Work Minus Fear of AI With Jim Sterne

03 Apr 2019   |   Technology

Jim sterne

Work Minus Fear of AI With Jim Sterne

03 Apr 2019   |   Technology

Jim sterne

Today, our guest Jim Sterne. He’s the producer of the Marketing Analytics Summit, and this is Work Minus Fear of AI. Hi Jim, how are you?

I’m great. How are you doing?

Very good. Jim, you do a lot of things and you’ve been doing them for a long time. So, why don’t you give us a little bit more background about who you are?

Was in sales and marketing to make a living, and then when I tripped over the internet in 1993, I got very excited. Started public speaking and writing books, put on a conference which is now the Marketing Analytics Summit, and the audience created the Digital Analytics Association. So, I’m co-founder of that. And, yeah. Consultant, public speaker, and just trying to keep up with what’s going on.

And you’ve written a ton of books too, right?

Well, twelve. So, yeah. It’s a problem.

Nice. So, you’ve been in this game for a long time. You talked about 1993, that’s going way back. So, tell us about the waves you’ve seen in the digitization of marketing, how they come and gone, and how have you managed to stay ahead of the curve?

Just consuming lots of information, asking people lots of questions. I put on a seminar series in ’94. Webmaster from Sun Microsystems, webmaster from IBM, head of an internet access company, guy who wrote Navigating the Internet, and we introduced the internet to people who never seen it before. I remember, when we finally got to Silicon Valley that year, show of hands how many people in the audience have an email address. It was about half. How many people had seen the World Wide Web? Like, ten hands went up. So, the conference was mostly showing them websites and explaining how this will change everything, and then guessing what might happen, and figuring out if that’s true or not. So, at the end of every book that I write is a chapter called What the Future Will Look Like. And if I go back to my 1995 World Wide Web marketing book, yeah, I missed a bunch of stuff but, on the other hand, I’m still waiting for a few things. You know, where’s my flying car?

Yeah. Have you reached the point in your career when you’ve seen so many changes that you just kind of like you get tired of you feel like, okay, I’ve done my part, I’ve done enough, let somebody else take it? Are you still just as excited as you were at the beginning?

I am absolutely excited because I learn new stuff every day. The joy of being a child is everything is new. And when you become a teenager, the only thing that brings that level of excitement to you are sex, drugs, and rock and roll. So, now that I’m in my 60s and I’m still excited because I learn new stuff every day, I’m hooked. I’m not giving that up.

It’s awesome. Okay. Well, let’s talk about marketing because, obviously, internet technology is lots of different things. We’re talking a little bit even closer about AI later in the show. But, why is it the marketing seems to be on the edge of most these technologies that they seem to adapt it faster than other departments?

Well, that’s a relatively new phenomenon. Marketing was, you know, printing catalogs and mailing out junk and putting up billboards and putting ads on TV. When the World Wide Web came along and email marketing came along and search marketing came along, that’s when marketing said, oh, wait, there’s technology. I mean, we had database marketing and that was exciting. But then, now we’ve got interactive, live, pull marketing rather than push marketing. And that’s when marketing got on the technology bandwagon and has been pushing it forward. And I think it was about ten years ago the articles came out saying that the marketing department was outspending the IT department. And now, you can’t do anything in marketing without technology.

Yeah. It’s amazing to think somebody coming in new to marketing, it seems like almost a fully digital job, like there’s no way to get around that.

It is. I cut my teeth on command line IBM and basic foreign data general computers. And then personal computers came along, and then graphical user interfaces. And I look at then to now in the learning curve. If you’re in your 20s and you think you’re going to master marketing, there’s a lot to get to know.

Well, let’s further confuse it and further push it out in the future by talking about our topic of automated intelligence and artificial intelligence, looking into the future about how this can interact with marketing. So, give us a brief primer about where you see these convergence is happening.

