Episode 208 | Posted on

Best Practices in Messenger Marketing with Larry Kim

Is email the fax machine of our era? Are we living through a tectonic shift in how people communicate online? Can a chatbot really get you conversions? There are many commentators who have proclaimed chat and chatbots as the future of marketing. 

But today I’m talking to someone who is actually creating that future. Larry Kim is the founder of Mobile Monkey, a Facebook Messenger marketing platform. And he’s the founder of WordStream, a PPC marketing software company employing over 300 people and managing over $1B dollars in ad spend for tens of thousands of customers. Larry has been working at the coalface of AI innovation and natural language processing and has gleaned many important insights about the benefits and limits of automation and its role in marketing, along with some startling stats on the effectiveness of email marketing versus chat marketing. 

If you’re interested in cutting through the hype and hearing from someone who has grappled with the real-world implications also implementing these technologies, then this episode number 208, is a must-listen. Similarly, if you’ve been on the fence about chat and chatbots, you won’t want to miss what Larry has to say.


Larry, it’s so great to have you on the show.

Awesome. Thanks for having me. Great to be here.

Let’s first of all, let’s talk about MobileMonkey. You had a really good run there with WordStream and then you exited. I’m sure it was a successful exit but we could get more details in a bit. Then you jumped right into a new startup. Wow. No time off or anything.

I think that’s pretty common for entrepreneurs. The goal wasn’t just to make a ton of money. It was the journey and I was very interested, as a personal challenge, to see if it’s possible to build an even bigger company in perhaps half the time or less time than the last. So far so good.

All right. Let’s hear about what MobileMonkey is and what’s the value prop.

Sure. MobileMonkey is the world’s fastest-growing Facebook Messenger marketing solution provider. Think about how you use messenger or other chat platforms to communicate with your friends and family. What MobileMonkey lets you do is it allows businesses to connect with their customers on the chat platforms that they use everyday.

We’re currently a messenger platform although, the hope is to expand to other chat platforms like Google, Apple, or SMS in the near future. It’s really exciting. There’s a big shift in communication preferences in terms of how people like to communicate. People overwhelmingly prefer messaging over email. Email feels like a fax machine where you have long emails with attachments and stuff like this.

For more informal communication, it’s always messaging on one platform or another. If you’re a business and you’re still sending out stupid emails with 5% open rates, that’s silly. Sure, you should keep doing that but you should also be investigating messaging because that’s the new marketing channel where people actually engage and read content. You know how it is in marketing. When a channel is new, there’s a ton of value in being first. That’s where we’re at and what the company is about.

You’re lucky to get maybe 20% or 25% in email. What makes your platform unique or more valuable compared to ManyChat or whatever else is out there for Facebook messaging platforms for marketers?

Half of our customers are actually people who gave up on ManyChat. Usually, you’re supposed to not be negative. Let me just say, MobileMonkey is a very stable platform. It will not crash or have crazy unexpected results. The bot builder is the easiest to use in the world. It’s not like you have to create this huge spaghetti code. It’s way easier than these crazy tools out there and in terms of features and functions. We’ve got a really powerful integration with WordPress in terms of natural language processing. We have an AI that will figure out what the user is asking and make it easy to surface the correct answer to the users using natural language processing. These are all things that are unique to MobileMonkey.

In terms of pricing, other chatbot companies charge by the number of contacts and we think that’s really unfair because most of these contacts aren’t even reachable. It just seems like a raw deal. We have a free forever plan but there are some advanced features and we don’t charge more if you have 50,000 contacts or 100,000 contacts. It’s substantially cheaper. Generally, I think that those are the key things—more features, easier to use, better pricing.

Where there’s a human workflow that needs to execute queries manually, synthesize data, and formulate a hypothesis, lies an opportunity for automation. Click To Tweet

What sort of pricing might a business owner expect to spend with you?

Most of them are zero. Maybe we’re giving it away from free. It sounds like a high 90% in the 90s. The people who are paying for this stuff, there’s certain bells in that swells around Facebook Messenger ads and stuff like this. If you’re spending money on Facebook Messenger ads, it’s $49 a month. You’re probably spending more than that on Facebook ads so it’s not really a significant cost. Otherwise, most people use it for free and it’s not like buying some marketing automation platform for $5000 a month or something. It’s between $0 and $50.

This isn’t the Marketo of chatbot solutions and from the sense of the pricing.

I think it’s an interesting analogy from the perspective of features of functionality and that you can reimagine most of the email marketing playbook but delivered via messaging as opposed to emails. Drip campaigns, that’s something that we support. Sending out a sequence of whatever, five messages over a certain period of time or based on certain triggers or activities. This is a pretty common thing that people do in email marketing like if you buy something, they’ll try to upsell you a warranty for that thing or some related products. You can do that through messaging as well.

