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

TrendSpotting Episode 5: Magnus Unemyr

Episode Overview

Today, we’re going to hear from Magnus Unemyr a marketing automation and AI consultant. During this episode we will discuss AI, Machine Learning, Data Science, & Marketing Automation.

iTunes       Overcast      Spotify      Stitcher

Episode Transcript

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

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

Ben:                 This month, we’re going to define and investigate some of the overlapping new trends in marketing like Artificial Intelligence, Machine Learning, Marketing Automation and Data Science.

Ben:                 Joining us to kick off Modern Marketing Month is Magnus Unemyr who is a marketing automation consultant, speaker and the author of a series of marketing focused books including “Mastering Online Marketing and Data Driven Marketing using Artificial Intelligence”.

Ben:                 We’re excited to welcome Magnus to the TrendSpotting podcast to help us differentiate between the various trends in marketing today.

Ben:                 Here’s our interview with Magnus Unemyr, the author of Data Driven Marketing using Artificial Intelligence.

Ben:                 Magnus, welcome to the TrendSpotting podcast!

Magnus:           Thanks for having me!

Ben:                 It’s great to have you here and we’re excited to talk about some of the, maybe the most highly used buzz words. Maybe some of the newest and most effective way to drive marketing growth. A lot of ground to cover but let’s start off with you telling us a little bit about yourself, your business. What’s your area of focus and who are your customers?

Magnus:           Okay, so I have been in the international software industry for about 25 years. Most of that time in various marketing positions including the VP of Sales and Marketing. I have been co-founder and a co-owner of a company on the internet of things space and I now work as Marketing Automation Consultant, speaker and author with a specific focus on more advanced marketing automation and marketing with Artificial Intelligence.

Ben:                 So, you’ve been in the marketing field for at least two decades, at least two and a half decades and have seen the emergence of technology, you’re active in content marketing, you’ve worked in SAS business product, you have a well-rounded set of experiences. I want to talk to you a little bit about some of the trends that are popping up in marketing over the last three to five years and I mentioned them in our introduction but, we’re seeing the words Artificial Intelligence, Marketing Automation, Data Science and Machine Learning, pop up regularly.

Ben:                 You’re invested in that field. How do you describe that sort of conglomerate of terms, areas of focus, marketing functions?

Magnus:           Yes, quite fussy even to us who work in the field I would like to say.

Magnus:           To start with, Artificial Intelligence, to me at least, is more like an umbrella term. But in comparison to several specific areas including predictive analytics or machine learning so we don’t really have AI currently, at least not strong AI where machines have feelings and opinions and so on but what we have currently is weak AI whereby pre-programed certain applications become quite good at solving a specific task but they can’t really start on new things to which they haven’t been pre-programed into doing in the first place.

Ben:                 So you mentioned that in some level we don’t have true AI and it is being used as an umbrella term. What do you consider true AI? Is it marketing with feelings? Is it, you know, Sky Net and the Terminator? What is, really the, I guess not necessarily the future of AI but what is real AI and what is what people call AI? Help me understand the difference between the two.

Magnus:           So real AI in my opinion would be software solutions that among other things can learn to do new tasks that they haven’t been pre-programed to do. They may also be very intelligent and some to develop their own feelings, their own opinions, perhaps to take over the world. We are very far from that but perhaps that is a good thing. So what we have is narrow AI which is pretty much machine learning, software solutions that are pre-programed to do a specific task that they do incredibly well but they can’t really start to do other things.

Ben:                 So, it sounds like the definition of, I’m using air quotes that no one can see but, “true AI” is when the machines and the software that we’ve created ourselves have the ability to tackle problems that they have not been programed to be able to answer and that’s when we get into topics like the singularity obviously not really the topic for this show, we’re going to talk specifically about marketing but you mentioned that we have Predictive Analytics, Machine Learning, and that to me all kind of goes into the Data Science bucket.

Ben:                 Talk to me about what the current status of Machine Learning is? How do you define that as a term?

