Your boss told you to do some “social listening.” So you set up some Google Alerts. And you get a few hits. But miss most of them — especially the most important of them all: unbranded mentions. Business intelligence expert Cam Steed from Smashed Abacus tells you what you’re doing wrong in this episode.

About Cam Steed:
Cam has worked in and around social media for more than six years. Initially starting in the Agency world, he moved over to the social media monitoring company Sysomos, where he founded their in-house reporting service (working with clients like Google, Apple and Coke). In 2014, he left Sysomos to start his own social media business intelligence agency, where he helps clients manage their online brand reputation, strategy and social listening.

Tod: I talk to a lot of people who are in sort of junior to midlevel management digital marketing and their bosses told them, “We need a social listening plan put in place.” And then, they kinda go back to their desk and they don’t know what that is, how to do it, what tools to use, how much it’s going to cost, how to measure them. I mean, this whole category of social listening is one of those phrases, I think, that people… It’s just like the word multimedia used to be 15 years ago. It could be anything. So what is social listening to you?

Cam: Yeah. And that’s a great question. In fact when I present at conferences, that’s one of the things I try and do at the beginning, is present almost a glossary of terms. And so going forward just we’re on the same page as I casually make this refer to these phrases regarding social media, just that we’re all understanding. Social listening, as I would define it, is the act of understanding or reading or trying to interpret the content that is generated on social media by everyone except you. And so the way we break it down as an agency is, is by talking about inbound versus outbound social media content. And what we mean by outbound is any content that you, yourself, are responsible for publishing on any of your social media channels whether it’d be Facebook, Twitter or LinkedIn. This is the content that you have 100% control over. And that is outbound. You are contributing it out in to the social media landscape.

Inbound though, is any mention that talks about your brand. It’s a reference to a product or a service or stuff or a store or your company or people within the company. That is the inbound. And then, there’s this third component which really doesn’t have a name which is just the other. It’s the social media content that everybody is producing every single day that may not necessarily reference your brand, you may not have necessarily published but exists and obviously is driving social media engagement up and up and up, week over week, month over month.

Tod: We should say that not only are you founder and CEO of your own company, prior to this in your previous life, you are one of the people behind Sysomos which is one of the single most respected social listening tools out there. I keep trying every month to get to complete account from them and never seems to work but it a phenomenal tool and I have looked at Radian6. I have looked at SM, SM2. I have looked at all of them and I actually do believe that Sysomos is the best, best in brand out there. So let’s just talk a bit about that, that social listening component. What is the single biggest mistake that people make when they are doing social listening? How are they getting it all wrong?

Cam: The great question. The answer, in my opinion, is simply that they are too focused on mentions around their own brand or, or it’s just related to their own and not tapping in to the vast amount of content that is, is being published by everybody else, that other content that I referred to earlier. It’s something that we have time and time again whether talking to a four to five hundred brand that’s a household name or PR agency of two people representing a small company in their hometown. It’s just not something that a lot of people prioritize when it comes to their very limited social media resources or listening resources especially.

Tod: Why, why do they care about unbranded messages? Why do I care about people talking about the category I’m in, but they are not really talking about me?

Cam: Well, I think fundamentally, there lies opportunity. In this day and age as we strive to become more original and unique the way in which we champion or leverage social media, so too then lies the requirement to try and really seize or take advantage of content in a new and different way. We see any number of presentations about how to create or captivate a content through your social media channels or how to drive your response rate or engagement rates where people will talk about your brand or post your Facebook page. But, the one thing that nobody is really talking about is how to maybe leverage or take advantage of all that other content that may or may not feature a brand name or reference at all. It’s really gonna help you gain a competitive advantage and I would imagine a year from now or would be maybe a little more of common knowledge, but right now, really it’s not something at least in my experience that a lot of brands are doing.