For the same reason why it works in email, it works five times better in chat because they’re more likely to see it. It’s a good analogy from a functionality perspective. Although, somehow these marketing automation companies went bonkers and started charging crazy prices along the journey. We view this as more of a Mailchimp or a Constant Contact where it’s free or $50, that kind of thing.

All right. Very cool. Do you have any examples of really clever campaigns that clients and customers have been using trigger-based messaging for?

There’s the whole e-commerce sector using it like crazy for abandoned shopping carts. 10 minutes or 5 minutes after you leave the site, if you have something in your cart, fire off a message. There are upsell bots, five minutes after buying something, send them a related product. That’s one. Other areas are in professional services like dentists, realtors, or car dealers where you’re trying to book appointments. You can expose your calendar via messaging so that people can book appointments and then have the messaging platform automatically trigger a reminder message like, “Hey, by the way, your dentist appointment is tomorrow,” or something. Then having them confirm the appointment, cancel, or reschedule the appointment using messaging rather than having to phone.

That’s pretty interesting. Half of the economy is based on this services industry where you’re trying to book time with a train, a music instructor or something, or a B2B software company, these are all appointment setting use cases. It’s basically like email. Just as email is used to promote just about anything, chat’s just a messaging channel and there’s nothing you can’t do in chat that you were previously doing in email.

Work smart and look for ways on how to innovate and automate the business.

Okay. Those are all great use cases. Do you have an example of a really innovative company that you can point to, that maybe someone who’s listening can try to engage with that company and see what they’re doing that’s quite innovative?

I thought the LEGO Bot was pretty interesting. LEGO has a gazillion, like 15,000 different types of legos, different spaceships or whatever, castles and things like this. They have a nice bot that’s designed to help you figure out a gift for somebody, if you feel to give a LEGO gift. It asks you a couple of questions. The conversation is through chat and it says, “How old is the kid? Is it a boy or a girl? What’s his or her first personality? What’s your budget and price range?” kind of thing, and then it’ll narrow it down from 15,000 to 10 or 5 different creations.  You can even view the product and buy it. I thought that’s a really simple way of exposing this thing through messaging.

Of course, there’s re-engagement. That’s the initial experience, but you can do all those re-engagement tactics like, “Hey, there’s a new creation in this family of products that you were perusing,” or reminding them to finish the purchase, etc. Tons of interesting things going on there, that’s for e-commerce example.

Assuming that you eat your own dog food, I’ve been onto your mobilemonkey.com website and I see that you do invite people to use messenger to communicate with you, but you also have pop-ups that are maybe using OptinMonster or something to fire those up, that drives people to free online training like an on-demand webinar. How do you integrate those sorts of experiences together? I’m curious.

Most of the sign-ups on MobileMonkey, all the marketing calls-to-action and stuff, are actually chatbots. When it says, “Sign up for today’s webinar,” it won’t take you to a desktop or mobile landing page where you fill out some form. It’ll push you into messenger. Then, the messenger session will just ask you, “Can you confirm your name, email, and phone number?” because the thing about messenger is we already know your phone number, name, and email. We just need your permission to release it so that we can pre-populate that information and have people click on buttons to confirm the information on file. Then, using an integration with the webinar provider like Zoom, GoToWebinar, or whatever, upon completing that question in chat, we will then fire off that data to the webinar providing platform and get them registered for the webinar and then send them the link to join the invitation. Did you have a chance to look at that by any chance?

I didn’t. I didn’t try to sign up for your webinar but I did get the pop-up to invite me to it. I thought that was an interesting pop-up that you had there with a cute little magical unicorn.

That link also exists on the homepage where it says, “Sign up for a webinar.”  If you were to click on that, the registration will be taken through chat, not as a web page. There’s a lot of advantages to doing it through messaging rather than a web page. The data will be correct as opposed to goofy names, emails, and phone numbers. We can pre-populate this information with the information from the user’s Facebook page.

You can send reminders. The day before or 10 minutes before the event, we can send a reminder to remind them that the thing is starting in 24 hours or 10 minutes. We can send out the links to the recordings via messenger. These are things that you do in regular email marketing as well. It’s just that the open rates are very bad. You’re 5–10 times more likely to see these notifications if sent through messaging rather than email.

Right, because that’s one of the big problems with doing live webinars or even on-demand webinars where people don’t show up. They sign up but they don’t show up and it’s gotten from a 50% show-up rate to under 30%. It’s crazy.

