LeadCrunch Webinar: Use AI to Find Your Best Prospects, July 15, 2021 - Register Here

How AI & ABM Are Improving B2B Marketing

Olin Hyde

Listen now on these popular platforms

Share this podcast

Highlights from this episode

Podcast Transcript

State of the Industry – B2B Demand Generation talk with Olin Hyde, CEO and Founder of LeadCrunch. Today we're going to discuss the state of the B2B Demand Generation industry, how we got to where we are, and what exciting things lie on the horizon.LeadCrunch[ai] uses artificial intelligence to drastically improve the performance of B2B demand generation campaigns through account-based "lookalike" modeling. Click the link for more information. https://leadcrunch.com/solutions/

Posted by LeadCrunch on Thursday, June 27, 2019

Hosts: Dave Green & Jonathan Greene

Guest(s): Olin Hyde

Topic: B2B Marketing 

Subtopic:  Artificial Intelligence 

Duration: 27 minutes

In the past 10 years, B2B marketing processes have multiplied but ROI is stagnant. In today’s episode on the Green and Greene Show, LeadCrunch CEO Olin Hyde joins the podcast to discuss why – and how AI can fix it.


  • Introducing Olin Hyde, Co-Founder of LeadCrunch
  • Why B2B Targeting Is So Terrible
  • What’s Not Working: Firmographics, NAICS Codes, & Intent Data 
  • The Limits of Cognitive Load and Fighting the B2B Status Quo
  • Massive Improvements: From a 40% Error Rate to 15%
  • What ABM Could Signal For the Future of B2B Marketing 
  • Nuts & Bolts: How AI Improves B2B Targeting

Podcast Transcript


[0:00:08.2] ANNOUNCER: Live from the city with the most perfect weather ever, San Diego, California, all the way to the gleaming shores of Jacksonville, Florida, it’s the Green & Greene Show. Here are your hosts, Dave Green and Jonathan Greene, goofing off instead of working, while unlocking the mysteries of demand gen. The Green & Greene Show is brought to you by LeadCrunch, which creates B2B look-alike audiences.


[0:00:33.0] DG: Jonathan, I told you we had to edit out that goofing off part.

[0:00:38.3] JG: Not going to happen. 

[0:00:42.1] OH: What do I pay you guys for?

[0:00:44.7] JG: Oh, God. Straighten up, Dave. Straighten up. The boss is on this time. We have to act like we actually know what we’re doing here.

[0:00:51.1] DG: Oh, boy.

[0:00:54.8] OH: Hope somebody does.

Introducing Olin Hyde, Co-Founder of LeadCrunch

[0:00:57.6] JG: Well, listen. It’s the Green & Greene Show, if you haven’t guessed. I’m big Greene, he’s little Green down there, at the bottom. Today, we’ve got Olin Hyde, President, Co-founder of LeadCrunch. He’s doing big things. He’s looking good doing it. Welcome on the show there, sir.

[0:01:12.2] OH: Thanks. I’m a little confused. You guys couldn’t get another guest today? How did I end up here?

[0:01:19.7] DG: We need more material, man.

[0:01:22.0] JG: These slots are highly coveted, but we figured we’d carve one out for you, since you write the checks.

[0:01:27.9] OH: My third-grade teacher told me, “Empty cans make the most noise.” It’s going to be a loud show.

[0:01:35.8] JG: Right on. What we want to do today is get a state of the industry of B2B demand generation from you, because you’re well-positioned to know these things and have a perspective on it. We’re going to just jump right into that, talking about B2B demand generation as a whole. As I look at the industry today, probably the overarching thing that jumps out at me is that there are a few things that are not great about B2B demand generation. Targeting is one of them. It’s probably the biggest one as far as we’re concerned. How did we get into a situation where B2B targeting sucks so hard? How do we find ourselves in this position?

Why B2B Targeting Is So Terrible

[0:02:17.0] OH: Well, you nailed it Jonathan. It is awful. B2B targeting is absolutely awful. There are a lot of reasons for it. First, very few marketing technology solutions have focused on targeting. Most of them focus on channels, like how to get better social engagement, how to get your e-mail to be more effective, or how to get whatever channel to work. There hasn’t been a whole lot of evolution around targeting. The few tools that do exist on targeting really are limited because they look at the world the same way as everybody else does. That’s through firmographics or intent. Does everyone know what firmographics are?

