Being Smarter with the Long Tail of Social Monitoring and Influence


When social media first began gaining popularity with brands, the first thing they wanted to know was, “What are people saying about us?”

It’s understandable – whereas before brands would only get to find out what the public’s perception of them was when the cash registers slowed down, now they could get insights on what was being said before it became a problem, and the perception of their response to that conversation (positive or negative).

This led to a booming market in social monitoring platforms. Companies sprung up with technologies that could monitor millions of conversations, send alerts to brand managers, and define the changing sentiment around a brand and the subsequent buzz that went with it.

All for a very nice premium, too, with licenses running into thousands of dollars per month for just a single license.

But the truth of the matter is social monitoring is flawed, and will continue to be flawed, while we still think in bits and bytes when it comes to human behaviour.

We Are Not Machines

The basic premise of a social monitoring platform is simple:

  • Choose your industry;
  • Choose keywords to monitor (brand, product, person);
  • Set up your alerts;
  • Define your goals (buzz, sentiment, volume, leads);
  • Gather data, report, refine, rinse and repeat.

Sure, there are other areas of data that brands may set up, depending on their goals – competitor intelligence, for example. But even this basic set-up of goals and tactics highlights the flaw in social monitoring – we’re hoping humans behave like machines to tell us what we want to know.

For most social monitoring platforms, the technology is still fairly basic in that all it does is monitor online conversations for certain keywords (much like Google scans the web for your search terms and then gives you a series of results).

The problem with this approach is it requires linear thinking on behalf of the target. Instead of true natural conversations like the ones you have with friends on Facebook, monitoring tools are often looking for non-connected scripts. Take the following example.

I’m in my house, freezing my butt off in a typical Canadian winter. I go online to moan, and say something as simple as “Being cold sucks.” There’s nothing really there for monitoring software to pick up. Or is there?

If the software was advanced enough, there are multiple reasons for me being cold. Is there a hole in my window? Is my roof insulation not working properly? Is my furnace broken? Am I struggling to pay heating bills?

Immediately, there are now four opportunities for four vendors to take an interest in me – glazier, roof insulators, HVAC companies and maybe even my bank, to see if they can help financially.

All from three little words, instead of a monitoring solution looking for me going online and asking “Know any good HVAC companies in Southern Ontario?”

You can see why we still have a way to go when it comes to monitoring. But that’s just a part of it.

Social Monitoring and the Influence Effect

Let’s take monitoring one step further, when it’s being used as part of an influencer outreach campaign.

When brands use influencers, they need to know who’s making the buzz happen and who’s creating action from intent. Otherwise, they’re just shooting in the dark while paying thousands of dollars to social scoring sites for putting them in touch with the influencers in the first place.

So, as a campaign unfolds, brands use monitoring platforms to see where the conversation is stemming from (influence solution partners can offer this information but you should be monitoring for your own needs as well). They track the times an influencer speaks, and whether this causes a trickle or ripple effect.

And this sets up another problem with monitoring at that high level – it doesn’t take into effect all the disruptive factors that help a decision be made, positive or negative.

Influencer disruptor paths

For example, I see an influencer talk about the new Ford F-150. I trust the influencer (he or she’s a car geek, just like me), and I like the mix of fuel economy and torque that the F-150 offers. I’m sold, and I mention as such to the influencer on their blog, so that goes down as a positive net.

But I’m not the decision maker when it comes to finances – my wife is.

So, as much as I love the truck and as much as I give off the vibe that I’m moving beyond intent to buy to actually buying, based on an influencer’s write-up, I don’t buy, because my wife has rightly said we need to go on vacation this year to unwind, and the money needs to go to that.

The effect of that decision isn’t felt, because the monitoring only stayed with me until I was a positive result for the brand and influencer. Had the monitoring or influencer program stuck with me for a week or two, they would have seen me jump online to say, “Vacation this year, truck next year.”

Instead, the brand wonders why there was a positive effect that didn’t correlate into a sale; the influence program is questioned for effectiveness; and the monitoring solution fails to follow up on my secondary conversation.

