Scott:               Because they have to type something into that search box on Google or Bing. What we try to do is make sure that we’ve got a lot of good coverage on the keywords, especially around the brand terms of the hotel. Because if they’re looking for your hotel, you want to be in that spot, right? Like, you don’t want a competitor or someone else being there, or an online travel agency. You want that direct booking.

Scott:               The other side we be more of the traditional prospecting terms, so if someone’s just searching for hotels in Times Square, places to stay in Downtown San Diego with a pool, the cool thing is we have some additional data partnerships where we have amenity data, nearby location and point of interest data, neighborhood data, where we can create super long tail terms of hotels, so that if someone’s looking for, even if it’s not the specific hotel, but they’re looking for something your hotel offers, we can prioritize targeting that user on a search ad.

Ben:                 It’s interesting. Initially you started off saying marketing automation is primarily around the campaign launch and there’s your ROI optimization, your bidding strategy, and you’re also doing some creative management. But really outside of just using event based triggers when you’re going to launch a campaign, you’re able to do some personalization and then you’re also increasing your engagement through your remarketing. You’re managing not only your impression levels, you’re optimizing your creative to create the right experience for a specific person and then you’re launching your campaigns when you think there is a high propensity to buy.

Scott:               Yeah, I think that’s right on track. Basically, we try to have a mix of prospecting, right? Like, going after more upper funnel users, to maybe if they weren’t thinking about that hotel specifically, say we could get that ad in front of ’em, so they see that and start thinking about it or click-through, visit the site.

Scott:               Then, there’s the lower funnel stuff which is, “Hey, they’ve been to the site before” where we can really go after them with RLSAs on Google and Bing.

Ben:                 Sorry, you said RLSA?

Scott:               Yeah, so that’s a good thing. It’s a cool feature that Google has and Bing have that basically you can set up tags and track user behavior on your website. Then, you can either, you can both customize your ad copy, based on their behavior on your website, and/or modify your bidding tactics.

Scott:               Say they’ve added something to their cart, that’s when you can do something like, “X item’s waiting for you in your cart.” Then you could also bid say twice as much, ’cause you figure this person’s almost at the finish line, I’m willing to pay.

Ben:                 They’re farther down the funnel, so you’re willing to invest more.

Scott:               Exactly. Your conversion rate’s obviously going to be higher on average for customers like that.

Ben:                 Yeah. Tell me a little bit about how you work cross-functionally. You’re in the product marketing organization. How does your role overlap with your data science team, with your engineering team? The show’s positioned for marketing executives, what do the marketing leaders need to understand about putting together a team that is capable of launching and scaling effective marketing automation services?

Scott:               Yeah, totally. It’s a great question. As you say, I’m on the product team where I work pretty closely with our product marketing team, but then I also work closely with our data science and engineering team. That’s what I really love about the product role, is you need to be very cross-functional, where a lot of different hats. I coordinate with our marketing team, too.

Scott:               On the product marketing side, it’s working with them to make sure we can educate the customer on the value we’re driving them. Taking a new product to market, always super exciting, but I’m not great at writing out copy or making things sound fancy, so I love their expertise and they’re great to work with.

Scott:               Then on the engineering and data science side, that’s more working with them to figure out how do we scale all of this? We can develop our AdWords and Bing API services, there’s a lot of piping that needs to be put in place. You don’t really anticipate how much work that can take, and then my favorite part personally is just really digging in with the data science team to come up with algorithms, to boost campaign performance, just because my background is as I was saying, in search, and I did a lot of hands-on optimization.

Scott:               Nothing feels better than you make some changes and you come in and a week later, you’re-

Ben:                 The goal posts are just that much closer.

Scott:               Exactly, yeah. That’s probably my favorite part but it’s all fun just ’cause I take the Marty Cagan approach of you’re the CEO of your product. That’s what we do here, and again, nothing’s better than coming in and seeing your month over month revenue increase, and it’s a great feeling, and a lot of fun. I think you definitely have to make sure that everyone’s in alignment, make sure that data science and engineering understand what you’re doing, and you’ve got a clear roadmap, and plan to execute what you’re trying to do.

Ben:                 I think that’s the complicated part, or maybe the intimidating part for some marketers who let’s just say are less digitally inclined, or traditional brand marketers, when they think about marketing automation and they think about data science, they’re thinking advanced mathematics, they’re thinking about teams of engineers, data scientists, this all sounds big, complicated, and expensive.

Ben:                 What are some of the ways that you suggest marketing brands that have more of a traditional approach can leverage some marketing automation best practices, and start integrating data science into their marketing?

Scott:               That’s a great question, yeah. I definitely don’t have a perfect solution for the expense side, right? Especially if you’re in the Bay Area, it’s going to be tough to find any engineer data scientist for under six figures, but I think the key would be investing wisely. I think you definitely need at least one person on the platform side, right? Who can build you the piping into the AdWords API, or Marketo, Salesforce, whatever kind of system you use.

Scott:               Someone who can look at some API documentation and figure it out, and build that stuff well and reliably. I think that’s something you for sure need to invest in if you want to do automation, or at least the approach we’ve done where we’re directly pushing bids to AdWords and Bing. Obviously there’s different approaches, like you can use a Marin or Kenshoo, things like that. But that’s the way we’ve chosen to do it, because we have more control over it.

Ben:                 Talk to me a little bit about that trade off. I was just ask about to ask you as a follow-up what are some of the tools and technologies that are available to people that don’t have the access to-

Scott:               Totally.

Ben:                 … building an in-house team? You mentioned Marin and Kenshoo. What do you know about those platforms and how do they work?