[0:02:56.4] JG: I would venture a guess that everyone does not know what firmographics are.

What’s Not Working: Firmographics, NAICS Codes, & Intent Data 

[0:02:59.1] OH: Firmographics are things like industry, headcount, revenue. You can think of these things as descriptions of the company you can look up in a database. Almost all these descriptions are fundamentally flawed. Take industry for example. What industry is Apple in? Are they a software company? A hardware company? A music publisher or retailer? They’re all of that and a lot more. What industry do you put them in if you’re a government bureaucrat coming up with NAICS or SIC codes?

When those codes were developed, Apple was only a computer company. The whole idea of having these online marketplaces didn’t even exist when those things came up. Firmographics are going to be flawed just because they were developed a long time ago. Companies are complex, and every company does more than one thing.

The other way a lot of companies do targeting is with intent. This is flawed because intent is saying, “Okay, these are the people who are searching for you on the web, so they must be interested in you.” What if they’re searching for you because they’ve already looked at your competitor? The competitor has already set the frame for how they’re going to view you. You’re already starting from second place or beyond in intent marketing. I think that’s fundamentally why targeting is bad. There hasn’t been a lot of evolution in it.

It’s also bad because it’s really hard to do. There’s no silver bullet. There’s a lot to consider. People like to think about companies or industries or products or people, and it’s all that and a lot more. On top of all of that, even if you can figure all of that out, you might not have any influence at all on a given company’s ability to make a purchase decision. I know that when my board of directors comes and tells me, “Olin, you’re going to buy X,” I get hammered for X. If you’re selling Y, it ain’t going to get bought.

I think there are a lot of reasons targeting sucks. Fundamentally, it’s because everybody’s looking at the world the same way through a distorted lens of trying to fit complex data into a very simple model.

[0:05:24.8] JG: What do you think about that, Dave?

[0:05:28.0] DG: Yeah.

[0:05:31.6] OH: You don’t have to agree with me. I think it’s more interesting if we have a debate. I would love to hear other people’s perspective on this. Don’t just agree with me. I mean, is that true?

[0:05:43.3] DG: Intent data is one of those things where I do think, to your point, everyone has seen it as some new magic bullet. Just because somebody’s reading an article, you have no idea if they’re a good fit for your company or they have any role in the purchase decision. It’s one little tiny faint signal. To hang your hat on that, I think, is maybe crazy. 

Also, I was talking to the head of our data science team the other day and we have a huge database here. I think it’s like 200 million contact names in the US or some crazy number. He went in and was analyzing all the different titles, because that’s the other method people use to target. They have titles. There were 250,000 different titles. When you put “IT Director”, guess what? You’re probably not going to get the guys you want. You’re going to get some of them.

[0:06:44.5] JG: I’m pretty sure my title on LinkedIn right now is Marketing Ninja or something like that.

[0:06:49.4] OH: Yeah, Demi-God. I think you’re part God, part human, something like that.

[0:06:55.0] JG: By the way, I hear Apple has a really nice doggie daycare as well. I’m not sure that’s how they should be segmented, but “Software company” definitely doesn’t capture everything that they do.

[0:07:05.5] OH: I walked by Dave’s office earlier. You brought your puppy in today, Dave?

[0:07:09.3] DG: I did. Hey Mish. He wants to pick up my camera.

[0:07:14.4] JG: Mish is older than half of the client success teams. I’ll throw that down to you.

Anyway, we’ve got a status quo and the status quo is that things are marginally crappy, like Dave’s first marriage. No, I’m just kidding. What are the core problems with the status quo and B2B targeting? What’s going on underneath the surface there?

The Limits of Cognitive Load and Fighting the B2B Status Quo 

[0:07:41.9] OH: Status quo almost always sucks. I don’t think anybody says, “Wow, the status quo is awesome,” particularly if you’re working at a startup. I start startups because I don’t like the status quo. The status quo of targeting is bad because you have to find out exactly why a company is a good target. That’s what targeting is all about. Why are they a good fit for you? “Why” is not something you can just look up in a database, as we were just talking about with firmographics.

Firmographics is something I love to beat up on, because we spent more than two years fixing that problem. Forty percent of all firmographic descriptions are flat-out wrong. We did a really interesting study where we had a panel of people go on websites and compare the firmographic identification of the NAICS and SIC codes. The most recent information were from take a really great database. I like DiscoverOrg. I think they’re great database. ZoomInfo is great database. There are some good databases out there.