Tod: So practically, what you are talking about is, let’s say that I own a Domino’s Pizza franchise in Burnaby, BC and unbranded mention that you are talking about this unbranded content or people saying in my market region, saying, “I’m really hungry right now.” And then, I dive in because somehow I’ve got social listening tools in place to look for the word hungry or pizza or famished. Is that what you mean?

Cam: Yeah. And that’s exactly and so what put you on the tools first and foremost because maybe some people are immediately writing this off, thinking, “I don’t have the time,” or, or, “We don’t have the tools in place to do that.” There are any number of tools that can tap in to the type of content that we’re talking about here. Some of them being free like Hootsuite for example and some of them being paid like Sysomos map platform. Those are two examples of the different ends of a spectrum in terms of a listening software that would allow you to identify exactly that you are fundamentally looking for content that is being published by users and almost real time that address a concern or an issue or a problem or a need that they have. But you as an organization really could be delivering on if only you had the ability to see it.

Tod: What if it’s just too much information? Like for example, my company engageQ, we just wrapped up working with the Grey Cup candidate superbowl essentially right. So, we, we handled all their social media. I supposed I could have used a tool like Sysomos or, or any of these things, Sprout Social or Hootsuite to look for the phrase football. But what, what am I gonna do with… You know, I just gonna put it in to a massive Google spreadsheet somewhere and review later for sentiment or something? Like, what is the actually val… Or, am I going to look for the high influencers and talk to those people? What do, what do you do when, when you got just an enormous volume of stuff in your category?

Cam: Yeah. And that’s a good question as well. The truth is, believe it or not, that’s something we still wouldn’t have quite readily in the social listening space. The jargon or the industry jargon phrases is “trying to boil the ocean”. And I would say in that scenario, looking for a term like football in a conversation about the Grey Cup this past Sunday is just too vague. You’re intuitive too much. It’s just like Ford is trying to say, “Show me every mention of car.”

Tod: Yes, exactly.

Cam: “…that’s happened in the United States this week.” You’re gonna get millions of mentions. Some of which could be seemingly valuable. Others are absolutely useless. And so what we try and focus on is instead of saying, “Show me everything about cars,” it’s to try and identify what your objective is or more specifically to try and identify questions that people are asking, that you may have the opportunity to then answer or at least enter in to the conversation as a trusted individual. So, for Ford, rather looking what everyone is saying about cars, what is it you’re interested in specifically and if it is in fact maybe tapping in to some of that unbranded content that other content, let’s maybe try and focus on questions or issues that people are having that you may try to assist with, like for instance, which is a perfectly likely tweet to see. Somebody texting that their car just broke down or that they needed help with something.

In fact, today just doing a quick look, I did see in the last 24 hours, a 193 Americans had posted that need an oil change. I don’t know if Mr. Lube or any of the major chains are listening to that kind of conversation. But, there is potentially now just over 200 opportunities that a brand like that could engage with and even influence or that is telling you in real time, “I need help. I need a service that you provide.” And, if you can come back to me with a, you know, “Hey, I hear you needed an oil change. Just want to let you know 10% off right now,” what a timely and I think effective use of your social media listening effort to, to try and tap in to those types of people, reliable.

Tod: Yeah. How reliable Cam, is geotargeting here because let’s use that example. Let’s say that I own an OK Tire store. I got to get an OK Tire instead of Kal Tire because they’re our client. So, let’s imagine that OK Tire is in Toronto and we’re in Etobicoke or something as there is in fact a store in Etobicoke. And, you hear the197 mentions in the US. Let, let’s just assume for, you know, comparison sake that a 100 mentions in Canada of oil change. How reliable are the tools geotargeting for me to be able to say, sure, there’s a 100 people talking about it in Canada, but I just wanna know the people that are within say a 20-kilometer radius of where my store is.

