Gaming the Web: How Google’s Search Engine Got Smarter – The INBOUND Series


Daniel Russell serves as COO of Go Fish Digital, a cutting-edge marketing agency serving small, medium, and large companies and providing everything digital except video and email marketing.


In this interview, Daniel, an attorney, explains how law school provided him with not only “a business background,” but also “the rules of the game.” He notes that Bill Slawski, Go Fish Digital’s Director of Research, is also an attorney. Bill studies Google’s patents for insight into how Google is changing things. The company’s founders started their careers as FBI contactors before becoming marketers, so the company, now with 42 employees, has some interesting roots.


Daniel presented, “Google’s AI Is Smarter than You: What Does That Mean for AdWords and SEO Campaigns?” at HubSpot’s Inbound 2018. He talked about how Google, as part of its DeepMind project, developed the AlphaGo computer. AlphaGo beat top players at “Go,” an ancient Chinese boardgame with a “near infinite number of moves and possibilities” and no rules for “your next move.” The learning algorithms, neural networks, and artificial intelligence that came out of AlphaGo research have dramatically increased the effectiveness and accuracy of Google’s voice search. Daniel also talked about AI in relation to Google ad quality scores, SEO, and SEM, how AI might potentially affect them in the future, how AI has changed the search process, and the nonconformity of AI use on different platforms.


Influencer marketing, currently the fastest growing part of Go Fish Digital’s business, intrigues Daniel. Consumers often tune out advertised content. When an influencer, a trusted “name” with whom target consumers identify, promotes a product/company, these consumers find the information more credible, even though they know that the advertised companies sponsor influencers. Today, influencer marketing is not necessarily just celebrities talking about your company. It may YouTubers talking about your website on their videos, Instagram postings, and less so, Facebook pages. LinkedIn can be effective for B2B services.


Daniel is available on Twitter at @dnlrussell or by email at His company website, Go Fish Digital, can be found at

ROB: Welcome to the Marketing Agency Leadership Podcast. I am your host, Rob Kischuk. I am live in the


WeWork at HubSpot’s Inbound Conference, and I am joined right now by Daniel Russell. Daniel is the COO of Go Fish Digital. They’re based in Washington, D.C., and a new office in Raleigh, North Carolina. Welcome, Daniel.


DANIEL: Thanks. Glad to be here.


ROB: Great to have you on the podcast. Why don’t you start off by telling us a little bit about Go Fish Digital and about what makes Go Fish great? What are your superpowers?


DANIEL: We are an almost full stop digital marketing agency. We do just about everything digital except for video. We don’t do video, and we don’t do email marketing. But everything else – advertising, SEO, PPC, display targeting – pretty much everything you can think of on the digital side, we do.


In terms of our specialties, we tend to be pretty cutting edge. Part of what I’m speaking about today at Inbound is about how artificial intelligence is affecting SEO and SEM, and that is our MO – to be ahead of the curve if we can.


We have a Director of Research who is an attorney, like me. He studies Google’s patents to give us some insight into how Google is changing things. So that would be probably our main superpower. We’re really on the edge.


ROB: That is pretty interesting. There’s been a lot of changes in SEO from this thing where you can have a top 10 list of “things you should do.” Are there still things that people should definitely do for SEO? Is the rest of SEO lost?


DANIEL: That’s an interesting question. There are reports, I feel like every week, that SEO is dead, the old way of SEO is dead. It’s not quite true. Those reports are overstated. [laughs]


The big invention that Larry Page came up with when he was inventing the Google search algorithm was page ranked based off of link authority. Backlinks are still the biggest deal. It’s been that way since the ’90s, and it’s probably going to keep being that way.


The things that are newer is that it used to be that you were playing almost a chess game with Google’s algorithm. There were very specific moves that you could make – but to the algorithm. You were doing optimization to the specific algorithm.


The algorithm’s gotten smart enough now that you can optimize to people – which, to be honest, is sometimes easier because we work with people every day, and Google doesn’t reveal everything that’s in their algorithm.


What Google usually says is that if you create content that people will like, then Google will like it. I think for the most part, that is true. That is definitely oversimplified; obviously it still makes sense to pay attention to algorithm updates and those sorts of things. Links still matter the most, as they always have, but now you can have more of a human element.


ROB: There’s sort of an interesting race to authenticity. You mentioned backlinks still work, but increasingly – I think, and you can tell me what you know, because you know better than I do – but inauthentic backlinks are increasingly unhelpful, right? What I think I would say is being helpful to real people using the right keywords is modern SEO. Is that what you would say?