I think it’s actually worse than that. Attendee rates for webinars are probably more like, on average, closer to 10%. It depends but the longer the registration is, if I tell you to sign up for something a month from now, less than 10% of those people are going to show up. It just depends on various factors, but in general, however crappy your webinars are doing, if you do it via chat, I guarantee the attendee will be 2–3 times higher. That’s my experience; just through the mechanism of the reminders.

Conversion works five times better in chat because of people more likely to see it than email. Click To Tweet

That’s great. I’m going to have to try that out. My team does this all for me but they have set up Twilio and they send out reminder text messages to the registrants, make sure they show up because email reminders are insufficient. Of course, we send out email reminders, too. Our platform for emailing is Infusionsoft. It’s painful and it still doesn’t get very high show-up rates, even when we hound them like that.

Yes. It’s because your emails aren’t being seen. It’s like they’re stuck in limbo like Google Promotions tab or Outlook Clutter box. Even if you’re email list is 100% white hat, opt-in, deblocked in, whatever, if 10 people complain, then as soon as that happens you’re in the doghouse and there’s no way out of it. You’re stuck forever.

Yes. It makes me wonder if people specifically sign up for their competitor’s email newsletters so that they can turn them into spam.

And click the spam button and all of that stuff. It’s pretty diabolical.

That’s pretty evil. I’ve heard of some crazy, negative SEO techniques. Not only spamming with low-quality link building your competitor which is evil and totally unethical but also go after legitimate links that your competitor has built and then pretend to be the competitor, demanding that the link gets removed because the site is considered low-quality by Google. That apparently works. These negative SEO attacks seem to be working where they’re getting quality links removed from their competitor’s link profile. Crazy.

I don’t know. It’s basically manual link outreach and that’s so annoying. It’s so hard.

Although, there are tools that help scale that like Pitchbox. I had the co-founder and CEO of Pitchbox, Michael Geneles, on this podcast talking about scaling link outreach. It’s crazy how you can automate pretty much anything these days as long as you’re still completely white hat which is the kind of clientele that Pitchbox has. They don’t want spammy companies become users. It’s amazing what you can do once you’ve automated a lot of what can be automated and you’ve set up wizards and things that make it super easy to scale the rest of it and the workflow that manages all that and shows where things are at like salesforce.com, but instead of customer relationship management for outreach, for links, and PR. Pretty cool.

Anyway, enough about the seedier side of negative SEO and people getting their competitors in trouble with email providers and so forth. Let’s talk about the AI side of this because that’s really exciting and it’s a great buzzword. In fact, I remember hearing the first time, maybe a decade ago. I’m big into going to futurism events, conferences, and things like Abundance 360. I’ve been to foresight.org’s Senior Associate’s Gatherings and I heard AI is not going to mean artificial intelligence. Eventually, it will mean autonomous intelligence because where is the line where you decide that this “sentience” is artificial versus not? It’d just be autonomous. Pretty wild. Where do you see AI? And what is the digital marketer’s toolkit, and not just in your platform but other tools? Where do you see AI currently and where do you see it heading?

Sure. That all sounds cool and I hope that happens soon. I think I’m a little more realistic about the capabilities of the stuff right now. Basically, we think a lot of the times AI is being used, it feels like a clunky search engine like Alexa, Google Home, and Siri. These are interesting but still pretty frustrating to use if you have them. I actually think there’s a lot of challenges and I hope they get better in terms of how they work with MobileMonkey and chatbots.

We actually have a natural language processing feature and not all bot companies have this capability where it’ll, using natural language processing, infer the intent of what the user was asking for. We’ll try to serve the correct answer based on previous answers and content that you’ve provided to users. I think that’s a real time saver because you want to be able to scale messaging.

The neat thing about a website is that people can read about your products and services without you having to talk to them manually. If you’re doing chat and you have to manually respond to every single inquiry, that’s pretty terrible. That’s totally unscalable. I think the technology, as I see it, in terms of a way for this particular use case of chat and chatbots, it’s trying to automate the frequently asked questions where it’s long tail.

There’s going to be commonly asked questions that come in many different variations. These absolutely need to be automatically answered. What are your hours? What’s the refund policy? Directions. These are questions, regardless of how the user is asking the question, we need to be able to answer them. Understand the question and answer those and I think that we’re already there.

The problem is that there’s this long tail of very unusual questions where we don’t have good answers for. That’s still 20% of the questions we’re getting. That’s the key, getting something from 80% to 90% is enormously difficult, from 90% to 99%. Just like how people complain about Google search quality today even though we used to think that it was such a miracle 20 years ago. Perfecting these things is very difficult. I don’t know if this is reality or not or if we’re talking 5 or 10 years.