Cam: Yeah. Now, that’s a good point. It really depends on the tool. I know that Sysomos, it’s map platform does give you the ability to triangulate or geotarget down to the city level, but I also know that other tools like Hootsuite will allow you to put filters on that say only notify me if this type of conversation has been found to be with, being posted by an individual that has been identified as being from this particular area.

Tod: But are those reliable? I mean, how are they doing it? Are they doing it by the IP address and doing it at reversed look up to say, okay, well, this person’s server is in Burnaby or Etobicoke or is it based on their Twitter bio? I want to get a sense of just how reliable that geotargeting actually is.

Cam: Yeah. So, Sysomos which I think and maybe I’m biased having been involved with them earlier in my career, Sysomos does it based on individuals who identify their location either through things like their bio or through their conversations, really publicly disclose the location. We don’t go by server location, otherwise, every single Tweet would say that the user is from somewhere in California.

Tod: Yeah.

Cam: Then, it is to be taken with a grain of salt and it is to be, I think set something that maybe at an individual store in the middle of downtown in Etobicoke if you can call Etobicoke having a downtown. It might be a little bit challenging, but we duly speaking and it’s interesting you mentioned that. One example that I’ve heard that I love was for a disinfectant, a company that makes disinfecting wipes, not by any stretch of hot topic on social media and teens are running to their smart phones to statu-, or post status updates about how they just wiped down their counter with a disinfecting wipe.

And so when it comes to social media listening, knowing that brand mentions where maybe filling far between, one approach that was undertaken through, through our sort of joint collaboration, was to try and listen geographically throughout the United States people saying on a week to week basis that they were getting sick or coming down with a flu or getting a cold or homesick with a flu. And then, week over week over week, identifying which cities in the US had seen an increase in that type of conversation week over week over week. That company was then able to market or tailor their communication to those specific areas to try and influence or at least become front of mind prior to a purchase decision. Because really, other sales information withheld on is the people who had actually chosen a product and purchased it. They didn’t know what markets they should be focusing efforts to try and become front of mind then. And so here we are focusing on content that has nothing to do with the brand, its products or any of its offers, just people saying they’re getting sick which we found categorically there were a lot of on social media.

Tod: It’s really important and that sentiment analysis can, can provide a really good kind of early warning mechanism for things going wrong. Back when we were doing the Grey Cup a couple of days ago, we started seeing in our listening tool that we use, Sprout Social, a whole bunch of people complaining about the concert stage as it turns out the stage for the halftime show was pointed at about 10% of the audience. The in bulk of the stadium couldn’t see. It was looking in the back of the stage of something like that. So, we kind of started picking those conversations up and realizing, “Okay. This is an issue we need to dive more into.” But how do I mean, we’re a great fan in our company orfZapier which is a middleware, Zed A-P-I-E-R. And you can connect one tool to another. So you could, for instance, any time a client of ours logs in to review their invoice, my entire team gets notified. When they pay their invoice, the entire team gets notified. And, that is FreshBooks on the invoicing side, Squiggle on the chat room side and to connect those two is Zapier. It’s a little over technical, but it’s essentially kind of middleware. How do I get that kind of functionality through social listening tools? Because I don’t wanna be in Hootsuite, in Sysomos every hour running queries just in the odd chance that, that a branded mention or an unbranded mention that I could take advantage of pops up. Are there ways to be alerted to that as opposed to just kind of always being in there searching?

Cam: Yeah. Again, speaking from Sysomos standpoint again based on my experiences, I know that one of their platforms, product called Heartbeat, allows you to create a notification or alert system based on a quick configuration or keyword configuration of your choice. So if you set up within the tool that you wanna be notified every time somebody from Ontario or Toronto mentions or posts something that features certain types of phrases or keywords, they can notify you either daily, weekly or even as it happens. But, I also know that that functionality exists in almost any listening tool that, that’s out there any platform collecting content from an online space will likely have some type of a notification system. Then again, if there’s not the automated group, there is always, and Smashed Abacus provides these services, agencies that are willing to take that work on yourself. Agencies fundamentally exist in the concept that clients are just too busy or don’t want to do it themselves. Maybe they don’t have the expertise. That’s where agencies like ours try to help those clients.