Even the best of the best are all pulling from basically the same data sources. They’re saying the industry is wrong 40% of the time. We only get it right about 60% of the time. We came up with a way of looking at industry as a group of features. If you can think of this in math, you’d call it a vector. To what degree is Apple a software company? To what degree is Apple a retailer and everything else they do? When we looked at this, the error rate was about 15%.

Going from an error rate of 40% to 15%. Fifteen percent sounds really high, but when you combine that with other things, you can actually get to high levels of precision. The problem with the status quo is they’re thinking about things in a very structured, linear way. We live in a world that’s very complex. The human brain, the way we think, if I can geek out on this a little bit, is we all have something called bounded rationality. The human brain can only keep track of about seven things in working memory. Think of that as seven variables. You can’t think of a thousand things at one time.

The degree to which you fill those seven available slots to think about is called your cognitive load. A marketer has a lot going on in their day. For them to think about the thousands of attributes of the company, why they would be a good fit, and think about all the other things they’ve got to do in the day, there’s not room for that to happen. Your status quo does not account for this limited capacity of the human brain, the incredible complexity of the world around us, and the fact that we are surrounded by data that is factually wrong. That’s a pretty bad place to start, and that’s the status quo.

Get the ebook “How to Scale Your Demand Gen Operations “

[0:10:46.3] JG: That’s why robots are dope, because they can do all kinds of crazy calculations at one time.

[0:10:50.7] OH: Right. It’s why Google is dope. Imagine if we went back 40 years ago and said, “Someday, you are just going to a little place where you can type in a question and get magical answers to almost anything you can think about.” You’d think that’s science fiction. The good news is that we have the data science and the technology to not only change the status quo, but replace it with a better status quo, which allows people to really deeply understand why companies are good fits, why they should be targeted, and how to engage them.

[0:11:27.1] JG: Yeah, that’s crazy if you think about it from a tech perspective. I was alive when there were rotary phones and you had to go stretch that long cord across the room and under the door to your bedroom and shut it, so you could do the angsty teenage thing. I’m sure you guys pretty much did the same thing with clay tablets when you were young.

[0:11:48.7] OH: Like cuneiform. Yeah. We’re actually still fluent in cuneiform and Aramaic.

Massive Improvements: From a 40% Error Rate to 15%

[0:11:54.6] JG: Look, we’ve glossed over the point for a second, but this is important. You said you took the error rate from 40% down to 15%. That’s a lot of waste in marketing budget. If you start looking at the universe that you’re targeting with your advertising and 40% of it is BS, and you can take that 40% to 15%, I don’t know if people really understand the compounding effect of that impact on the bottom line, but your waste is being reduced by a lot—roughly 60%.

[0:12:27.4] OH: Marketing waste is the tip of the iceberg. The data is changing constantly. Twenty-five percent of Americans change jobs every year. Human knowledge doubles every four years. The data accuracy for contact information, for example, decays by about 3% to 4% per month. As you pointed out earlier, there’s no standardization for titles. There are 250,000 titles and approximately 160 million people in America with a job. That means there’s probably an average of about 160 or so people per title. That’s just insane. The world is not that complex for there to be that many different jobs, but it just shows how difficult it is to find the signal for whom to sell to.

[0:13:21.1] JG: Yeah. It’s not just human knowledge that doubles every four years, human stupidity does as well. It’s true. The terrible analysis of that knowledge and the terrible application of it by people who really don’t know what they’re doing floods that channel and every data source. It’s not just for the human brain to find connections in data. It’s also to sift through all the BS that isn’t right, or that isn’t measured. Robots are just way better suited to doing that than we are with mass applications of linear algebra and stuff like that.

[0:13:56.8] OH: The robot actually makes our jobs a lot more interesting, because we can use the human brain for what it’s really great at, creativity and building meaningful relationships and the human side of it. The inspiration for starting this company was really around how we can use artificial intelligence to make the human life better. That’s a pretty cool thing to come to work to do every day. Hopefully, we’re doing that for marketers.

[0:14:26.2] DG: I felt you really nailed it right at the outset. I’d just like to touch on this idea of “why”. “Why is somebody a good potential customer for you?” is a question that I think marketers need to spend a lot more time and think more deeply about. Part of that is trying to get the right data to even know who the heck you should be targeting. That “why” question is one that, because they’re so overwhelmed with minutia, is sometimes very difficult to just step back and ask the more important questions about why. Why did they respond, and why should I go after this target versus that target?