I’m not actually building that technology. I’m leveraging technology. You build using the components that are available to you as a builder. That’s where the technology is today. It’s still a little clunky. That’s the reality. Any thoughts?

Chatbots, when set up with the right intention and great content, can cut through the chase with prospects even without a real person involved.

It is clunky now. It’s not really AI. It’s just machine learning which is the very beginnings of full-blown AI. You need usually large data sets for machine learning to do its job well. Then it learns from that big data set but will be evolving into other types of AI, deep learning types of AI over time, and as I said, eventual full-blown autonomous intelligence. What sort of road map are you looking at to incorporate more and more machine learning and then full-blown AI into your platform? I’m curious.

It’s a very important strategic differentiator. It speaks to a customer’s user experience with your chat functionality. We’ve already invested more than our competitors have in this space. I think it continues to be a top-three priority for near and long-term investment in the platform. I would put it up there with needing to support other channels, advancing the AI. These are very important strategic imperatives.

When you say you’re investing more than the competitors in AI, how do you gauge that? How are you measuring that? Is it staff, programmers that you’re employing to work on an AI? Is it a percentage of revenue? Is it just a total dollar spent on AI?

In ManyChat, for example, if you type in the word “yes” or something like this, unless the user exactly uses the phrase that is required, the magic password, it will do no natural language processing. If you say, “Yeah,” or if you say, “Hit me,” or something like that, it’ll only work if you put in the exact text that it is expecting to receive. When I say we’re investing more in this area, what I mean by this is the technology doesn’t exist in these other platforms.

It’s like Google trying to figure out what people were searching for 10 years ago. Before Hummingbird, before Rankbrain, it was a lot of pattern matching, exact words, stemming, and all that stuff instead of using machine learning to understand the intent of the query.

But I’m saying some of these products don’t even use stemming. It’s an exact match only.

Okay. We won’t bash the competitors too much here. I get your point. Where do you see the other non-messaging platforms evolving in terms of AI? Not many of them say that they’re using AI. MarketMuse is one that says they’re very heavily reliant on AI for their technology but I can’t think of a whole lot of tools, at least in the SEO space, that are touting how valuable the AI components are in their toolset.

You know, every marketing software application that requires you to set up roles or manually do things could be reimplemented as an AI. Rank checking, this is a ridiculous manual thing where you’re supposed to pick certain keywords and see how they’re doing. An AI version of a rank checker would just give you a report about what you need to know like, “Hey, we noticed there’s a really great opportunity here that you’re missing out on. You had this really great thing but you got killed here.” A lot of these marketing tools are literally these tools where you, as a human, need to process. You have to know what questions to ask, then you have to analyze the data, then you have to process some kind of hypothesis, some kind of actionable insights from that analysis.

What I’m saying, that’s the case in a lot of marketing software like pay-per-click advertising tools, SEO tools, CRMs where your salespeople are combing through leads to find the gem. All of this could be rewritten using AI to really figure out what is interesting, what to focus on, what should be done, and who to call, etc.

For example, I’m thinking SEMrush, they have a great feature where you can see all the featured snippets of a competitor that you would name. You’d put in the domain name of that competitor and then you change in the filter settings to select only featured snippet keywords from the list of organic keywords that they’re ranking for. Boom, there you have it, every single featured snippet that they have, but that doesn’t tell you which ones are the ones that you want to necessarily go after first because there are some things in there that are not valuable to you or maybe even to the competitor who just happened to have a featured snippet for it. You’re saying an AI could give a much better data set to the company doing the competitive research.

That’s kind of right but you’re suggesting that it could improve picking of featured snippets. I’m saying it could even work at a higher level. It might not just be featured snippets. It could just like, “This is a really great link opportunity,” or “This is a lost link that you need to really get back.” It could analyze every possible thing that you could ever do for SEO using the data set that it has. It could surface three ideas that are the most leveraged and most likely to come to flourish.

Anything, where there’s a human workflow or a marketer, needs to manually execute queries and synthesize data and formulate a hypothesis, all of this should be automated away. I’m surprised. SEMrush is an amazing company. I’m amazed that there are that many people who know what to do with that data. The reality is people are not rocket scientists and to be able to accurately analyze and process this data is very difficult and that whole step in the equation should be automated away.