DANIEL: Yeah, I would. Trying to match the searcher’s intent. And per your backlink point, it does seem that Google is getting better at telling links apart. Obviously, a long time ago they introduced the whole follow link versus no follow link to try to get at that at first.


But now they’re using more advanced methods like what publications it’s coming from. They supposedly have a list of publications that are offering paid links that they devalue. So Google supposedly is getting smarter.


ROB: How do two people go from law school to digital marketing, by the way?


DANIEL: [laughs] For me, I always wanted to end up in business. I knew that since a younger age. The question was, how should I get there?


I started asking around to different businesspeople that I looked up to, and some of them had gone to law school instead of going to business school. Some of them didn’t go to any school, some went to business school, and some went to law school.


The ones that went to law school would tell me, you’d be surprised, but you learn a lot of the same things you do in business school. At the end of it, you come out and you can actually do something with it. You can go to court if you need to. It helps you understand the rules of the game. I talked to a couple of people that went through business school who actually said the same thing. They said, “I kind of wish I’d gone to law school.”


So, the goal was always to end up in business; the question was the route, if I wanted to be an attorney for a couple years or not. Luckily, this opportunity with Go Fish Digital came up. I had had some background in marketing before, in market research, so it made a lot of sense. I leveraged my legal experience into joining a company and leapfrogging the attorney steps.


I did work for a couple law firms while in law school, and actually before law school, and I’m technically our in-house legal counsel. I’m barred in Maryland, so if need be, I am still an attorney.


That was my route. And then Bill Slawski is our Director of Research, who also went to law school. I think his story is – less planned out is not quite the word, but I don’t know if he just knew he was going to end up in search. I think what ended up happening is as he was in law school and as he was getting out of law school, he started looking into patents specifically.


When Google came around, he was just really fascinated by it. Before long, on all of these old school, 1996 forums, he was participating every day, talking about patents and Google, and someone said, “Hey Bill, you should do this for your job.” [laughs] So that’s how he made the leap.


ROB: When did the agency start? When did you join, and what was your role when you joined?


DANIEL: The agency technically started 12 years ago. It was cofounded by my coworkers Dan and Brian, who started it in Washington, D.C. They were both working for the FBI at the time.


ROB: [laughs] This is a high test agency here. You’ve got FBI, you’ve got lawyers. I’m a little scared.


DANIEL: Yeah. Well, they were contractors for the FBI. They were consultants for the FBI. They started it 12 years ago. The rubber really hit the road about 6-7 years ago, and I joined 5 years ago. When I joined, I think there were maybe six or seven people in the agency. Now we’ve grown a bit. We’re not a massive agency, but we just hired I think our 42nd employee last week.


ROB: Congratulations.


DANIEL: Thanks.


ROB: Did you know these guys before you joined?


DANIEL: I didn’t.


ROB: So, you go from being one of the first 10 employees to now being a very trusted member of the leadership team. That’s sometimes an interesting dynamic with two partners/founders in the business. What do you think that journey was to that degree of trust and that shared leadership that you have now?


DANIEL: I started out with them doing some side consulting from a legal perspective. I was introduced to them. I said, “Hey, I have a market research background. I understand marketing in a big way – not all the details, but in a high level way, and I now have all this legal knowledge and time to use it.” That was the initial part. I think that’s where we probably started building some trust.


Then I started working for them the summer in between my second and third year of law school. For a start, they were letting me talk to larger clients because it was a small agency, and before long they were having me be the one to go speak with everybody.


Sometimes I think they thought that I could talk with the room maybe even better than they could, or if our powers combined, their technical knowledge with my background, we could make a good team. And I think we did make a really good team. That’s where it started, and it’s grown from there.


ROB: Very, very interesting. You mentioned you do just about everything except for video and email. How have you come to decide and embrace the things that you do not do?


DANIEL: We like to be very frank with ourselves about our abilities. I think that helps us, and then it helps our clients. If we’re real with what we can’t do, then we never over-promise, and if we really double down on what we can do, then we over-deliver. We under-promise, over-deliver on those items. That’s the motivation behind it.


Part of that, too, is that things like video are just really hard. [laughs] You need equipment, you need people dedicated to it. Once a video gets online, we can probably still handle that. That blends pretty well into our other verticals. But actually creating the video to be used on a digital space is a big undertaking that we don’t have.


ROB: Do you partner in those other areas? Do you refer things out? Do you white label what other people do to execute on those sorts of things?


DANIEL: Email, we have not executed on and amazingly haven’t had too many requests. I don’t know if it means that the companies we work with just are not interested in email or have email down internally, but not as many requests there. We get more requests for video.