Well, first of all, we need the definitions. So, artificial intelligence is an umbrella term that covers natural language processing, computer vision, self-driving cars and robots. And the thing that marketers need to understand the most is machine learning. So, natural language processing computer vision are allowing us new ways to communicate with consumers. You talk to your smart speaker device and you can order stuff, and that’s a different kind of marketing. But, machine learning is a different kind of programming and marketers need to understand the difference between just writing a piece of software or creating an app versus creating a system that can we get smarter. And, on the one hand it’s an amazing step forward, on the other hand it is just statistics, deep statistics all the way down, that has a way of changing its mind. So, you send out an email and it looks at the results and it says, well, next time you should do this and you do that, and it looks at the results and, says, well, the next time, you should do this – ad infinitum. So, it is a different kind of software. That’s part one.

Part two is this idea that it is going to be intelligent is completely wrong, I’m afraid. It’s really good at doing very specific things. Put up a headline that people will click on and it will discover the best headline to put up for the right audience at the right time. But, it can’t do anything else. And it can only do it amazingly for a while and then the model will eventually fall apart because times change and competition changes and the competition is using different kinds of models. So, it’s an arms race. But, never are we going to see all of these systems coming together to create something that looks like human intelligence. It’s not in the cards.

And this is kind of like what you talked about in terms of weak AI versus general AI, right?

Well, weak AI is the professional term as for data scientists. I prefer narrow AI or functional AI which is solving a specific problem.

Okay. So, marketers have all sorts of problems. So, what are some of the initial ways that AI is going to help marketers solve some of those problems and which ones are going to be less solvable?

The solvable ones are the things that are a little too complex for a spreadsheet and require human interaction. And right now, you get an intern to do it. You know, I hire somebody to sit down and go through all of my contacts, and looks through Linkedin and find out how they’re connected to each other, and create a list of who is the most likely person that I should call tomorrow as a sales person, or which people are most likely to respond to a direct mail action or an email or an ad on Facebook. And that takes a little bit of knowledge. It takes, not knowledge. It takes common sense. Those things that are highly repetitive, that are specific, you can’t really make them rules-based. That’s what a straight programming piece of software does. It’s a whole bunch of rules because the rules are a little mushy and they’re too fragile, and a model that you build gets complicated and falls apart.

But, machine learning can look at all the data and show you the patterns. It can show you the anomalies. It can show you the people who are most likely to respond and it gets better at that over time. So, the repetitive, annoying, tedious tasks that you actually need a human for today are the first things that machine learning can do. Lead scoring, finding lookalikes. And here are my highest customer lifetime value. Customers go out in the world and find people who look like that so that we know who to advertise to, and then segment those people by the type of message we should send. And then here’s a database of pictures and a bunch of headlines to try and see which headlines and images work best with which segments, which a human could do in months and months and months and months or the machine can do in a matter of hours.

I think, you said something about, you know, should hire an intern to do this. I think, as we go forward into this new world, this question of who should I have to do this is going to become one of the essential questions for marketing and for every other department too. Should I hire a human? Do I need to get a smart person to be able to do this or do you need to get a smart system that’s going to be able to do this?

And that’s always been the case with technology. Should I hire another bookkeeper or should I give them QuickBooks? Should I have secretarial pool or should I buy Word processing for my executives? Should I hire a couple of people to do lead scoring or shall I — And we’ve reached the point where it’s I don’t need to hire a data scientist to create a machine learning model. Instead, I can go out to the market place and find a startup that has done that already and then just use the all-new tool. So, this is you know the, I don’t need to fear that my job is going to be replaced because there’s a whole bunch of human things that this machine will never do. But also, I don’t need to become a data scientist tomorrow. I need a better understanding of statistics so I can talk to data scientists just like I needed to know enough about Windows or how MAC operating system works so I can find my files once I save them. But, I don’t need to be a programmer to be a marketer.

How technically literate do you feel like most marketers are today? Do you feel like we’re decent or do you feel like there’s a long way to go to be able to build up his vocabulary and understand how to talk to data scientist?