Scott:               Yeah, so they’re platforms I’ve used before in my previous marketing life. They’re great, because if say you need to manage several campaigns, like thousands of keywords, it provides a lot of tools to really enable that. I think it can help out if you have someone who needs to manage a bunch of campaigns. To be honest, I don’t think that there’s as much opportunity for pure automation.

Scott:               They have some cool tricks that can make things better, but if you’re looking for an automated solution like we have, we have an algorithm that calculates bids every day and pushes it via API to the engines, it’s amazing, I love it, but Marin and Kenshoo would be more like if you can’t really invest in that kind of team.

Scott:               On the data science side, I think that you talk about how it can be a little intimidating for marketing folks, and I totally agree. I think that data science can be a lot of people with PhDs talking a lot of theoretical stuff, and at times it is like that, but that part’s actually pretty interesting to me, too, and I think the key is finding the right team who can speak to that stuff, and explain it to you in a way that you can understand, that’s also not condescending.

Scott:               I’ve lucked out in terms of the team that I work with. They’re amazing. ‘Cause a lot of times, at least we’ve seen so far, the algorithms don’t have to be insanely complex. They just need to do the right thing. They need to be smart. I’ve manually managed campaigns for a big part of my working life, but we have 800 hotels doing search marketing with us, so there’s no way I can do it for everyone.

Scott:               Basically, can you find an approach that would roughly approximate what I would do on one account for 800 accounts? And obviously the savings in terms of not having to pay someone like me, you make back a lot of money real quickly if you can implement a proper approach. I would also say that to go back to maybe you can’t afford a data scientist and an engineer right now, there are also some approaches, like you can do through a Marin and a Kenshoo, but also through Google, and I think to some extent, Bing.

Scott:               They have some automated bidding tactics that you can employ on your campaigns, where it basically uses machine learning to come up with what they deem the right bid.

Ben:                 Sorry to interrupt you. I’m always a little hesitant to set the automated bid on any of my search marketing campaigns, because to me, you’re asking the service or person, it’s not a person, the service that you’re paying to provide you marketing results, to optimize for your best interests. When in reality, they’re running a business and I always just feel like they’re going to optimize for their highest ROI.

Ben:                 Maybe I’m skeptical, but that’s a concern for me. I think that what I’m taking out of what you’re talking about is data science is a complicated process, and it does require mathematicians and PhDs and a lot of the theoretical work that goes into building out a robust data science program, can be very difficult to communicate.

Ben:                 So, as you’re thinking about building a data science team, or if you’re trying to implement marketing automation, there is a middle layer. There is the set of tools and services, the Kenshoos, the Marin Softwares way of the world that have built out these types of software that are more plug and play, and once you start getting to the point where you have scaled those services, then maybe it’s time to start thinking about building those programs in-house, and when you do that, you really need to have a layered approach of the product managers, like Scott who’s kind enough to join us, and then the data scientists and engineers as well.

Ben:                 I do think that there is a relationship and a handshake between the business and the logic that goes into building the algorithm, and the data science and engineers that goes into operationalizing what you’re trying to accomplish.

Scott:               I 100% agree, and in particular on your first point around actually employing Google’s logic, because if you go to your mechanic and bring your car in, there’s a pretty good chance they’re going to tell you you need something fixed, even if maybe you don’t really need something fixed. They’re incentivized to spend more, and things around where you do need to be a little more aggressive on spend when you launch some things, so you can get more data points around conversions, so you can optimize better, but I am always hesitant to really rely on those.

Scott:               I think they’re fine to test out, and I think you can often see good results from it, but I agree, there needs to be just a healthy level of skepticism, because of the motivations.

Ben:                 Yeah. I totally agree with you. It’s not necessarily that Google or Microsoft, Bing, have bad intentions. They also don’t have the visibility into your business and the understanding of the difference of one conversion versus the next. At the end of the day, you as the business owner are going to have a better understanding of what drives the type of business results that you want, as opposed to making the assumption that all conversions are the same. You talked about incentives, there’s just a misalignment of incentives when you’re handing over the way that you allocate your budget to the person that is receiving that allocation.

Scott:               Totally agree. That makes sense.

Ben:                 Any last words on what marketing execs need to think about as they’re trying to rationalize the difference between marketing automation, machine learning, data science, just any last tidbits about how to segment those three topics?

Scott:               I think at the end of the day, it’s try and scope out what your end goals are. Depending on how ambitious and large those are, then you want to probably consider going after building a team, a data science and/or engineering team that can implement that. But I think if your goal is to do something that’ll only affect revenue by X amount, and the investment isn’t quite worth it in terms of salaries et cetera, just think about what you hope to get out of it, and then find out if the investment’s worth it.

Scott:               Then, do a little research into what has potentially been done in the space before, right? You’re not necessarily reinventing the wheel if there are solutions out there that can get you part of the way with much lower investments.

Ben:                 Okay. I appreciate you giving us your thoughts on marketing automation and how it relates to machine learning and data science. That wraps up this episode of the TrendSpotting Podcast. Thanks again to Scott McLin for joining us. If you’d like to learn more about Scott or Sojern, you can click on the link in our show notes to Scott’s bio, or visit

Ben:                 If you’re interested in spotting more marketing trends, or if you’re interested in Searchmetrics, the creator of the TrendSpotting Podcast, click the link in our show notes to see our podcast content archive, or you can go to If you have questions or if you’d like to be a guest on the TrendSpotting Podcast, feel free to fill out the contact us form on the website, and if you’ve enjoyed this podcast and you’re feeling generous, we’d love for you to leave us a review in the Apple iTunes Store.

Ben:                 Lastly, if you like this podcast and you want a regular stream of marketing driven insights in your podcast feed, hit the subscribe button in your podcast app and we’ll be back in your feed next week. Okay, that’s it for today, but until next time, remember, it’s a data driven world out there, and the team at Searchmetrics is here to point you in the right direction.