In this Marketing Over Coffee:
Learn about the state of Media Monitoring Sofware, AI, Fake News and more!
Brought to you by our sponsors: ahrefs and LinkedIn
The 2020 Media Monitoring Buyers Guide
What made you write the report?
What are the key features of Media Monitoring Software?
The transition from clippings to business intelligence with internal and external data sources
Quality of the data inputs
8:39 We’ve got a special offer for Marketing over Coffee Listeners, you can redeem a free $100 LinkedIn ad credit for your first campaign: Click here to get a $100 for your first campaign!
Protecting the brand when the generated content is beyond the human ability to consume it all
Reccomendation Engines vs. Natural Language Processing (NLP)
Tracking message penetration
AI vs. Fake News
17:30 MoC is brought to you by ahrefs – learn why we use ahrefs for SEO, check out some of the things that only ahrefs can do!
Case study – tracking climate change opinions by country
The secret to rapidly training your model
The legal fight to stop TV scanning
Sign up for the Marketing Over Coffee Newsletter!
Get episodes early, get the back catalog – get the app for iOS! Or for Android
Our theme song is Mellow G by Fonkmasters.
Machine-Generated Transcript
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.
Unknown Speaker
This is marketing over coffee with Christopher Penn and John Wall.
John Wall
Good morning. Welcome to marketing over coffee. I’m John Wall. Today our guest is Eric Schwarzman. I’ve known Eric since the earliest days of podcasting. You were on board, you know, 2005 2006 era, but always focused on PR today, and you’ve got a new report that we’re going to be talking about. So we’ll jump into that. But thanks for joining us today.
Unknown Speaker
Yeah, thanks for having me. It’s great to be back with you.
John Wall
Now, you’re known as the founder of IPR software, you’ve worked in the PR and PR monitoring space forever. Give us a just the overview of this report. What What’s this report you’ve put together and what drove you to do this research?
Eric Schwartzman
So this is the 2020 media monitoring Buyer’s Guide and I was down at the prsa International Conference in San Diego and I was talking to some of the different vendors out there with their products. And there seemed like there was so much misinformation out there and really no easy way for people to get a sense of what products can and can’t do. And originally, I was just going to write a blog post, but it spiraled out of control and became a 54 page Buyer’s Guide. I analyzed and reviewed, I don’t want to say the top 10 media monitoring platforms, because it’s not a definitive list. But I basically started with the vendors that were exhibiting at the PRC international conference, added a few more created a side by side features comparison charts so you can you know, quickly see who does what I did a user ratings charts you can see you know, what their customers think about them. And, you know, for buyers, I also created like a list of questions you can use when you’re talking to sales reps, about their platforms. So you don’t necessarily have to test drive it for a few weeks and see what it does. You And sort of get a sense of their capabilities before you agree to actually give it a try. And then I also analyzed some of the most common challenges that marketing PR people have around media monitoring. So there’s a lot of useful information in the report.
John Wall
Yeah, no, it’s fantastic. I mean, I love the fact that you hit all the major things, you know, you go into the features that people want, you do a great overview of kind of what the state of the nation is, as far as what these tools can do and can’t do and what to expect. But then the big meat to is nearly half of it is digging into that list of 10 or so vendors that are on the edge. And so anybody that’s going to be buying into the space, this just saves them thousands of hours of having to wade through all kinds of, you know, extra stuff. But for folks that aren’t familiar with these products and kind of what they do, what are the big the shortlist of features that people are looking for the software to do for them?