It’s a spectrum. Some are very advanced and have become data scientists themselves, some do not speak the language and don’t care too and that’s their prerogative. Data literacy is necessary now and statistical literacy. And I don’t mean you have to go back to college. But, you need to know what statistical significance means. You need to be able to understand the technology well enough so you can describe your problem properly to either a data scientist who’s going to create something or to the marketplace so you can find the best solution for any given problem. So, yeah, it’s new enough that marketers are so busy doing what they do, that they haven’t had time to turn in. I’ve got to worry about social media. I’ve got to worry about how are we going to measure the effectiveness of advertising on Fortnite? There’s just new stuff every day and you want me to go learn artificial intelligence? There are only 24 hours in a day and yet, this is exactly the same thing I heard when I was out there pounding the bricks, telling people they needed to have websites. Oh, we don’t have time. We don’t have money. Nobody’s actually going to be on the internet. It’s like, “yeah, they will.” Well, my competitors aren’t going to fill in the blank, be on the internet, use social, use machine learning. Oh, yes they are and you need to compete with them.

So, we’re talking about the fear of AI. So, what scares marketers about AI right now?

Two things. First of all, it’s a mystery. It is just a stone mystery and it is sold as being the most complex thing on earth, whereas, you know, a handful of articles, or you could read a book,  will give you enough knowledge to allow you to work with it. But the other fear that is trotted out all the time is this is going to take my job or artificial intelligence is going to eat the world. And, just not true. If you are a marketer and you are worried about your position, the thing that you should do the most and make sure you set aside time for it every day is to understand your industry and your customers and your product offering because the ability for the human mind to use reason and common sense and empathy and applying experience learned over here to the problem over there, these are uniquely human that we have integrated by cognition. I can read a science fiction book and it will give me an idea about something I might do tomorrow in marketing to sell my product. That’s something machines don’t do.

So, if you know more about your industry and your customers and your product offering, and then these new tools come along as they do in waves, you are the value, you are the human value because of your reasoning and common sense. Your ability to recognize the problem that you want solved, the question you want answered is the big deal. Machine, you have to tell the machine what problem you want to solve. You have to tell it what data to consider. And then you have to look at the results to see if they make sense, what I refer to as the smell test. Those things are uniquely human and will never go away. So, is your problem that nobody opens your emails, or nobody clicks, or nobody converts? Is your problem that your customers are not loyal? Is your problem that etc. etc. Identify which problem you want to solve. Tell the machine which data might be the most predictive so that it has things to look at because it doesn’t know about anything except what you tell it, and look at the results. So, if it says, the best way to get people to open an email is to send them a million emails every day and eventually they will open one. Well, yeah, that’s true. But, common sense says, no, thank you anyway.

This is the smell test, right?

That’s the smell test. So, the ability for humans to relate non-related information and experience is amazing. We don’t know how that works. We call this artificial intelligence and we call it neural networks because it imitates the way we think the brain works. But, we know that we don’t know how the brain works really. But, they came up with this kind of model of how it might work, then they created this programming capability that mimics that and, by golly, it works great! But, the human mind is far more amazing than we understand. Now, when you take six or eight or ten human minds and put them in the same room and have them collaborate, there is no computer that could ever approach that. The better your collaboration, the better you are with human interaction, the more knowledgeable you are about your marketplace and your products, the more valuable you are as a piece of the system that includes a bunch of technology tools. But the human, your job is not going away as long as you’re focusing on reason and knowledge and not just on, oh, I’m an expert at using a tool. I don’t need an expert at PowerPoint. I need an expert at the subject matter that’s going to use the PowerPoint to impress people.

Yeah. So, we’ve been talking about how individual marketers can use one of the tools that are underneath AI to get their job done. But, let’s talk about the enduring human parts about how a marking happens in general. Is it conceivable that a company, all of their marketing is purely system-driven, system-led, there’s there’s really just one human maybe sitting back in the back directing traffic or are there always going to be parts of marketing that need to have a very human touch to them?

If I thought about it enough, I might be able to come up with a type of company selling a specific product to a specific market place where it’s all automated. Seth Godin talks about Permission Marketing and the ultimate extent of that is when you’re unconscious in the hospital with an IV in your arm, you have given the hospital permission to put things in your body without asking you anything further. That’s the ultimate. So, is there an example in marketing where it could all be black box driven? Yeah, maybe. But, the difficulty is that you’re dealing with humans on the other end. So, we all know that when you have a problem that you can’t answer online and the app won’t help you and the chatbot won’t help you and you have to call, you do not want to press 1 for this and 2 for that and 3 for that. You just want to go right for zero. I need to talk to a human. I have the unique problems. It’s not just problem 1B-4. This is a weird situation. Can you please help me? We will always need humans. We also will always need humans for the creative side. Can I give the machine a bunch of examples and it’ll find which ones of those are the best? Sure. But, it cannot think out of that box and that’s what we need humans for, to say, you know, this issue or this product connects to people emotionally on this level. Here’s a type of message we can try.