So that disinfecting wipe example that I gave you was, was not a client weekly looking at which cities in the US that’s seen an uptake of people complaining that they were getting sick. That notification took the form of a weekly report deliverable that they had on their desk Monday morning that we provided to them. So there are any number of different ways that somebody can get that information without having to be in tool always on themselves. But I will say, with all that being said. I think the most effective way to try and get to people who are expressing an immediate need or issue or concern is to have that always on mentality. And for companies or brands that have the resources to do that, I think they see a tremendous value or benefit in that approach.

Tod: Sysomos costs a lot of money, thousands of dollars a month, incredible support, great query, they help you out with it, but it is not a cheap tool to have. What you’re describing sound a little bit like Google Alerts which is free. Why wouldn’t I just use Google Alerts for all of my social listening?

Cam: You absolutely can. It’s entirely up to you and your requirements. There are features and functionality of both that are different depending on your needs or objectives that maybe of value to you or not. That’s what I said, going back a few minutes ago, that there’s a spectrum of tools that exist out there and really in, in, in creating our own agency this past year, one of the biggest and primary objectives of ours was to try and listen or at least sit in on a demo for many of the different software, social media software tools that are out there. Because what we found is that whether all “social media measurement tools,” they run the gamet in terms of features and functionality, cost, fee for use, really it’s, it’s a wide open landscape out there. And so there are options from free to thousands per month. Just like cars there is usually one that fits your budget. You just have to do a little research to try and figure out which one is the right one for you.

Tod: The car that fits my budget is a 1981 AMC Eagle.

Cam: Nothing wrong with that car.

Tod: [laughter]. Let’s talk about sentiment as well because this is a pretty crucial part, sentiment being the mood essentially of the, of the piece of content out there. This is embarrassing. Literally, 15 years ago, I started a company called Mindfly that developed sentiment analysis technology. We didn’t know that’s what it was going to be called. But it basically would scan through. This is before APIs, XML and stuff, RSS. So we would screenscape Yahoo Finance pages and look for the words that people were using. So if they started talking negatively using negative terms about a publicly traded company, we would send that publicly traded company an alert saying, “Wait a minute.” And that essentially is kind of sentiment analysis as it exists today and you could see some kind of low range versions of it in tools like Sprout Social. You get a very general sent-, Sysomos who provide you with sentiment analysis on any number of queries you want. But here is my question Cam, how reliable is that? Like how good is it at detecting the subtleties of the English language to know that when someone says, “That was really sick,” that they’re talking about sick in the way young people say good and not sick as in I have a horrible disease. How, how reliable is sentiment analysis?

Cam: You know, it’s interesting. I get asked that a lot and having been a former employee at Sysomos was one of the questions I could get most often and not having seen a number of other tools. The sentiment issue I think is really one of the biggest or I think the most wide open aspects of social media listening. But I have yet to see a company automatically or an automated approach that’s really champion. Regardless of the accuracy that they may tell you a tool has and it varies greatly depending on what you’re doing. There is never going to be someone… sorry, a system out there that are as sophisticated as somebody being able to manually or humanly interpret that information.

Tod: Right.

Cam: I have worked with clients that throw automated sentiment right out the window. We don’t, we don’t care about it. It’s not accurate. Even if it is accurate, there is a little we can do about it. We’re not even gonna bother. And I’ve had other clients say it’s the biggest or most critical aspect than the listening that they are looking to do. From an agency standpoint, and I can only really speak from our own expertise, we’re gonna do it manually. We have found that a manual approach to sentiment analysis is the best way to do any sort of listening.

Tod: All time consuming that .