Eric Schwartzman
So you know, most media monitoring services have access to the same traditional and social media news. That’s pretty much square one, right? You sort of take in and ingest all these feeds. at square one, the interfaces will let you introduce bolian keyword filters to sort of hone in on the right information. And then after you’ve done that, you know, you can start to leverage some narrow AI, artificial intelligence algorithms to evaluate sentiment at the entity level, tag articles by concepts, detect geographical origin, stuff like that. You can even you know, try out relevance and sentiment filters. But you know, you need to be prepared in step three to really check those last two for accuracy. Used to be back in the old days that media monitoring was really about sort of aggregating and counting the feature stories and media mentions you got for your client or your company. But now it’s become much more about you know, strategic insights, by sort of finding the intersection of your internal data and actions. journal data, which is why the ability to import stuff like call center transcripts and sales figures, I mean, even stock market data or the Consumer Confidence Index is so important. So you can start to spot patterns and figure out how that activity correlates to your KPIs. So, you know, it used to be about sort of seeing what people are saying about you, and trying to get a sense of what widespread beliefs are. And it still is, to some extent, but it’s also about comparing that now to external data to try to get some actionable business intelligence.
John Wall
Yeah, that’s interesting. It seems to have leveled up from more just, you know, clipping service to full on business intelligence, kind of knowing what’s, what’s going on there. How about so as far as the media sources, you know, you mentioned that a lot of these tools are pulling from similar channels as far as just the raw data. Do you find any variants across the board? Or is it pretty much like everybody in these 10 players have all are all at the enterprise level and you don’t need to be worried about that? Do you do need to dig in and say like, Okay, well, we only follow this little thing here. Like, let’s see how good this tool is at that niche?
Eric Schwartzman
Well, you know, it’s crazy because the answer is unknown. And you know, that may be partially by design, but none of the services Well, with the exception of one, none of the services have a public database that you can search and see what they cover.
John Wall
Yeah, nobody wants to share the secret sauce.
Eric Schwartzman
No and and most media monitoring platforms claimed to have access to more media than their competitors. But just as you necessarily don’t know, all the stations, you get in your cable media package, because it’s changing all the time. They don’t even really know what they get, because they’re dealing with the same source feed providers, Factiva LexisNexis, critical mentioned TV eyes, and that’s changing all the time, too. So it’s not like there’s a schema for a pivot table with all the news. that’s available in Every database from every outlet, it’s it’s changing all the time. Interestingly enough, not only do they not know what feeds they get, they also don’t know how much of the news they get from each of those feeds. Because some may just give you a snippet, some may not give you all their coverage, some may not give you there photos may not give you their graphics. So I did go through, you know, with the different services and figure out, you know, what they got what they didn’t get and put it in that features comparison chart. But the truth is, if you really want to figure out who gets what outlet, specifically, you’ve got to get with a sales rep, and they’ve got to manually figure it out.
John Wall
Okay, so you do that as part of your buying process, you’re gonna have to go in and actually dig in there. And
Eric Schwartzman
if you even know, you don’t really know, particularly in crisis communications, where a conversation is going to break out. So how could you possibly anticipate the sources that you’re going to need from your media monitoring platform,
John Wall
you’d mentioned one of the things in there is you’ve got to This comprehensive list of 17 things that, you know, you should be asking as part of this project and where you go, what are the biggest ones on there? And, you know, kind of what do you suggest? Is it literally, you know, engaging with the vendor and picking off these questions and making that your place to start?