Teasing out the most valuable actionable insights out of everything they could possibly be doing is something that would be great for an AI to handle. I agree with you. One of the biggest problems, at least that I see being heavily in the SEO space, working with clients and talking to prospects and so forth, is the focus on doing everything that is a “best practice” even if it’s not likely to move the needle. So much emphasis and time put on rewriting your meta descriptions when there’s no rankings benefit to doing so. It’s only to improve the click-through rate of the user reading that snippet and your search listing saying, “Oh, that sounds really good,” and then proving the click-through because of that.

What really moves the needle is moving up the rankings and meta descriptions don’t do that. Why are we spending so much time on all of these different things just to tick a box that says, “Yes, I did that best practice. I did that one. too. I did that one, too. And that one, too.” It’s being activity-focused instead of outcome-focused.

I agree with the sentiment. I wish you hadn’t used that example.

Okay. Why is that?

Editing your meta description absolutely helps your rankings because higher CTRs most certainly improves your read. You’re higher than the expected click-through rate for a given position will most certainly improve your rankings.

Okay. So CTR will improve your rankings, yes. There is no direct rankings benefit to a meta description. It’s not used as a ranking signal so it’s an indirect benefit. There are so many other things that you could do. Optimizing your title tags, for example, that’ll move the needle massively more because that’s the first thing that people look at in the search listing. It’s also heavily weighted in the ranking over them as far as on-page signals. If I had to choose between my implementation folks working on optimizing 5000 title tags across 5000 pages, optimizing the meta descriptions, or doing just a little bit of both, I would everyday, just choose the title tags over every other option.

I would choose to increase the click-through rate.

Every marketing software application that requires you to set up roles or manually do things could be reimplemented as AI. Click To Tweet

Okay. How would you do that? Through the meta description and leave the title tag alone?

Whatever is needed. Sometimes the title is great and you already have a great title. The problem with that situation is the better your title is, the harder it is to find a better one. There’s going to be situations where your best bet is to work on the meta description. Other times, if you have a crap headline then obviously, you should focus on the headline. Just like how you said you want to focus on the outcomes, I would also want to be focused on the outcomes which are where the goalpost is as CTR. You’re calling it indirect factor but I think it’s a direct factor because it directly controls ranking.

That’s how all these machine learning and AI systems work. There needs to be some feedback loop from the users for the algorithms to understand if they made the correct guess or not. For all of these platforms, SEO, AdWords, the stuff on Facebook Newsfeed’s algorithm, these are all click-through signals that either validate or invalidate the hypothesis of the algorithm that they surfacing. As a piece of content, it didn’t do well so we’re going to diminish in the rankings or the opposite where it does really well so we should probably make it the featured snippet or give it a better ranking.

I understand what you’re saying. I still disagree with you because if it’s up to Joe the search marketer whether to dedicate X amount of resources to rewriting title tags or rewriting meta descriptions, and we’re talking about a pretty big website, almost never do I see perfect title tags. That’s something that will directly increase the click-through rates. It will directly improve the rankings because that is a ranking signal directly whereas meta description is only indirect through CTR.

It’s the first thing that people see when they look at a search listing. The second thing about that would be the URL. The third thing after that would be the snippet underneath. That’s just my position and I understand where you’re coming from, but this is how I advise my clients, to focus on the things that are going to move the needle the most.

The thing that would move the needle most is brand signals. SEMrush had a study a couple of years ago that was widely panned. I think they said something like brand signals is the biggest SEO ranking factor and a bunch of people, a bunch of cross should be SEOs jumped on that one, kind of debunked it.

That’s my position. Everything matters, 200 signals and they all rank. I think what you’re saying, title tag, that’s on-page SEO and impacts the click-through rate. I think there’s another layer of signals around brand and user signals. That’s what I would be focusing on.

I’m agreeing with you that I think brand signals, user signals, the usability of the site, and so forth, that all matters for sure. Where do you stand in terms of EAT?

Can you remind me? Just so we’re clear.

EAT stands for Expertise, Authoritativeness, and Trustworthiness. It’s an acronym that Google uses. They use it, for example, in their quality rate or guidelines. I’ve had conversations with Googlers talking about how is it used, particularly, and specifically being used in YMYL websites—Your Money or Your Life—that’s another acronym that Google uses internally, especially in their quality rate or guidelines. If you have a financial advice website, you have a medical advice website, holistic alternative health website, you’re absolutely being held to a higher standard, and EAT—Expertise, Authoritativeness, and Trustworthiness—is something that’s incorporated into the algorithm to keep all the shotty charletons buried in the search results.

I think we saw that problem with fake news in the search results like Obama is an alien or whatever. These things would actually show up in Google search results because the algorithms, these engagement and brand signals went awry. Believe it or not, there’s a huge audience of people who would want to read that, that Obama is an alien because it confirms a bias that they have or something like these. It gets a very high click-through rate and therefore, ranks very highly.