Cam: And there in lies all, I think competitive advantage. It is an extremely time consuming process. There are only a few companies out there that are willing to do it, but I can statistically say that we are phenomenally accurate with our sentiment analysis and it’s something that provides a tremendous value to clients. And one of the reasons that we’ve been as successful as we have is we are working with the clients that we have. If you have the capabilities to do a manual analysis, it is without doubt the only way that you can really get that truly accurate sentiment read.

Now, with that being said, I will say some general best practices. You’re gonna have higher sentiment if you do an automated analysis, would be more specific of listening exercise you’re doing. So if I said, “What is the online sentiment around an iPhone 6,” and I’m gonna get a number. Let’s say, it’s 25% positive, 50% neutral and 25% negative and we can poke holes all day and whether or not that number is true or not. But, I will categorically say that the number would be a lot more relevant or accurate if you said highly specifically what people are saying about the iPhone6’s battery life or the earphones. Then, when you start to get in to specifics, I find that the sentiment really, really changes because iPhone6, people are saying millions of things about, whether it’s that they have lost it or it’s complicated or they hope to get one for Christmas or save $2 on a Black Friday deal. There’s just too much conversation out there. But when it comes to something like the iPhone6’s battery life, this is a product feature that is highly polarizing. Some people really hate the battery life and the fact that they can’t go more than 8 hours without plugging in. And generally speaking, social media is a good medium for them to show that opinion.

And so it kind of tying back to the whole not trying to boil the ocean concept. When it comes to sentiment, I truly believe your best bet is to try and get very specific with what you are trying to understand the sentiment aroung knowing that if it’s just your brand, you’re gonna have a heck of a time trying to figure out exactly what people like and don’t like and probably will end up bucketing conversations by product features or attributes any way which is what we’re saying you do right from the deck.

Tod: I have to say Weston is incredibly good at unbranded mentions or they do mention them but they’re looking for the word Weston and that’s a great term because it’s not Vancouver Hotel, you know. It’s something that’s easily searchable. A funny story in our book is, a couple of years ago, I was down in Seattle giving a speech. I was staying at the Weston and as I often do, I have forgotten all of my toiletries at home. So I called down at the front desk and I said, “Listen. I had forgotten like some stuff. Would you mind sending like a little cheap toothbrush and toothpaste up?” And they said, “Yeah, no problem. That will be 2495,” and she hung up.” And I said to myself, “Twe-, twenty, 2495? for a cheap…” Like Holiday Inn would give that to me for free. So because I’m kind of an asshole, I went on to Twitter. You know it’s coming, right?

Cam: Oh, yeah.

Tod: I went on to Twitter and I said, and I did not use the phrase, I didn’t use the username Weston. I didn’t specifically address it to them, but I just said, “I cannot believe that Weston charged me 2495 for a freaking toothbrush that I can get for free at any other hotel.” Within, Cam, three minutes, three… No, no. My bad. I’m gonna say five minutes. Sorry. But very quickly, the general manager of the hotel I was staying in had called and said, “I just got a call from…” And he didn’t say the social listening team, but that’s basically who it was. “I just got a call from our social media people.” He said, “I understand there’s been a massive miscommunication here. The number she was giving you, 2495, was not the price. It’s your room number.” She was saying, “Okay. That will be 2495.” That will be room number 2495. And then I felt horrible and of course I tweeted out all sorts of apologies and I flagellate myself whenever I can in public about that.

But interesting kind of story because that’s a brand that is not just looking for their own mentions, their own usernames, but just looking for the brand in general. And maybe they have some sentiment analysis where they detected that the tone of my tweet was really bad and that went up on the queue faster.