Eric Schwartzman
I mean, of those 17, you know, I wouldn’t skip any of them. And so, you know, when you’re getting into the sales process with a sales rep, I think you know, that having that list of questions in front of you as part of your RFP is really going to save you a lot of time and energy in terms of figuring out what questions you need to ask, so that you don’t find out after you’ve made the purchase that the system doesn’t do necessarily what you need it to do. That was the idea with that. If there was one big takeaway from the report, I would say, you know, it’s probably that media monitoring used to be about gathering, you know, clips, and now it’s about discovering business intelligence. Right. It’s about measuring your online footprint. How many people Have you engaged versus your competitor? What percentage of feedback that you receive through socials positive? And how does that compare to your sales pipeline? Right? It’s about finding the intersection between internal and external data. Right? How does what your competitors are spending online impact your revenue? You know, is there a consistent relationship between those numbers? Here’s an interesting one. If you analyze the boilerplate paragraph of your competitors press releases over time, what strategic insights can you get about their product roadmap? Well, that’s a great
John Wall
point to bring out there. We will dig into that more. But we have to pause just for a second to think LinkedIn marketing solutions for their support of marketing over coffee, time and places everything especially in marketing, but in today’s age of a million messages per minute, and not enough hours in a day, how do you really catch people’s attention? Fortunately, there’s a simple way LinkedIn can help you speak to the right professionals at the right time with over 62 million decision makers on LinkedIn, you’re able to connect with the right business leaders who are relevant to your company. And with LinkedIn ads, you can make sure your messages are getting through to these relevant people. LinkedIn targeting tools help you focus on reaching your precise audience down to their job title, company name, location, and more. People on LinkedIn want to learn they’re on it to grow. In a recent survey. 71% of people said they use the information they find on LinkedIn to inform their business decisions. Marketing on LinkedIn means you get to tap into that hunger to know more. In fact, we do the same thing we’ve been using our Chris’s you ask I answer videos on LinkedIn all the time to generate discussions about business intelligence, artificial intelligence and marketing. Everything from attribution to just getting your Google Analytics to work properly. It’s been a great channel for us for generating traffic. At the end of the day, LinkedIn ads are helping businesses of all shapes and sizes get bigger results. You can try it for yourself LinkedIn is offering a $300 LinkedIn ad credit to launch your first campaign, simply visit linkedin.com slash MLC that’s linkedin.com slash MLC terms and conditions. Apply, and we thank them for their support of the show. Yeah, that’s interesting that you’re talking about the all those competitive advantages, you know, being able to see what’s going on. And, you know, the report opened up a bunch of things for me that I hadn’t even thought about it like internationalization, you know, I never thought that, I mean, that’s huge. If the name of your company or some of your products are different than other languages, if you’re not scanning for that alternate, you know, which is kind of an alternate spelling, you just totally missed the boat, all the way up to virality, being able to, you know, see how quickly things are spreading and where they’re spreading. There’s just a bunch of other things you can get on over there any gut or feel as far as when you go through these 17? Is it kind of a classic thing where, you know, the bigger guys pretty much hit all of them? Maybe some of them not that well, and then there’s smaller ones that, you know, hit a couple of them and specialize in that are good. What’s the state of the industry there?
Eric Schwartzman
Well, I mean, the truth is, you know, there are a ton of media monitoring platforms out there, you know, I covered 10 of them, but there are many more, but there are also a lot Use Cases, right? Like, the reason there’s a lot of cars is because people have different preferences. So I would say, you know, most of the platforms I reviewed do excel in one aspect or another over the others. So picking the best platform, it really starts with defining your objectives. Square One is probably asked the CEO or the CMO, what success looks like. But most of the time, they’re going to tell you your objective is to protect the organization’s reputation. And that’s easy for startups in b2b is because you know that you can manually consume that content and sort of assess what the widespread beliefs are, you can read it but but for established consumer brands, you know, there may be more content out there than you can manually consume. Right. So if the volume of content exceeds your ability to manually read and analyze it, it can be tempting to rely on artificial intelligence. But I mean, the Fact is in that’s that’s still a really risky proposition.
John Wall
Yeah, we’ve talked a lot about that. I mean, there’s so much going on in that space. And yet there’s just a huge wide range in the ability of AI algorithms to be able to sort and properly classify stuff that again, it just seems like it’s right in the same bin. And you’ve got to take that stuff for a test spin and see how well it works for you and see how it matches whatever you got going. you dug in a lot into artificial intelligence and even artificial general intelligence. What were some of the things that surprised you there?