Take it one step further – let’s say the software really digs into who I’m talking with and can filter them into relevancy, as well as alerts if there’s a follow-up to our original conversation online. They would have picked up my wife speaking with her friends online and saying, “Danny wanted a truck, but we really need a vacation this year, so we’re doing that instead.”

The result would be immediate – the influencer program clearly worked, it’s just priorities that take precedence and, in this case, a vacation was a higher priority. But the message about the F-150 came through loud and clear and, had the vacation not been the disruptor in this case, the sale would have been completed.

We Need to Be Smarter with the Long Tail

Now, these are hypothetical examples, and there are companies that are trying to identify not just the main conversations, but the secondary and tertiary ones too. In our book, we highlight the ones we feel are making great inroads, and dedicate a chapter into using these platforms for your influence campaigns.

But as hypothetical as they are, they also clearly illustrate where we need to go, and that’s into the Long Tail of monitoring and/or influence. We can’t just stop at the result – we need to understand what made that result happen:

  • What diverted an action (my wife being the logic to my emotional decision);
  • Where the follow-up should be (in this case, reminders that I’m in the market for a truck in 12 months time);
  • What language tipped the emotional purchase (prior to the vacation becoming a disruptor);
  • Where the true result came from (in this case, a few weeks after the perceived success).

We’re not there yet, and while social scoring continues to be the lead when it comes to measuring influence online, we won’t get there anytime soon. The good news is, companies are moving away from scoring and really digging into all the data that’s available to us.

When monitoring catches up and combines its resources with the knowledge we get from identifying true influence, business will never be the same again.

Ready to start the next wave?

Sign up for free weekly content

Enter your first name and email below to get my free weekly newsletter with the latest posts, recommended reading, content tips and more.

(I respect your privacy and will never spam you)

Blog consulting with Danny Brown

    Leave a Reply

    Your email address will not be published. Required fields are marked *

  1. says

    You make a really good and interesting point here, Danny. However, as someone who represents a social media monitoring company I have to disagree with some of what you said (but not all of it). Yes, it’s true that software like ours works off of the keyword system where we find the keywords that a client has deemed to be relevant to them and bring those in for engagement and measurement purposes. Of course, as technology evolves we’ll be able to go a lot deeper into the conversations, but currently a lot of what you’re asking for can’t be done at a level of excellence worth sharing with the public.
    Now, where I see the real problem is in the laziness of some marketers (or communicators or whatever) these days. A lot of them don’t think to look past that single piece of information that the system has brought in because it contained a keyword. Something I always do after I’ve found a piece of content (let’s use Twitter for the example as it’s the most public conversational back-and-forth network currently) is look back at the users involved in the conversation to see how they arrived to the tweet that’s been brought into the system. I need to the context of why someone has mentioned our company or one of the terms I’ve deemed relevant enough to track. After taking in all of this information do I make my decision on how to act on it. The problem is that too many people don’t take the time or make the effort to do things like that. They think they’ve paid to license a tool, that should be enough.
    Like in your example of the truck, there is a lot that can be missed by a tool or even by a person. But there’s also a lot that can be learned by doing a bit of research and putting a little more time into what they’re doing. I think that this is where people are falling short. They’re not putting the effort into things that require effort.
    Tools can be very helpful, but you can’t always rely on them to do all the work for you.

    Sheldon, community manager for Sysomos & Marketwire

    • says

      Hey there Sheldon,

      Thanks for dropping by, always enjoy your perspective mate. Completely agree with the laziness aspect – and marketers are some of the laziest people out there (at least, some of the ones I’ve come across in the past are). However, I’d also flip that just a little and say that can also come down to the education (or lack of) from the monitoring vendors.

      I’ve used a few platforms where essentially you’re just dropped in with the keys to the race car and told, “Off you go.” No training, poor materials, crap customer service for support questions. When you have that lack of service and knowledge base to start from, it’s definitely going to impact your effectiveness at digging beyond the basic information.

      There are some really cool developments happening in the natural language space – nModes from NSM being one particularly interesting concept:

      Interesting time to be in this space, for sure.