Magnus:           I would like to explain this free technology such together make up this field. To start with, we have big data which is about finding very, very small patterns and correlations in historical data and harvest insights from that. So, for example if you have a bunch of historical credit card transactions in the database, the key is to take data from the text which kind of credit card transactions were in fact fraudulent in the past. I would say predictive analytics is the next step up. Predictive Analytics, it can detect such patterns in the future known data so with having the data from some credit scores, that historical credit card transactions were fraudulent but with predictive analytics it can detect the currently ongoing credit card transactions by determining it isn’t fraudulent. The problem with predictive analytics is that it keeps making the same predictions even if the underlying world changes. So, if credit card transactions from Country A becomes less fraudulent and most fraudulent transactions come from Country B, predictions algorithms that change and increasingly would start to make worse and worse predictions over time. So, what we need is basically a feedback loop that’s basically retrains the prediction algorithms as new data comes available over time. This is in effect machine learning, this is prediction algorithms that can retrain themselves or optimize themselves automatically as new data becomes available.

Magnus:           So in effect, machine learning, we have software that can adapt and optimize themselves as new data becomes available over time and we therefore have almost self-optimizing software.

Ben:                 Okay, so just to re-cap: Essentially there’s a three-prong phase to what we’re calling machine learning which starts with big data, a large data set that which you can use to start to understand patterns of behavior. You then create an algorithm that we’re calling predictive analytics which takes those patterns and understands when something is likely to happen in the future and then what really what machine learning is is adjusting that algorithm or those patterns based on a real-time feedback loop so you have a dynamic way to try and keep your algorithm relevant over a long period of time while the data set changes.

Magnus:           Yes. Pretty much so.

Ben:                 Let me ask you question. You mentioned big data. Is there any way to quantify what is big data and what is just data? Is there a certain amount of data transaction records or volume that you can use to start to really understand when you can create these micro patterns to start getting into a machine learning feedback loop?

Magnus:           So big data typically means a very large data set. But it also means that the data may come in very unstructured format. So it may not be just database records or an excel sheet with no baseline of data. It may be some site log files, it may be sound clips, maps, video clips, photographs. So in addition to being a large set of data, normally the data can also be very unstructured data in multiple forms. But I think it is, as we’ve said, machine learning algorithm for machine running solutions without big data in the sense that the amount of data needs to be massive. But at least the unique algorithms detect certain patterns and correlations and data that typically comes from the data feed.

Ben:                 Okay. So in the underlying machine learning landscape, you have to have large sum of data we’re “Big Data” and there’s not necessarily a rule of thumb in terms of how much data that has to be but generally it’s large enough that it’s unstructured and diverse.

Magnus:           Typically yes but it depends on the application as well.

Ben:                 Let’s change the subject a little and talk about Data Science. It’s a prevalent term and to me I think of scientists, people in lab coats, with droppers and beakers. Then we data which is something that’s in a computer. Let’s join the two of them together. How do you define data science and where does that fit in relative to understanding machine learning and artificial intelligence.

Magnus:           Yes. They are very closely related but you can use data science also for average tasks which isn’t necessarily machine learning. And a data scientist is basically a position someone who is very interested in mathematics and statistics but also having software knowledge and understands the domain feed the business needs of organization. So it’s across the main position where you have a very strong mathematical foundation but also need to understand software science. You need to understand the business and the business value your organization wants.

Ben:                 What are some of the ways that data science is being applied in new formats and new field of marketing?

Magnus:           So I think that in this situation, data science is mostly used in machine learning, predictive analytics, and in my book I outlined I think well over 100 field cases of out use machine learning, predictive analytics in marketing. For example, there are tools that use machine learning for market and competitor intelligence, business intelligence, commercial optimization. There are many of these machines running in email marketing or even adapting customer journeys or AI software where it’s found and which person shall receive which task points in different channels of book cadence with what content. There are many machine learning solutions that can be used in sales prospecting and enriching the data’s current systems to automatically get more data about customers and companies and build the database.

Ben:                 One of the things that you mentioned in describing data science is, is getting into marketing automation, understanding how and when and why to reach out to a customer. Talk to me a little bit about how you define marketing automation.