We have met a couple of groups that do video well, but usually the requests that our clients have vary widely. We once had a client request our help with a Super Bowl ad, and we said, “that is so far out of our wheelhouse…”


ROB: [laughs] That would be fun, but . . .


DANIEL: Yeah, it would be a blast. I understand they’re very expensive to buy. But that’s just so far out of our wheelhouse. And then we have other clients who are saying, “Hey, can you make an explainer video for us?” It’s like, no, you can go on Fiverr and find explainer video services.


ROB: Is there a sweet spot of a typical client for you, or some examples you can share?


DANIEL: Yeah. I think our path has been probably a fairly traditional agency path. We started with a lot of groups that were either local to the area or referrals from past clients who were local to the area. Then about 5 years ago, right around the time I joined, we became the Agency of Record for GEICO, doing their SEO.


Since then, we’ve added a lot more of those larger brands, international brands. We’ve done work for Lulu Lemon and Marriott Hotels and different groups like that.


I would say we really enjoy a mix, to be honest. Our teams are usually working on a combination of different projects just because we find that sometimes the experimenting you can do on a smaller website really benefits bigger websites where it’s more locked down. Our teams actually really enjoy being able to work with a small local consulting group’s website, and then at the same time be working on GEICO’s website.


ROB: Very cool. That’s a nice little bunch of logos there on the website.


DANIEL: Yeah, it’s great.


ROB: You are giving a talk later today entitled, “Google’s AI Is Smarter than You: What Does That Mean for AdWords and SEO Campaigns?” Tell us a little bit about what you’re going to be sharing and what we should be thinking about.


DANIEL: Well, the impetus behind all this was actually back in 2016, when Google’s AlphaGo – if you recall, they were in the news for a while; it was part of the DeepMind project – Google’s AlphaGo beat the top Go players in the world at their own game.


I was really surprised by that. We had made chess algorithms that had beaten people before. Humanity had made chess algorithms, not me. [laughs]


ROB: Not you personally.


DANIEL: But Go specifically has a near infinite number of moves and possibilities, so there’s no way to have rules for every step. So when I saw that, I thought, why is Google so interested in this board game computer? What is it that they’re figuring out?


Obviously other people had already been talking about this. I started looking into that, started looking at Google’s patents, what they were filing that had to do with artificial intelligence. I started to realize and put the pieces together of how they could use the same system that beat a board game to amplify voice search and to make that a thing that mattered.


Before, voice search was so difficult and no one used it because it would return incorrect results to your queries half the time. But now, largely thanks to learning algorithms and neural networks and artificial intelligence, people use voice search all the time. It’s actually fairly effective.


That is largely what I’ll be talking about. Things like voice search, things like quality scores in Google ads, and how artificial intelligence is affecting those different items and how it might in the future.


ROB: How does that then play into – I think notably, Facebook uses a good bit of AI in terms of targeting and refining campaigns. How does that apply on the targeting side to Google or social platforms when we’re doing paid ads (LinkedIn, Twitter, etc.)?


DANIEL: I would say, on those social platforms, it’s a little interesting. As we know, for the most part Silicon Valley does not share very well. Elon Musk has published some of his patents for free use before, and they have had some sharing initiatives, but for the most part they’re trying to lock down as many of their own personal AI wonks as they can.


Supposedly one of the highest paid positions – it used to not be, but is now one of the highest paid positions – is a coding neural network professor, because the competition to get these guys from Caltech and Berkeley and MIT is so high because everybody wants their own private team so they don’t have to share their AI knowledge.


Facebook’s AI is surprisingly different than Google’s. It’s been very photo-centric – which Google started to get into with Google Images, but it was not something right off the bat that they were concerned with like Facebook was.


I think a lot of the same stats—that they rely on to test and change up their algorithms—that the neural networks use are similar. They’re largely looking at things like time on that page, click-through rate for ads, those types of things. But then how they change it is actually a very private thing for them based off of that.


So unfortunately it’s not a 1:1 translation from Google to those social platforms.


ROB: Sure. Some of these tactics you have out there – you have your website audiences, you have your email audiences, you have lookalike audiences built on those, and you’re optimizing for objectives. Are there tactics that people should just stop that were effective 5 years ago, in terms of targeting ads that are kind of pointless and nonsense, and you can just accept that the AI is better than you at achieving your goals?


DANIEL: Yeah. It’s a bit of a tough battle. In my opinion, some of the least effective old means of tracking ad success are things like last click. Attribution models, largely.