Eric Schwartzman
Well, I mean, I learned a ton about it, because, you know, I talked to academics at Stanford at Oxford, I talk to Chief scientists from Pinterest. I read a ton of studies about sort of where we’re at with AI. I spoke to experts and I mean, I think it’s understandable that people are confused about it, right? Because like if Google knows what I want to search, Amazon knows what I want to buy Netflix knows what I want to watch shouldn’t immediate monitoring platform. Be able to Find me relevant news and online conversations. I mean, it seems like a pretty simple question. But you know, when you dig into it, you realize that recommendation engines are very different than machines that can understand natural language. Right recommendation engines analyze specific patterns for narrow outcomes, like a search term product recommendation, right. But recommendation engines deal with structured data and provide concrete recommendations, whereas news and social media are unstructured, right? You’re dealing with a broad base domain of sources. There’s no consistent standards. And it’s complicated by sarcasm, slang and emotion. So while recommendation engines are certainly impressive, they are incomparable to humans when it comes to natural language processing. If you think about the goal of media monitoring for public relations is to track message penetration. Right. So in other words, you have a message you have targets rather than just counting key Word mentions in articles, right? The more important goal of media monitoring is determine if your targets are repeating your messages. So you’re basically tracking message penetration. And that’s less about aggregating articles. And it is about understanding and identifying concepts and ideas, right that are being repeated in natural language. And unfortunately, that requires a level of natural language processing. and common sense reasoning that automated solutions still can’t deliver, right? This requires what we call artificial general intelligence. And what we have today are neural networks that can achieve narrow AI objectives. Right? They can compare apples to apples, but they can’t compare apples to oranges. And this is what’s required with the true analysis of concepts and ideas, right? And these aren’t just my opinions. By the way, this is the opinions of scientists and academics I interviewed in the report who are leading the natural language processing in AI Charge, I spoke with the chief scientist of Pinterest, who’s also Professor of Computer Science at Stanford about why AI can’t solve the fake news problem. And what he said, is that in order to debunk fake news, basically what you have to do is build a machine that knows all the truth in the world, so that you can say what is truthful and what is not. Right. Essentially, you have to build a machine that knows everything. That’s truth, because only then when you understand everything, that’s true. Can you say whether something is fake? And at this stage, that’s still a bridge too far for AI? I don’t know if you’ve seen Kai foo Lee’s book AI superpowers. It’s a New York Times bestseller. He’s the former president of Google China. And what he basically says that is in order for machines to think as we humans do, that would require a multi domain learning, domain independent learning, natural language processing, common sense reasoning, planning and learning from a small number of exams. apples. And taking the next step to emotionally intelligent robots would require self awareness, humor, love, empathy, even an appreciation for beauty. So these are really the key hurdles that separates what AI does today by spotting correlations in data and making predictions. And that’s what he calls artificial general intelligence, the ability to think on par with humans. So how in the world can you rely on artificial intelligence to determine relevance and analyze the sentiment of traditional social media? If you can’t build machines that understand natural language? Right at the same time, you know, there are some very impressive AI features in many of these media monitoring platforms. But if you reduce the number of results first with Boolean filters before introducing algorithms, you know that there certainly is something you can do to parse and interpret the data. But if accuracy is mission critical, right, you’re going to need to allocate human resources to analyze the data. Yeah, and make sure the machines get it right.
John Wall
Yeah, exactly. And I just see this as a general AI problem across the board is that it takes so much effort to train, you know, training a machine to know all the truth that’s even beyond is you just some of our organizations are Google leading the charge for that. It’s still so much for one company, it can only go so far. And then it’s not going to, you know, that becomes a corporate asset. It’s not going to go into the public good or be shared. So there’s a lot of question marks on how far that can really go, and how well it can work. As we mentioned, digging into data though, we also have to thank hrs for their support of marketing over coffee href makes competitive analysis easy. There tool, show you how your competitors are getting traffic from Google and why you can see the pages and content that send them the most search traffic, find out the exact keywords they’re ranking for in which backlinks are helping them rank. From there, you can replicate or improve on their strategies. It’s got everything as far as if you’ve never done any SEO work for your site. You can get in get running, run some basic reports, see what’s working for you. You can also Check on all your competitors, you can check out all the unique features h refs has got, we’ll have a link in the show notes so you can get over there. And again, we actually use them for a bunch of additional analysis, Chris has been exporting the data from reports that we’ve been running because they do such a great job of scanning everything out there. And we can use that data for further machine learning to get insights that we can’t get from anyplace else. So h refs is our tool of choice for SEO, you want to check them out? They’ve got a seven day trial for $7. link in the show notes, and we thank them for their support of the show. Yeah, so debunking fake news. That’s a big part of it. Do any of the vendors actually take a stand on that and say that they’re able to do that kind of stuff? Or is nobody playing that
Eric Schwartzman
card? No, I didn’t hear anyone play that card. It’s such a common problem. I mean, I was recently at a dinner party. And you know, that was sort of the topic of conversation. Everyone was sort of incensed, and these were not folks in our sector, but people were incensed that the social networks can’t debunk fake news. But the truth is, I mean, that’s a hard problem. I mean, the only way you could debunk fake news at this point No social network is shut down the social network. Yeah, yeah.