[0:15:06.9] OH: Yeah. Why is a really fun thing to work on. That actually allows you to have that meaningful conversation, where you’re immediately relevant. You get to the endpoint a lot faster, which, financially, is far more efficient. From an emotional standpoint, I think we all want to be relevant. People love to buy; they hate to be sold. That really gets down to the why thing for marketers to understand their customers more deeply. That can really only be done at scale with artificial intelligence.

[0:15:45.2] JG: Yeah. There are a lot of people who heard you say that and they’re rolling their eyes right now, and I’m going to tell you why, and then I’m going to give you an opportunity to cement what you just said because I think you’re right. This is not the first time that B2B marketing land has heard that a tech solution is going to change everything.

For all the tech advancement that we have—ABM, advanced segmentation, artificial intelligence, CRMs, digital retargeting, all the iterations that marketing has gone through in the B2B space for the last 20 years—we ostensibly are no better in terms of ROI than we were when we were buying single-page ads in whatever magazine. Tech is the do-all-and-be-all and the answer to all these questions. Why hasn’t ROI skyrocketed for B2B marketers in the time since the tech revolution began?

[0:16:42.8] OH: Yes, you’re right. Actually, ROI has gone down. The reality is ROI sucks for a very simple reason. Targeting technology has not improved in decades. What we’ve done is scaled channels while conversion rates drop. The noise floor is deafening, and that’s because channels make money by selling volume. Conveniently, marketers are addicted to volume. Audience volume, traffic volume, lead volume, you name a volume, if you’re selling more of it, marketers want to buy it.

This addiction comes from the horrible conversion rates. I guess it’s what’s called a feedback loop, which just keeps getting worse and worse and worse the more technology you add on to it. That’s why the ROI sucks. We can go through a specific example. One of my favorites is display ads. Holy shit. Only about half a percent seeing your display ad clicked on. That’s why the marketers call it banner blindness. I think it’s more accurate to call it banner pollution, because you’re showing ads to people who really don’t care.

The story gets worse. Most of those clicks are going to bounce because even the ones that click don’t see that you’re relevant. For sake of argument, let’s say everyone who clicks sees irrelevance. That’s not true, but of the ones who do click, only about 2% of them share their identity. You’re left with 98% having no idea who they are. Ninety-eight percent of your 0.5% are anonymous. By the time you factor that in, let’s say half of those are relevant, now you’re down to a 0.003% of your ads, or about one in 300,000, have any value to you at all. You’re buying 300,000 to get one. If you’re really good, 1% of those will close. You need to buy three million ads to get one converted customer. That’s why you do paid search, and that’s why Google’s worth a trillion dollars.

[0:19:10.2] DG: I think if we could just get it down to maybe a billion ads, or something.

[0:19:19.2] OH: These numbers are like what you’d expect from the spawning of salmon or something. 

[0:19:29.9] DG: This is one of those pieces of data that, when the aliens are looking down at us, they’re thinking, “Yeah, no real intelligent life here on Earth. Let’s keep going.”

[0:19:39.5] OH: It reminds me of mayflies, fly-fishing on rivers in upstate Michigan. There are billions of them. They’re all going to die, right?

[0:19:50.4] JG: We might as well be a bear sticking our face in the stream is basically what you’re saying.

What ABM Could Signal For the Future of B2B Marketing 

[0:19:54.8] DG: One area where I think there’s a huge opportunity, and I think it was all with the right intent, is the whole movement with ABM, Account-Based Marketing. Finally, it occurred to someone that not all the leads have equal value, after throwing over tons of sludge that the salespeople hate. Somebody said, “Hey, you know what? There’s a smaller set of accounts that actually buy a lot and maybe we should put more resources on those.”

Yet, they use the same tired approach. What’s your opinion of which companies we should go after and things like that, as opposed to using all of this modern technology to really have a more dynamic ABM list based upon a lot of different factors that would guarantee more revenue for you in the long run? I think that is a huge opportunity for marketers into the future, getting a lot more sophisticated about ABM targeting and using artificial intelligence to help with that rather than mere opinion.

[0:20:55.6] OH: Yeah, I think ABM is just the beginning of a long evolution. We’ve gone from what’s segmentation-based on firmographics, which is crazy, to segmentation-based, like let’s take a universe 10-X and just focus on 10% of it and focus our efforts where we think it’s most likely to generate return. That’s called ABM.