Magnus:           So, marketing automation is about automating the touch point in the leader or customer. So it is the software logistics that size which person should get what content at what time, in what channel. And dependent on how someone behave they get their task points or dependent on some information we have about them in the data base they can adapt the behavior and create different customer journals and different task points on an individual level. For example, using machine learning we can deploy it using predicative lead scoring that is basically a software solution that can guesstimate how close someone is to buy and by comparing the behavior of leads to people who already bought in the past, we can guesstimate how close someone is to buying. So they get a score between 0 and 100 percent that explains that this particular lead may be 72 percent ready to buy because the behavior of this person correlates to the behavior of people who later become customers in the past.

Ben:                 It sounds like going back to how you described data science, it’s the statistician that understands the practical applications of mathematics to business. And they are able to calculate and quantify patterns in marketing and business behavior to understand how to move the business forward. And then marketing automation is setting up a set of triggers or processes that make it so you can reach out to the right person and the right time with the right content.

Magnus:           Yes and by listening to an negative footprint of every person in the database and adapting the customer journey automatically, dependent on how people behave. But I would like to say that not very many people would actually create their own AI solution or machine learning solution I think for internal use in their company. More likely, companies will just go out to the open market and purchase a ready-made software tool but used as AI to improve their marketing. So there are already hundreds of marketing automations or solutions on the market that use this machine learning or predictive analytics to optimize things like their email send time or the email cadence on the individual person. So, I think most companies wouldn’t exact set up their own data science department and hire data scientists to run their emails and conduct marketing campaigns.

Ben:                 So talk to me a little bit more about which of these functions between data science marketing automation, machine learning, artificial intelligence being the umbrella term. When a marketing executive hears that these are some of the things that they need to master, accomplish, build into their organization. Which one’s are more prevalent and further along in their development cycle? And which terms are likely just to be buzz words and targeting, meant to a marketers attention?

Magnus:           I think that we have to realize AI is already here, machine learning is already here in marketing and there are several hundred of these solutions on the market already that do also different things. I can take a quite unexpected example of how machine learning can be used in marketing and the very software tool that can assess public speeches made by speakers on scene or on stage and analyze what they do well and less good and recommend automatically how those public speakers should change their way of making speeches to improve their results.

Ben:                 As someone who makes his living recording podcasts that sounds terrifying.

Magnus:           Yeah. Well it can also be business leaders who just want to improve their presentation skills. To me that is a very unexpected use of AI. But you can already buy such a solution on the markets. Other things that are much more down to earth I think on most emailing marketing solutions would use machine learning the send time of automation so that it could send 100,000 emails at the same time, then rather send one email at 100,000 different times to optimize the chance that every recipient would open the email because different people may prefer to get the email at different times.

Magnus:           There are email solutions that can adapt the cadence of emails such that a specific person can get them more often as they like that or other people who refuse having emails coming in more often than they would get them unless frequent cadence. We actually have got supports which is very common already, I think most larger companies already have that set up on their site. But I think in the future, you still should consider the chat groups we have up now the text base one where you type the keyboard will largely be replaced with voice-based systems like Siri or Amazon Alexa or Google Home SMART Speakers. And that will change a whole lot of things.

Magnus:           If most internet services aren’t accessed from a keyboard and screen in the future, or they build voice interface while you are driving a car or jogging then that means that search engine optimization will suddenly die. Because I don’t think people googling for something will accept 10 or 20 ads being read out a loud before the search is mentioned by the system. And perhaps they can read out one and that’s very short. So there will only be once for a few search, that needs to be the top one. Everything else is irrelevant because the OIS systems will not read out your ad a loud or perhaps you can’t even have any ads being read out because the people won’t accept it. And that may change the revenues for companies like Google or something.