John Wall
So there’s really no, you pretty much have to have humans in the loop somewhere there. And even at that point, you can, you know, have triggers so that if stuff starts spreading, you can look at the big stuff. But there’s just so much out there that it’s impossible to tell. Can I give
Eric Schwartzman
you an example? Yeah. So I served as special advisor on climate communications, the US Department of State during the Obama administration, just before the climate conference in Paris, they called cop 15. The negotiators wanted to get a sense of what climate issues were resonating with each of the different countries because if you if you break down climate, right, you’ve got conservation water. You’ve got environmental quality and trans boundary issues. You’ve got global climate change. You’ve got international health, biodiversity, you’ve got marine conservation. You’ve got oceans and polar issues. You’ve got science, technology, cooperation. You’ve got space and advanced technology, and you’ve got wildlife trafficking, these are sort of the nine areas that are affected by by climate change. So what we wanted to do was get a sense of which countries are most interested in which issues so that when the negotiator sat down with those countries, they could be able to address the issues that resonated most with them. So you think about what goes into this type of immediate monitoring effort, right, I had to put together essentially a Boolean keyword corpus for each of these issues with you know, it contains also contains and it does not contain, we also did some keyword proximity. And then we had to monitor those feeds over time. Basically go through 1000 results, and manually tweak the filters to try to make sure that you get the off topic results out but not choke it down so much that you miss good results, which is the big, big challenge. And so, social media monitoring and in trouble media monitoring, you know, it’s one thing to get it so that you’re only getting relevant results. But how do you do that? And not choke it down, you know, so narrow that you also miss good results. Right. So that was sort of the the issue there. In a situation like this, you’re dealing with a mission critical outcome. Right. I mean, it’s our survival as a planet. So, I mean, you can’t relegate something like this to AI. You’ve got to actually manually get analysts on the computer reviewing this stuff, making sure it’s accurate. So that when the negotiator sits down with, you know, whoever is the minister of the other country, they know whether or not it’s oceans or whether or not it’s air quality or what the issues are that resonate most with them, so that the negotiation can move forward productively.
John Wall
Yeah, and how did you feel from all the work that you did to get that built and to train it for doing it for the next round? Was it at least you know, half as much work because you had built everything and at least knew where you were going, like, do you have any feel for what kind of leverage you’re able to get and how much effort it would take each time you do it?
Eric Schwartzman
You know, it’s really a manual process, john. I mean, you know, it was me, I mean, a team of analysts, you know, analysts speaking different languages. I mean, it’s a Herculean effort to do that type of a monitoring exercise, I can sort of give you a great tip around this. So really, you know, the easiest way to approach this type of an exercise is to sort of find the most definitive, comprehensive and current book about the subject that you can, and then use that book to create your corpus. So in this case, Naomi Klein had just come out with this book, This changes everything. And it was hugely helpful because we’re able to get the names of individuals that are behind certain efforts that people who are the climate deniers and everything we were able to build The corporates essentially starting with her book. And then after we sort of watched those feeds that came through, mostly it was the does not contain category that we added keywords to try to eliminate irrelevant stuff. But, you know, it was difficult to do that without missing relevant stuff. I mean, interestingly enough, LGBT content kept coming through in for oceans and polar issues. And so we actually had to put LGBT in as a negative keyword to avoid that type of content. But then we miss some content from LGBT organizations who were talking about climate change. So you know, it’s a challenging exercise.