  2. Sara Chi says

    First of all, kudos to you Danny, smart men always listen to their wives :)

    Secondly, while majority of marketers/PR practioners/vendors are still scratching the surface (and let's face, it's much easier this way), time is something people don't have and reluctant to invest in, like where is the ROI? :) Sentiment analysis and True influencers identifying are two codes waiting to be decoded and it'd interesting to see some pioneers in this space.

    • Danny Brown says

      I may be dumb but I'm not stupid, miss. 😉

      Ironically, the ROI question is one that can easily be measured in the influence space, if you move away from using social scoring as the main approach and go deeper with multiple platforms and informed analysis. Something a certain book talks about in depth… 😉

  3. says

    Very good points, Danny! At the moment, software isn’t a substitute for human understanding. E.g. Someone who takes the effort to look through someone’s tweets, and understands their situation.

    I agree with you that software needs to go beyond simplistic ‘sentiment analysis’. I.e. Ford looking through your tweets to know your wife holds the purse strings. It needs human beings to individually look at your situation.

    So until we can build a super-sophisticated megarobot can read through tweets like a good salesman – I think there will be no substitute for human beings.

    Therefore marketing/sales departments will be wanting more good salespeople who’ll read through your tweets and monitor/follow up where you are in the sales cycle.

    As aggressive and creepy as that sounds – maybe that’s where it’s going. What do you think Danny?

    • says

      Completely agree, mate, and it ties into the approach @40deuce:disqus refers to in his comment. It’s what we used to do, anyway – remember the old school car salesmen who’d get to know everything about you, or the classic headhunter that knows more about your life than you do?

      There’s a huge role for data intel analysts to play in this space, and will be for some time, until the machines can accurately predict an action based on a history.

      Cheers, sir.

  4. says

    I don’t like the term campaign. It is dumb and something the ad industry conjured up to suck money from their clients. It denotes beginning and end. And rarely is there an end even when a purchase is bought. What if you bought the truck and it sucked. It was a complete let down. Or if I bought all of @ginidietrich:disqus ‘s books, now she hopes I love her brand so much that I buy her perfume line for my wife. To me the only beginning and end exists when someone first encounters a brand and then when the person dies or the brand folds. In between is a lot going on. and it never ends.

    The problem with Social Media is it social. It needs people. And you can’t really automate people who refuse to act rationally all the time. I mean who needs a vacation instead of a truck to haul dirt?

    • says

      I’d agree and disagree, Howie. Campaigns (and the short durations you mention here) are a part of any longer term relationship between brand and customer. At some stage there has to be an introduction between the two, and the message that goes out will determine if the relationship flourishes or flounders after the “first date”. For me, that’s the campaign from marketing’s angle – making the introduction and introducing the customer to the next phase (customer relationship and brand management for the customer’s expectations).

      On that front, I think campaign is the ideal term because it is a short-lived thing. Yes, there can be ongoing marketing initiatives based around the same message or goal, but generally there are different campaigns for different goals – which brings us back to that short term approach.

      Cheers, mate.

  5. says

    The biggest challenge for such programs is the situational factors that surround the people involved in those conversations. Political, economic, religious, cultural, or geographic situations impact the meaning or context of the discourse. Theses are factors that are hard to glean from tweets regardless of how good the text analytics are. Analytics software, no matter how sophisticated, must allow for human-management of the data. There is no such thing as a turnkey solution in this space.

    Great article.

    • says

      Agreed, mate. As I mentioned to @40deuce:disqus, I’m intrigued by what the likes of NSM in Toronto is doing with nModes – although that still needs to be fine-tuned and ensure the automated response based on language and intent doesn’t miss the nuances of a slang statement, for example.

      While the technology may continue to improve, we’re always going to need that human gut instinct.

  6. says

    You know that scares me a bit. A software platform that can figure out from “It sucks to be cold” a bunch of possible opportunities and send me suggestions on how to solve that? What’s going to happen with the good ole banter that means nothing more than “it sucks to be cold”?

    BTW, you forgot one, “How about a nice trip to the Caribbean?” I figured that one out when you decided not to buy the truck. 😉

    I talk about booze a lot. This could be a problem……down the road.