That is clearly not the intended outcome. It’s more like a bug. Definitely, over the last two years, since 2016, what we’ve noticed is that they’ve gone the other direction and waiting the authority first. The way I look at it is flawless. It’s like in sports, there’s the regular season and then there’s the playoffs.

The regular season in search is like if there are 1 million results, how do you get to the first page? Going from 1 million results down to 10 results, that’s getting into the playoffs. What we’re seeing is that’s largely based on the link graph. The link graph is fairly a strong indicator of your ranking until you get into competitive rankings. When it gets to the playoffs, the ranking of the final 10 or 20, that’s where a different set of roles kick in. You check the box on your authority because you need a certain authority to get this far into the playoffs.

The final few results are then reordered based on user engagement and brand signals. That’s the model, how we see it working. I don’t polish everything that we do but that’s our understanding of how EAT works in combination with brand and user experience signals. It’s almost like a two-phased process.

Engage and re-engage to keep the communication line always open. This is the first step to great customer service.

When you say competitive keyword, are you talking about the ultra-competitive keywords like payday loans and viagra, that sort of thing?

Not at all. I’m saying, for any keyword we are ranking, there are typically millions of pages. It’s like 1 of 10 out of 3 million. Typically, what we’re finding is that just to get into the top 10 or 20, I’m just repeating myself here, the link graph is generally a decent predictor of your ranking.

When you say link graph, are you referring to the authority metrics or are you talking about trust metrics? Because those are different from each other. They’re both coming from the link graph.

A blended quantity and quality of links tends to fit the pattern of the search rankings excluding the top results. Where the correlation dies is in the top 10 or 20 results. Then it’s more to do with user engagement and brand signals. This is just what we found in our research. We think that both are important.

This was researched at WordStream or at MobileMonkey?

I do a lot of SEO research out of personal curiosity, but it was easier to do at WordStream because we had literally, tens of thousands of Google Analytics accounts and stuff like this. Basically, Google does this a lot. Can I give you an analogy? In paid ads, there was this time where there were ads on the right side and then there were ads above the folds. Do you remember that?

There were two things that had to happen in order for your ads to be qualified to show above the fold. One was that the ad quality had to be high enough and then you also had to pay more. Even if you were paying a lot of money, if you didn’t have ad quality score ranking and authoritative score, it still wouldn’t put you above the fold. You need to have both the high engagement but also the authority score in order to get these coveted spots. It’s pretty typical across different Google algorithms.

You’re generalizing beyond Google Ads to SEO as well.

Conceptually, there are similarities in that AdWords has always been a machine learning platform. There is no link graph. It’s always been AI. From day one, there was no link graph to order which ads should we show, in what order, and how much to charge. It was 100% based on click-through rate, end-user engagement, and landing page quality. Obviously, innovation isn’t done in a vacuum. As they were trying to incorporate some of these AI ideas into organic search, it made sense to look at what worked in their other channels if that makes sense.

Okay. I’m just noticing the time here. We only have 10 minutes left. I really wanted to talk more about what happened with WordStream. You had a successful exit and I’d love to get more details on that. What did you build over there at WordStream that was so valuable that it was worth $150 million to Genet? Could we jump to that topic?

Sure. WordStream was my first company. I started it in my 20s, literally from nothing, working out of a narrow bread which is like a bakery because of free WiFi and all that stuff. Over the last decade, it grew to be over 300 employees, millions of users, managing over $1 billion of ad spend for millions of users worldwide. If you think about Google Ads’s revenue, it’s shy of $100 million or something like that. The company was managing roughly 1% of Google’s total worldwide advertising revenue, so that was interesting.

I sold the business in July of last year and that was interesting. I actually had most of the common stocks. I had some investors but we didn’t raise much money, we only raised $20 million or so. I think it was a really valuable thing too, not only the business but also the learnings. I’m just trying to challenge myself personally to see if it’s possible to build even better things quicker. Any questions about that?

Messaging is the new marketing channel because it's where people engage and read the content. Click To Tweet

You’re the CTO there at WordStream. You were the architect of all the technology then from the beginning?

I’m the founder of the company. I built that, employee number one. I wrote the software initially.

Was there a point where you hired a CEO and then you stepped down to be the CTO to let the CEO grow the company to a place where it could sell for $150 million? How did that go?