The evolution beyond that thinking that it’s going to convert ABM version one is the next version of ABM, where you take a more scientific approach. That’s the big why. Why is that company on your ABM list? A good reason is company fit, supply-chain fit, that our solution is relevant. Currently, most of those ABM lists are generated because marketing goes over to the sales department and says, “Hey, guys. Who do you want to target?” Salespeople are better at knowing that than almost anybody, but it’s still not science. I think there’s still room to improve on ABM.

[0:21:57.9] JG: To make it really concise and clear and encapsulated in a nutshell, ABM is a great advancement in terms of thinking about how to approach the problem, but the way it’s done is subjective. It’s a supposition-based marketing. We’re almost saying that the answer to that is going to be intelligence-based marketing, in the sense that there’s an artificial intelligence curating who those targets ought to be. What is that curation based on? How does the artificial intelligence know who the right people to target are?

[0:22:32.2] OH: Well, it’s going to be different for every company. You need to have a way of looking at every unique company in their context and understanding, “Okay, how does this data apply to this particular company?” I think the easy way to describe this is to look at what the algorithms did to baseball. Baseball’s sitting on all the statistical information and suddenly, someone has the idea of, “Hey, let’s create a model to use this data to select and recruit and develop a baseball team.”

That’s much the way ABM is going to evolve. You’ve got a lot of data on the company, but people aren’t squeezing that data to get all the value from it. That’s the way you use the AI to get the right list. 

Nuts & Bolts: How AI Improves B2B Targeting

Now there are a lot of ways to do it. I would argue that the algorithm is always going to improve and getting into exactly how the algorithm works is probably really boring. I would probably want to bring in our chief scientists to talk about it. Basically, you can think of it as if you can pick out the shape of something, that’s called features. Think of it as pattern recognition. The pattern for every company is going to be unique for every company.

If you give me a list of 50 of your best customers, Jonathan, and Dave gives me a list of 50 of his best customers, they’re not going to be exactly the same list. Certainly, your two websites are going to be unique. I can look at the similarities and differences between each of your lists of 50 and your website. Now I can start to infer linguistically how similar your customers are to you and to each other. What are the keywords? What are the ways that they are similar and different? That’s just one of many cuts. There’s the cut on linguistics. There’s the cut on supply chain. There’s the cut on people. What type of people are working for you versus working for Dave?

This process goes on. Of course, machines never get tired and their math is perfect. Eventually, they spit out a model that we can use to measure how good the model is. The way we measure that is by looking at how accurately we can recall the 50 that you gave us.

[0:24:53.5] JG: The bottom line is that basically, the problem with ABM as it sits right now, as humans, we’re trying to model what we can’t possibly necessarily understand in terms of complexity of relationships between which campaigns are turning out well, which companies we’re working with are going well. There’s such nuance to what makes them successful relationships that we can never possibly unpack it with human understanding. 

You turn the robot loose on it, then all of a sudden, it turns it into a mathematical equation and extrapolates it to a data set that we couldn’t possibly consider. Suddenly, you’ve got a really cool list of look-alike targets based on people you’re already having success with. That sounds fancy as hell.

[0:25:38.5] OH: It took a long time to build. It wasn’t easy to build, that’s for sure. I sure hope it works.

[0:25:47.0] JG: What do you think, Dave?

[0:25:50.2] DG: I think that’s all for today, folks. I think we have totally crystallized the biggest problem that demand gen marketers face today: the enormous amount of wasted advertising they do. We’re here to help fix that.

[0:26:03.8] JG: Yeah, right on. Last word, Olin. You got anything you want to say to the market at large, all the beautiful people?

[0:26:11.4] OH: Yeah, I really hope that better targeting develops better relationships. I think that’s what we’re all in this for. I’d like to see a world where people are spending less money than what it takes to buy three million ads to get one customer.

[0:26:30.4] JG: All right, play the music. We’re done. It’s been the Greene & Greene Show. 

[0:26:30.6] OH: Thanks.


[0:26:39.8] ANNOUNCER: Thank you for tuning in to the Green & Greene Show by LeadCrunch. Green & Greene think differently about B2B and are starting a movement to transform demand gen. If you have ideas for topics or would like to be a guest, send an e-mail to david.green@leadcrunch.ai. If you’d like to find more customers, visit our website to talk to one of our demand gen guides, www.leadcrunch.com.