John Wall
Oh, that’s interesting. So certain organizations like that, that was part of their platform. And so that would show up, you know, more frequently than just the general content. Yeah, that’s I love that idea of grabbing a book and using that for building the corpus to give you a jumpstart as far because we view books as greater source of truth, you know, usually Respected authors have done a ton of research and you can, that’s just so much better than the general scan of Twitter today, as far as like what’s truth and what’s not, of
Eric Schwartzman
course, because you’re you are monitoring for concepts and ideas, and concepts and ideas are not keywords or key phrases. They are concepts and ideas. And they’re expressed through many keywords and phrases. And that’s what a bland keyword corpus is all about. It’s about finding all those words that in relation to one another will get you on, you know, the coordinates of that topic in a digital environment.
John Wall
Okay, how about just other interesting stuff, you know, putting the report together? Was there anything that got in your way you weren’t expecting or, you know, other surprises?
Eric Schwartzman
You know, there was, interestingly enough, you know, the company’s behind traditional and social media businesses, right, they’re in it to generate a profit based on their media, right? So they get their content or they make it difficult to get all their content. So when you’re trying to Listen, you’ve got all these password protected sites with content locked behind firewalls. You know, you’ve got all these premium traditional media sites and some social media networks that don’t want to make it too easy for everyone to monitor because it compromises their competitive edge. In one instance, there was actually a long standing legal battle between Fox News Network and TV is one of the top two broadcast media monitoring platforms. And it basically resulted in the inability for any media monitoring firm to deliver Fox News Network clips. So TV eyes and their competitor critical mentioned after a long drawn out multi year legal battle, and by the way TV eyes and critical mentioned are both covered in the report. They’re still able to monitor Fox News Network transcripts, and make them searchable and deliverable but they don’t have access to the video because Foxy is net blocked them from delivering video clips to their clients so they can essentially distribute fake news with impunity.
John Wall
Yeah, I was gonna say, right. I mean, this is obviously speculation on our part. But I do think that was the number one reason why they did that, because why else would they do it? I you could make some kind of argument. It’s like, you know, they only they want people just to be watching all the time. And that’s the only place to get that.
Eric Schwartzman
Does it track that if you block other people from redistributing your content, that people won’t come to you to find that content anymore? Because it’s news, right? There’s the classic cnn network. Right? There’s no shelf life. So how does blocking people from redistributing and after the fact hurt you? I think it helps you.
John Wall
Yeah, that’s a great point, right? By the next day, this stuff is dead. So
Eric Schwartzman
it’s just going to drive more traffic back to your channel.
John Wall
Yeah, yeah. You’re making yourself bulletproof by saying it’s okay. Whatever we say today because nobody’s ever gonna be able to go back and get it.
Eric Schwartzman
I guess. They were unhappy with how the comedy network reuses that content on The Daily Show?
John Wall
Oh, there you go. Now that Yeah, that’s a that’s, I could see that totally working to try and squelch that kind of press that makes a lot of sense. Oh, yeah, that’s very interesting.
Eric Schwartzman
What I like to do most is analyze and learn about new categories. So for me writing this report was a lot of fun. I mean, I have probably three months into it. It’s 54 pages. And I just hope people use it to make smart choices about, you know, their media monitoring investment.
John Wall
That’s great. Okay, we’ll have a link to that in the show notes. It’s at Eric Schwartzman comm slash monitoring mln it or IMG. But of course, you can sign up for the marketing over coffee newsletter and we push you these links on a regular basis. So you don’t have to remember them and they are of course in the show notes over at marketing over coffee calm. But that’s going to do it for us for today, Eric, thanks
Eric Schwartzman
for joining us. Hey, thanks for having me. It was great connecting with you.
John Wall
Alright, that’s gonna do it for this week. So until next week, enjoy the coffee
Unknown Speaker
You’ve been listening to marketing over coffee. Christopher Penn blogs at Christopher s pen.com. Read more from john J. Wolf at jw 51 fifty.com. The marketing over coffee theme song is called mellow g by funk masters. And you can find it at music alley from meh vo or follow the link in our show notes.