And here’s a totally different message we can try. Let’s give it to the machine and let it figure out which one is better. But the machine, it’s not going to come up with those on its own. It’s going to come up with a bazillion things that are a waste of money to try, and that’s a bad use of the machine. Let’s let humans handle the emotion part and the machine handle the math.

Yeah. You brought back actually interesting point I was thinking about of that when we’re selling to humans, that becomes an essential part that you need to pull a human element. But, there could be a day when we’re actually selling to machine. Procurement is taken over by AI itself. And so, is AI the best one to sell to AI on the other hand too? That’s the interesting question.

And this is where I think the world is headed and it’s where the robots really do take over. With privacy becoming a major issue and we’re losing our trust in Google, Amazon, Facebook, Twitter, etc. to take care of our data for us, there is a movement where individuals own their own data and allow access. So, I’m going to allow Amazon to access information about my credit card and my shipping address and I will get value out of that. I will let my doctor have access to my medical records. I will let my accountant have access to my financial records. But, I don’t want those people looking at that other stuff. That’s very specific. I’m going to give access to these people for this purpose. Then I wrap that with my own personal agent and my own personal agent knows everything about me and that’s okay because it’s my data. It’s not stored by Microsoft. It’s mine. I control it. And my personal AI system is not built by IBM. It’s mine and I can trust it. So, it will tell me that my dishwasher is going to fail next month and that it found twelve different dishwashers that will fit my budget, fit in the space that I have available in my kitchen, and will be a style that is compatible with the rest of my house, and is on sale, and can be delivered when I’m available, and would do I like the appointments set for next Wednesday? Yes or no.

That’s going to be marketing. Now, as a marketer, you need to be able to communicate with that AI system when it says it’s time to go shopping for a dishwasher. You’re listening, and your AI system pops up and says, here’s what we’ve got. What are you looking for? Negotiate. Negotiate. Negotiate special deals. We’ll send somebody out in the right time and win the bid if you will. This is something at Harvard they call vendor relationship management which is the reverse of customer relationship management. And it’s complicated and it is going to be a sociological shift, but it is the level of convenience that’s offered is so high that people will do it. They’ll run to it because it will make their lives easier, faster, cheaper. And marketing will change yet again. So, eventually, I’m going to have to write yet another book. Lucky 13.

Yeah. Well, it’s a great segue into the feeling like, okay, what are the small things that we can do as marketers, as other people, to be ready for AI. If somebody is listening to this, they feel like, okay, I’m scared now. I’m out of touch with all these technologies. Obviously, getting your book is the first step. But, after that, what’s one small step they can take?

As with all of these technologies, try something. Just try something. Just play with it. Go find a vendor who will do a proof of concept with you. The critical piece is that you’ve got good data for the machine to eat, that you’ve got people in the company who are willing to experiment, and set aside some time and some money to experiment and learn about the tools so you understand when they’re useful and when they’re not. With all of the technology changing voice ordering.  Hey, smart voice system — I’m not going to say her name because then everybody’s systems will wake up, I need some triple A batteries. Well, today, If I’d say that out loud, I am going to get Amazon branded batteries. But, if I specifically say I want Duracell, I will get Duracell. So, marketers, guess what — branding has become even more important than ever. We need to be top of mind. Yeah. So, get to know the tools, understand that building your brand is critical, and brush up on your soft skills because it’s all about collaborating with other humans. Not just customers, but inside the company and with your agencies.

Yeah. Absolutely. Well, the book is Artificial Intelligence for Marketing, practical applications of that. Jim, where else can we go to stay in touch with you?

I am at And you can find my event coming up in June, Marketing Analytics Summit. And, of course, Amazon because it’s a dozen books.

All right. Great. Well, Jim thanks so much for being on the show. We appreciate your insights and hope to stay in touch.

My pleasure, Neil. Thanks for having me.

Subscribe to The Digital Workplace

Join the journey to a better future of work