Magnus:           But I think also that the person using the voice assistant will develop a relation or loyalty to a voice assistant not the brand that you buy from the voice assistant. So for example you talk to the voice assistant and ask it to order coffee beans and the voice assistant will actually choose itself which coffee beans is best to buy but then it factors that price to online reviews and some reviews sites. So in a world where we have a lot of vocal systems perhaps brand loyalty becomes meaningless because people have loyalty to the voice system, to the machine learning which recommend which cocoa beans to buy or it will do it automatically.

Ben:                 I think that there’s a lot of ways that we could speculate what will happen within marketing as artificial intelligence and marketing automation tools get more sophisticated. The other show that I record for Searchmetrics is a podcast called “The Voices of Search” and we just wrapped up an entire month on voice search where we talk about the impact of “Position Zero” and talk about how the laptop or the sort of keyboard interface being less prevalent is going to affect not only the search industry but Google as a company as a whole. So if anybody is interested in that go look for “Voices of Search” podcast we spent a whole month covering that topic.

Ben:                 I think at the end of the day when we talk about the terms that are happening in modern marketing, we have artificial intelligence, machine learning data science, marketing automation. To me, there is sort of a ranking among them in terms of what is prevalent and what is likely to be a developed piece of software that’s going to have a software that’s going to have an impact on your business in the short term and then what is something that is a developing technology that will continue to be enhanced overtime but right now isn’t necessarily accessible to mid and lower parts of the market, sure. Google and Facebook and Apple they are actively working- Microsoft as well, actively working on artificial intelligence and understanding how to take these large data sets and do predictive analytics and build software that has the ability to make more decisions on its own.

Ben:                 I think for a practical application for most marketers, even outside of the fortune 100 and below, artificial intelligence is a service is still developing and right now is very buzz word heavy in my opinion. But things like machine learning, data science and marketing automation are definitely much more accessible and practical right now. In that order where you can have marketing automation tools that help optimize your marketing performance based on simple things like email send times, data science, looking at data sets and picking out the trends, and then machine learning which is understanding what trends are but being able to identify how your underlying data set is changing.

Ben:                 At the end of the day I think that this is a complex topic and we’re going to spend the next month talking about each one of these various subjects in more detail but, Magnus, before we let you go, any last words about sort of the confluence of all these terms related to artificial intelligence and how marketers should think about applying these practices in their business.

Magnus:           So I think that in my opinion we don’t really have AI or true AI, so what we have is predictive analytics and machine learning and some people use those terms interchangeably. To me, AI is very soon going to be a commodity. Everyone will use it. Very few companies will set up their own data science teams and do their own machine learning development I think. But rather marketing teams will just be able to buy commercial off-the-shelf products from the open market and pretty basic social solutions will have machine learning built in to optimize their behavior. I think that in 2-3 year’s time most marketing teams, even in very small companies, will use AI in their marketing software at least to some level.

Magnus:           I think also that after AI comes commonplace, the next wave will be harvesting machines from behavioral data. So we will have about 1 in 3 internet connected machines by 2025. And all those machines will generate enormous amounts of data and if we can harvest the data or insights from that data somehow, we can create enormous marketing systems that trigger marketing automation working automatically when a machine exhibits a certain behavior or pattern. And that will be the next big thing after AI becomes commonplace.

Ben:                 Yeah. It’s interesting when we talk about artificial intelligence and all of the subsequent terms that go into it and then now we’re mixing in well how the internet of things impacts the data set and this topic is going to be prevalent for a long time. It’s obviously developing and Magnus we appreciate you giving us an overview of how to understand some of the different topics of modern marketing.

Ben:                 And that wraps up this episode of The TrendSpotting Podcast. Thanks again for Magnus Unemyr for joining us. If you’d like to learn more about Magnus you can click on the link in our show notes to his bio or you can visit his website If you’re interest in spotting more marketing trends or if you’d like to learn more about Searchmetrics the creator of the TrendSpotting podcast, click the link in our show notes to see our podcast content archive or go to the website. If you have questions or you’d like to be a guest on the TrendSpotting podcast, feel free to fill out the “Contact Us” form on and if you’ve enjoyed this podcast and you’re feeling generous, we’d love for you to leave a review in the iTunes store or wherever you listen to your podcast.

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