Cam: Yeah. And that could be the case. When it comes to something like Sysomos or any sort of keyword based platform that will allow you to configure alarm words in addition to phrases, that was one of maybe, best strategies that we would imply or implore say, let’s look for your brand name, but then only show us mentions when your brand name is referenced along words like sucks or terrible or awful or miserable, you know. Let’s not rely on a system that decode what it feels are truly positive and negative words. Let’s kinda, you know, manipulate it or skew it by saying, “I only want the mentions where my brand is mentioned in relationship to that.” But you know, even that approach have its flaws because we would see tweets from people that said, “I’ve been waiting at my gate for three hours. Air Canada sucks, from now own, I’m flying West Jet.” And so tweet like that from Air Canada standpoint find it negative, from West Jet, it’s extremely positive. Good luck trying to rely on automated process to try and flag something like that to either company. That’s just the nature of the complex language that we deal with on a daily basis. So it’s not a perfect Science one that we are getting better at I think and people are continuously try to evolve but yeah. It’s whether they are one of the most challenging or dynamic aspects of social listening.

And again, to tie it back to that other or unbranded content, I mean, there’s obviously that people who say, “I’ve gone to your store and I had a terrible experience,” or “I bought your product and it’s awful.” But again, to keep in mind the companies that I think will gain a competitive advantage to go after the unbranded content are people who are looking for tweets or posts from people that said, “I can’t believe my hard drive just died. What the hell am I gonna do?” Put a tweet like that in front of Western Digital with any sort of urgency or frequency, there lies a huge opportunity for them to directly engage back to that use and say, “God, that sucks. What an awful experience. Sorry to hear that. How would 10% off a new Western Digital hard drive treat you?” That gives them a competitive advantage that I don’t think any other company in their spaces is leveraging right now. And nowhere in that tweet that say, “I hate Western Digital,” or, “They make terrible hard drive,” that’s a completely other area in the listening process when it comes to looking at brand reputation or brand reputation management. But those types of consumer criticisms or feedback for a brand that can be always on or whether through an automated tool, a manual process, an agency or a free tool, regardless. If a brand is able to monitor that, that’s where the real value lies in social media listening.

Tod: Let’s say that someone is listening to this podcast. They already have a social listening tool in place or let’s say think they do…

Cam: Yeah.

Tod: …what else can they do or what would you recommend they bolt on or change if they already have a social listening program?

Cam: We run in to it all the time as an agency. We’ll meet the people at conference. “What do you do?” “Oh, you’re Smashed Abacus. It’s a business intelligence agency.” “Oh, yeah. We, we already use it. We’re good.” Social listening is such a wide open landscape with so many different aspects or components to it that I have yet to meet somebody that’s doing it all. There is always another way that you could be listening or interpreting the content that’s out there. So if you’re a brand and you’re on top of the inbound, an outbound already or you’re looking at what people are saying about your brand and then using that in your marketing next to them influence the content that you’re posting, you are well ahead of the curve in terms of an online social media strategy. But are you listening to what people are saying about your competitors? Are you listening to what people are saying about the industry or products in general? Are you trying to take advantage of opportunities for people who are expressing frustration or annoyance or dissatisfaction?

People dissatisfied with service or product that they have experienced. There is always another social listening avenue or approach. And yes, you can bang your head against the wall trying to approach them all. So the best piece of advice is come in to this with an objective, to go back to your point. Medium, small-medium business brand managers had been told to adopt a social listening strategy for their product or brand for 2015, you need an objective. To simply go to a product whether it’d be Sysomos or Radian6 or Crimson Hexagon and start forking about money to listen to what people are saying about your brand, you’re just gonna bury yourself in quick sand. You really do need to have an objective and strategy and really steps in place to try and get the most out of the social listening that you’re doing. And as you trial and error, don’t be afraid to say, “Hey, listen. Nobody is talking about our brand.” That’s not a bad thing. Maybe they are just talking about it in a different way or the industry in a different way. Like that disinfecting wipes for example. Nobod’s talking about it, but doesn’t mean we cancel all subscriptions and give up on social. Well, different way or strategy, but we could take to try and get some insights around what people are saying about our industry or space that we’re active in and how we might be able to get an advantage by taking advantage of it.