AdWords and Facebook as well are trying to get better about that and trying to make it so that you don’t have to worry as much about attribution models. They just give you a great attribution model, and then you can step away and you don’t have to worry about last click or first click or touchpoints or anything like that. They can track things for you.


The problem with that is it starts to get into privacy concern areas. How are they going to create a much better attribution model than us without really keeping a tight eye on – all right, this person just saw your display ad that you’re running through the Google ad network, on a Google partner website. Now they just saw your ad in the search results for the first time, and now’s the second time and they clicked. You have AdWords on there, so we can see what pages they’re navigating to, but they haven’t called yet. But then you have call tracking set up.


Anyway, it’s one of those things where I think things are going to get a little more Big Brother-ish with things like AdWords as the artificial intelligence improves, but hopefully it will help us so that we can just trust it. It’s smarter than us. The attribution is better than what we could do.


ROB: What is the attribution model that people should . . . if they turn on one thing in Google and change it from last touch, what should they do?


DANIEL: That’s tough without seeing the setup of their business. To give a couple of examples, we work with some really large ecommerce websites that have a very – and for good reason – a different setup in AdWords and attribution setup than we do for a large law firm that we work for.


Part of that comes down to—you can’t advertise the same for a law firm that you can for a marketplace. There’s restrictions around that. But part of it also comes down to that a conversion for a law firm might be a 6-year life cycle for that client and might average a million dollars a year, whereas a click on to buy some new biking shoes will get you $80 and maybe a repeat customer – but then that relies on a totally different funnel.


So usually I don’t give a one-size-fits-all. But needless to say, last click is usually not my fave. [laughs]


ROB: Totally fair. So you’ve been with the company 5 years now. You didn’t start the thing, but you’re awfully responsible now. What are some things you have learned in growing in your role and in growing the agency that you would do differently if you were starting in your current role 5 years later?


DANIEL: Interesting. I thought you were going to ask what we did well, and I was going to say, “Well, let me tell ya!” [laughs]


I think what we would’ve done differently – we are very slow to hire, which I think is one of our strong points, but at the same time for a good chunk of the agency’s life, we’ve been trying to hire too far after we need people. We hurt. It’s like, “Oh man, we can’t take any more clients right now because we can’t hire fast enough.”


We’ve gotten better about that. We call it a bench. We have a bench of strong candidates that we could potentially have hired and onboarded within a month or two – for key roles, at least. That way, if a lot of referrals come in or we have some clients that really want to increase their scope, we can expand to reach that. Not in an irresponsible way, which is what we’re worried about, but in a more responsible way than doing it 6 months later.


If I could go back, I would start that sooner, because it’s definitely slowed our growth a bit. I can’t say exactly where we’d be had we started that 5 years ago, but it definitely slowed our growth.


ROB: Knowing, “We’re going to need somebody on creative, we’re going to need a strategist, we’re going to need an account manager” – you know those roles. What’s a role that you’ve had to hire for that you didn’t expect? A new role that you hadn’t been able to think of a bench person for, and you’re just like “gosh, we need one of those.”

DANIEL: One that surprised me but did not surprise my partners, just because I think they had felt this pain point before I started, was we needed some dedicated copywriting. The reason I was surprised by that is we had never offered copywriting as its own separate service. If we were going to do copywriting, it was usually done by the in-house team we were working with, and we consulted on the SEO front with that.


But adding some dedicated copywriting ability has really helped us quite a bit. A lot of that has come down to not every SEO [expert] is a good writer and not every in-house team is a great writer. When we can combine our SEO expertise with some in-house expertise on the writing front, it’s made our on-page adjustments quicker and more effective.


The benefits of that, I’ve been surprised by. And how busy the copywriting people are. [laughs]


ROB: [laughs] You thought they might just be on and off a little bit? “Do we always need copywriting?”


DANIEL: Yeah, “Can we work with contractors or the in-house team?” They’re like, “No, let’s get a copywriter.” Sure enough, it’s been great.


ROB: Really interesting. We’ve talked a little bit about AI; what are some other things coming up in marketing or, in particular, for Go Fish that you are looking forward to/excited about?


DANIEL: I’m excited for a couple of things. Influencer marketing is an area I’m kind of excited about. It’s the fastest growing part of our business right now. I think that’s, one, because it’s finally starting to reach more of a saturation point where people understand what it is.


But the other reason I’m excited about it is because it’s helping tackle this problem that advertisers have been facing for a while, which is people reject advertised content. For some reason, people want something more organic. They don’t want to see those ads. “Why am I going to trust this ad? I’d rather hear it from somebody that I trust.” Then the influencer comes in.