Typically, east coast VC likes to pair domain expertise with a business operator so we did that playbook. I was focusing, let’s say on product and marketing and my business partners were focusing more on business operations like sales, finance, the business of running the business, etc. I think that has some pros and cons. If you’re a Silicon Valley company, I think they would more likely want to empower the founder to build an operations team around that original vision, whereas in east coast companies, it’s very typical to bring in a higher gun CEO. It was a great partnership.

This was a difficult space. If you look around, there aren’t a lot of successful marketing advertising platforms out there. When I started this back in 2009, around then there were dozens of these pay-per-click advertising platforms, but they’re all dead or sold for nothing. We’re very proud to have not just survived, but actually produced a strong return for my investors, all employees, and myself as a founder.

Very cool. When did you bring on the CEO? That was somebody that was brought in by the VC, the investors, like an entrepreneur and resident sort of thing?

That’s kind of how it works. You’ve got to understand, I was in my 20s when I started this thing. You don’t really know what you don’t know. As a condition of the first institutional money that I brought in, the first $4 million, we brought in a CEO. He was a nice guy, but there were issues.

It wasn’t a fit.

Two years later, we brought on a new CEO to take over. In fact, today there’s another CEO there, a third one. The point is that a business is bigger than any one individual, obviously. I’m not sure what the point of this is.

First of all, I can relate to having a bootstrapped business. I started when I was 24 with my business, my previous agency, Netconcepts. Then I brought on a CEO at one point and that was helpful, then brought in a different interim CEO to help facilitate the sale of the business because I was ready to exit. That happened in early 2010. The exit, a year-and-a-half or two years earlier, the sooner we brought on that interim CEO. It worked out.

I can see how if you focus on your strongest gifts, maybe it’s technology, product development, or sales and marketing, whatever it is. It makes sense sometimes to bring in somebody who can grow your business to the next stage because it’s pretty rare that somebody who is the business founder, who can be good at going from $0 to $1 million a year, may not be very good at going from $1 million to $10 million and probably not very good at going from $10 million to $100 million.

Oh, for sure. That’s definitely the case. I wish people were more okay with that because if you grew a business from $0 to $20 million or $10 million and then they terminate you because they don’t think you’re the right person to go from $10 million to $50 million or something, that shouldn’t be such a bad thing because it just means that you have a certain sweet spot. I think on the west coast, they understand that better than out here on the east coast. I think on the east coast, people want to cling onto their thing forever. Their positions’ way past their expiry date or whatever.

In general, it sounds like you might have had a similar experience. I also brought on a CEO to focus on the things that I was interested in like marketing, product development, and that kind of stuff, engineering.

It also gives you room to breathe and not get burned out. I took a six-month sabbatical years and years ago but that was so valuable. It was a lifesaver for me.

This time around, I’m actually the CEO of MobileMonkey, so I’m running all the departments and everything. It’s a different experience to learn from your experiences and try a new shell initiative. Isn’t that how life works? You want to try something new instead of doing the same thing over and over?

For sure and you’re different now. Having gone through that first successful exit, you know so much more, you got so many more skills than you did when you started, and you get to leverage that. You get to work smarter instead of harder. Pretty much the only thing that’s hustle, seems to be the word of the day or at least since Gary V popularized the concept that you just, “I got to hustle in order to make it.” These 20-somethings are working themselves to death, working harder and not as smart. I think that’s sad.

I believe in the power of these weird loopholes. Like what you’re saying, working smarter, from time-to-time you discover these weird loopholes. You’ve heard of Wayfair, right?


That whole empire is a $6 billion e-commerce empire or $10 billion e-commerce empire. That was just based on the discovery of the exact match domain thing. Wayfair used to be 5000 different websites like buycouchesandbuysofas.com. This is what I’m saying. They found an interesting loophole and that allows them to go to market. Later, when that loophole closed, they consolidated under one name. It didn’t matter. By then, they already had the brand signals and the customer base to be the entity that they are today.

I can totally relate to this because I found my own weird loophole when I had Netconcepts. I was able to figure out how to use reverse proxy technology to do pay-for-performance SEO. I was able to get all those SEO with their proxy on these big client websites like APOS and Nordstrom and so forth, at their main subdomain, at www, not some weird other subdomain or even worse, on a completely different domain. They didn’t have to do any major invasive surgery to their e-commerce platform. All they had to do was install the proxy-based rewrite rolls and boom, they were done.

You were talking about working harder, hustling, or working smarter. I believe in the unicorn, these weird leveraged loopholes. I guess it’s what you’re saying about working smarter.

But I like your articulation of it, too. I like the unicorn. People didn’t even know to ask for it like, “Can I figure out a way to not have to implement SEO on my site because I’m running some convoluted e-commerce platform like WebSphere Commerce and it takes forever? Most of the time, it’s completely impossible to implement what I want because of the constraints to the platform. Is there a way around that?”