    • says

      Hopefully, smart brands will simply use the advanced software to truly identify an opportunity and then craft the right approach from there, and continue to build a connection instead of just a collection.

      Don’t tease me with images of the Caribbean while we’re enjoying the Canadian winter still…

  7. Devon Ellington says

    It’s also a good venue to find out what businesses DON’T deliver. “Customer Service” is a dying art, as are the ethics and values of a company. If I feel their ethics are counter to mine, I take my business elsewhere.

    And when I’m targeted by someone who obviously picked up a single word and didn’t dig deeper into my posts and interests, that also tells me it’s someone I won’t do business with.

    All of this can be a two-way street!

    • says

      Perfect points, Devon. It;s why I’m excited about some of the new platforms coming through that ignore scores and concentrate on impact based on a brand’s goals (which includes churn and service). And we delve into that quite a bit in the book, and share a new model that highlights current technology flaws and where a brand really needs to be working when it comes to influence.

      The real conversations are just about to start.

  8. says

    Interesting to think about how a social media post can be perceived by a marketing company. And yet there’s so much that happens offline so how much can we really believe in the influence scoring? What if the whole vacation vs truck conversation happened over dinner, got carried into your wife’s group of friends and then discussed with her family on the phone? How does all that really get factored in because you hit the like button and made a comment on a FB fan page?

    • says

      And there’s the crux of the matter right there, Penney – the disruptor factors (situation, micro and macro influencers, etc) that algorithms can’t prepare for. The current social scoring leaders like Klout really offer such a miniscule picture of the influence sphere that it’s scary how the words “standard for influence” are used. There’s a far bigger conversation needing to take place about this, as well as a new model needing to be introduced.

      Consider that the battle cry in the lead-up to May… 😉

      • says

        I agree about a new model but I get concerned just how much of our privacy will we be losing just so marketing folks can get a more accurate assessment of how a social post influenced our buying decisions. How close are we going to get to Big Brother (the book, not the crappy TV show)?

        MAY??? Is that when the book comes out???

        • says

          I think the privacy angle needs to be addressed by more than just marketers – the technologies themselves need to respect privacy settings. This isn’t happening at the minute. So I’m completely with you on the need for privacy and non-invasive approaches.

          It might be sooner, that’s up to the publisher. We finish our draft next month, so technically it’s feasible to be sooner. All depends on edits and their list of publications and partnerships. We’ll be doing a lot of stuff beforehand, though – webinars, etc, where we’d love you to share your smarts too, miss!

          • says

            Yeah, I think we can talk all day about this privacy conversation but maybe we should try to solve world hunger first since that might be easier then getting our answers to the influence scoring vs those disruptor factors.

            That’s so cool about your new book! Keep us posted …
            And webinars?!?! We can share? Does that mean you’re going to open up the floor and let us all make comments and say really hot and techy stuff? I loves me a good webinar :)

  9. Brian Vickery says

    I've been excited as I've been introduced to the Influence Marketing methodology that you and Sam are developing. I'm just getting teasers, but posts like this – along with the graphic – definitely point to a compelling book once you boys get it finished!

    Since we *are* a provider of social media monitoring/sentiment analysis, I'm also interested in vetting our functionality against your methodology…and then working to fill the holes as best as possible. See, you and Sam might not know how to program…but we sure do!

    • Danny Brown says

      And this, sir, is why we need to talk – Sam and i are providing the new roadmap. The technology is there in various solutions – but if one could be built that encompasses the roadmap, or at least provides the core engine…. 😉

  10. says

    This is a great article Dan. It clearly explains the challenges, so far, with social monitoring platforms. What are your thoughts on the effects of social media itself on consumer purchase decisions and can monitoring platforms evolve ahead, or will they always lag behind consumer needs?

    • Danny Brown says

      Thanks, guys. With regards the social media effect, I think we'll always lag when it comes to monitoring, until there is a really strong solution that can decipher natural language and slang. That may well be a pipe dream, since humans and our emotions are very hard to decipher at the best of times! So, for me, it will still mean manual legwork in addition to whatever the technology can offer.

      Thanks for the comment, guys!