Now, I don’t know – I think there may still be some flawed logic there, because you’re sponsoring these things. Companies are sponsoring these influencers. But still, it’s kind of like the old having Michael Jordan and Larry Bird competing for your McDonald’s ad. Now it’s having these YouTubers talk about your website on their videos.


ROB: Is that where you’re seeing the most momentum on influencers? On YouTube?


DANIEL: YouTube, Instagram. Facebook not as much. And then LinkedIn can be useful depending on the client. More for B-to-B purposes.


ROB: Anything on Twitch? [laughs]


DANIEL: Yes. Twitch can be extremely useful in the way that LinkedIn is niche for businesspeople, Twitch is niche for a very specific, more male, younger audience. But yes.


ROB: For sure. The influencer, I think it’s been through a trough of disillusionment. There’s a spike and then a trough and then some comeback. What are some of the other things you think have changed? People clearly need to find other ways to get messages out, but where’s the credibility coming back in?


DANIEL: I think it’s because it took a little bit of time for the influencers themselves to figure it out. Not saying that they were to blame at all for the trough, because nobody quite had it figured out because it was still fresh.


ROB: Right, and if someone writes you a check for $10,000, you say, “yes.”


DANIEL: Yes. That’s I think what it came down to. Influencers realized they had to say, “no.” They had to say ‘no’ sometimes to those $10,000 checks so they could collect the $20,000 checks later.


I think it’s matured enough to that point where influencers can maintain trust with their audiences because they can say, “This is sponsored, but I test the sponsorship first, and I don’t take the money if I don’t like it.” That way it’s a way for them to maintain trust with their audience, in theory, but still make it a business.


ROB: How do you navigate the potential for influencers – some of them are very vanilla. You could really have them push anything out there, but it might not be meaningful for them to advocate for one brand or another.


Then at the other end of the spectrum – just last night, Nike dropped their 2-minute ad with Colin Kaepernick, and I think Venus Williams is in there, and some nobody athletes too, and the defender from UCF who has one hand and he’s in the NFL. But that’s polarizing.


How do you navigate the polarizing/not polarizing/boring influencer, almost, landscape?


DANIEL: I think it’s interesting. To the point earlier about what’s started to click with influencer, I think some influencers realize that they can’t always be ‘full vanilla.’ If you are trying to curate an audience of everybody who would be interested in every product, that doesn’t quite work as well.


To give an example, one of the influencers that we’ve worked with in the past specializes in Las Vegas. She could’ve gone full into parties, into drinking, into that lifestyle everywhere. But instead she said, “I’m not into it everywhere. I am into the parties in Las Vegas only. That is my scene. That’s what I photograph, that’s where I go.”


That is, I think, smart on her part because there’s actually a lot of companies that are interested in Las Vegas specifically. Because she is Las Vegas specific, she has an audience that cares about Las Vegas. We had a client who specializes in these hangover cocktails the day after a party to get you going again. We partnered with her to do this influencer campaign, and the engagement was huge.


Even though she said, “This is an ad,” they got more likes than her typical non-ad post because it fit that very niche audience. People were so interested in, “Oh man, this is actually a really cool product.” I’m so happy this sponsorship worked out.


That’s one of the reasons why it can be dangerous to be vanilla.


For the polarizing, the old adage of “all press is good press” obviously has never been fully true, but to an extent, everybody has been talking about Nike. Everybody’s been talking about Nike, and at the very least, people who have been destroying their Nike shoes are getting them out of circulation now. [laughs] So now there’s more room for more Nike stuff.


I think generally, companies usually don’t want to be polarizing unless they already have clout like Nike does, so generally when we’re working with influencers, we’re not looking for someone to stir the pot. But that being said, I think it can still be a really effective strategy, for sure.


ROB: Very, very cool. Daniel, when people want to find you and find Go Fish Digital, how should they find you?


DANIEL: I’m on Twitter. I usually am talking about the stuff that we’re looking into at Go Fish, what we’re writing about. You can also usually find where I might be speaking coming up on Twitter. The handle is @dnlrussell. And then my email is public; it’s just Those are usually the best ways.


ROB: Fantastic. Thank you for dropping by, thank you for sharing. It’s been good to meet and learn from you.


DANIEL: Yeah, nice to meet you as well.


ROB: Thanks for your time.

Thank you for listening. The Marketing Agency Leadership Podcast is presented by Converge. Converge helps digital marketing agencies and brands automate their reporting so they can be more profitable, accurate, and responsive. To learn more about how Converge can automate your marketing reporting, email, or visit us on the web at

Leave a Reply

Your email address will not be published. Required fields are marked *