They never even thought that there could be a way around it. They’d hire us for consulting and then we would offer them the option of using GravityStream and like, “Wow. I wish I would have known about this years ago.” We were making more revenue off of the platform than we were off of our consulting even though we were known as a consulting agency in SEO. That was the Trojan horse.

Right. They’re usually unusual stuff that’s not conventional wisdom. It’s usually not hard to implement. They’re always obvious in hindsight and they’re very leveraged. It’s not a 10% increase, it’s a 10X increase.

The key things of MobileMonkey are more features, easier to use, and better pricing. Click To Tweet

Tell me, what was your unicorn at WordStream? What was the weird loophole that you were able to take advantage of?

If you go to WordStream, there’s this thing called an AdWords Grader. For the first three years of my company, we were asking people to try out our software and it was so dumb because 2% of people would buy it. Then the big AHA was like, “Well, maybe people are interested in trying out a pay-per-click advertising platform. Maybe what they just want to know is how they’re doing in their AdWords account.” I started asking people, “What was it that you were trying to do when you signed up for this trial?” Their answers were like, “I just want to know how I’m doing in AdWords.”

The lightbulb went off was like, “Then, I should build an AdWords grading tool rather than a pay-per-click advertising platform.” That grading tool, instead of spending 5 hours to learn a new platform, in 10 seconds it downloads your data from your AdWords account and gives you a report card of how you’re doing. It gives you a competitive analysis against other competing people in your industry with similar ad spends and gives you all the tips on how to fix things.

Those are the tools that were in the platform, except we just inverted it, exposed it in the form of an audit where the prescriptive actions were given as opposed to you having to use a toolbox to find the prescriptive actions manually all by yourself. That one little thing changed the conversation rates from 2% to closer to 40%. It made it from a low, single-digit company to the plus $55 million business it is today.

Awesome. That’s a great way to end this episode. Congrats, of course, on the successful exit and on your new venture. It sounds really exciting. I’m going to sign up for your platform. I’m going to give it a go. I’ve tried ManyChat and I’m very intrigued by the capabilities that you guys offer that ManyChat doesn’t. I’m going to give you guys a go and see how I do. If folks wanted to learn more about MobileMonkey, if they wanted to sign up, if they wanted to check out some of your wisdom online, where should they go?

It’s like twitter/larrykim. I spend a lot of time there still even though it’s just Twitter. That’s a good place to start or the MobileMonkey blog is where I’m doing all my publishing these days.

Okay. So the MobileMonkey blog, of course, mobilemonkey.com to sign up for a free account or a paid account. Any books in the works or anything like that?

No. I have two kids and a startup is just murderous.

In a fun way, right?

I meant it’s such a joy.

Awesome. Well, thank you so much, Larry, for joining the Marketing Speak podcast and thank you, listener, for taking the time to listen to this episode. I hope you’ve learned something that you can apply and get some great ROI and value for that time that you’ve invested. This is your host, Stephan Spencer. We’ll catch you in the next episode of Marketing Speak.

Thanks, guys.

Important Links

Your Checklist of Actions to Take

☑ Work smarter. Determine how I can automate my business processes as automation will save me time and money.
☑ Use chatbots to make communication with my prospects and customers more efficient and effective. 
☑ Deliver great content and customer service. My entire automation campaign should reflect my whole business’s mission and vision.
☑ Set up a messaging campaign for abandoned carts on my eCommerce sites so that whenever someone doesn’t fulfill a purchase, there’s a reminder about the product they were interested in. 
☑ Upsell if necessary, and always make sure you are adding value by ensuring it’s related to the customer’s recent purchases. 
☑ Share my calendar with important contacts so they can easily schedule a meeting with me if they wish. 
☑ Engage and re-engage with prospects and customers to keep the communication going. Always keep a line of communication open to my subscribers.
☑ Use Facebook Messenger for webinar reminders and sign-ups. People are more likely to respond to an IM than an email. 
☑ Use chatbots to automate FAQs so that queries can get answered in an instant without having an actual person involved. This will also save prospects and customers from having to read through the FAQs on my website.
☑ Check out MobileMonkey for my Facebook Messenger marketing needs.

About Larry Kim

Larry Kim is the CEO of MobileMonkey, a leading Facebook Messenger chat-marketing platform. Also the founder of WordStream, a PPC Marketing Software company employing over 300 people and managing over $1 Billion in ad spend for tens of thousands of customers, which was acquired by Gannett last year for $150 Million.



Leave a Reply

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