social media monitoring problems

The Problem with Social Media Monitoring Tools

social media monitoring problems

Social media monitoring is a critical aspect to any well thought-out social strategy implementation. Many practitioners often question whether or not today’s tools justify the spend and, even more importantly, what tools should be used for their company’s specific situation. The honest truth that most vendors don’t want to talk about is that most tools fall short on delivering what they promise. From my experience testing and using 20 + different tools over the course of the last 3 years, this is simply the case.

I did find, however, that in some instances, robust, six- figure/month social monitoring suites can be the tool that saves brands thousands of hours a month in manual labor hours sifting through conversations. Surprisingly though, this has been a rarity.

At the end of the day, you are a marketer, a communications or PR practitioner. You should ask yourself – what has this tool given me that is actionable? That is the point many bloated software suites so far fail to deliver on.

More often than not the tools themselves give you fancy-dancy GUI charts that look pretty for the CEO but fall short on delivering actionable takeaways. The best data that can be derived from social media monitoring comes with the active day-to-day listening. I am talking about reading every single post, every single Tweet and every single mention of your company, or at least making a concerted effort at doing so.

The drill-up capabilities, the word clouds, and the automated brand health analysis reports tend to be more of a distraction than actually helpful. Instead of spending time on those reports you should take time to understand the conversation with actual effort by reading through the conversations. It will take many man hours of doing so before you can adequately help your customers or garner intelligence to funnel back into your product life cycle.

Tip: Look for a tool that facilitates easy reading, browsing and collaboration (think CRM) of  insightful conversations your audience might be having. Pay less attention to the reports.

My biggest gripes with the current marketplace consist mainly of the following issues:

Over Priced Solutions:

The pricing models behind many of the popular, robust tools simply do not play nice unless you are Wal-Mart and have seven-figure budgets per year.

The alternative is to utilize a less expensive DIY tool which can run you around $100 month, and incidentally there are not many of them catering to this price point. Most businesses do not get the kind of volume that warrants a robust monitoring tool. For many, I recommend using a free solution such as RSSOwl or spinning off your own free dashboard for most situations.

Spam Filtering (or a lack there of):

You would expect when you pay $30,000 a year on a tool that you wouldn’t find yourself still sifting through MFA sites, Amazon affiliate clones and other irrelevant blog/website mentions for your companies name, right? However, with almost every tool I have tested, spam is rampant. Many times spam can become such a problem that you have to alter your Boolean variables in such a way that you end up missing over half of the conversations about your brand.

It seems that writing a piece of logic into your system that would actually work would end up looking something like this:

If site has Amazon affiliate AND > 2 AdSense blocks AND > 4 keywords in a domain THEN = Spam.

I have even witnessed one social media monitoring firm (that shall remain nameless) state that their tool is “Spam Free.” What is funny is their tool is one that produces quite a bit of spam in the data set when compared to other monitoring suites.

Automated Sentiment Analysis:

Monitoring solution providers have pushed this technology onto the marketplace.

Automated sentiment analysis is something many of the higher-end solutions state they handle. Many practitioners who are looking to gauge brand health quickly tend to lean towards these sorts of automated reports. From my experience they are usually wildly inaccurate.

Many of the tools claim around 70-80% accuracy. That simply isn’t the case whenever I peruse the mentions to cross check those sentiment metrics. I, for one, question the act of declaring something Positive, Negative or Neutral in the first place. Sentiment is subject to various interpretation, making it hard to train  a robot to decipher a long blog post that might have multiple tonalities to it.

The other problem I see with many of these automated sentiment solutions is that users who end up relying on the scores tend to lose sight of the purpose in the first place: reading what your customers actually say. If you automate sentiment, you are missing all the golden insights that might be derived from your customers.

Tip: The qualitative feeling around a keyword or topic is much more meaningful when you do your due diligence and read people’s opinions on the matter. Stating whether something is 76% neutral, 30% negative and 4% positive based off of a robot’s interpretation of it really is futile.

Social Monitoring Tools Need To Improve

The idea of using a tool to monitor conversations about your company on the web is not entirely new, however the underlying technology to monitor sources such as Facebook, Twitter and other media outlets is still in its infancy. We have to remember that many of these tools are less than 3 years old.

I would like to see monitoring tools put more effort behind fighting spam, being more open about what their products can actually do and being more transparent about how their technology works. Many practitioners simply don’t have the technical know-how to understand how things work, or don’t, and they are easily sold solutions that are not good fits for their needs.

Bringing to light the real conversations and delivering qualitative reports that highlight actual impactful data is something I really would like to see as this space evolves.

Have you had frustration with a monitoring tool? If so, I would love to hear your story.

Update: Seth Grimes has pointed me in the direction via his comment below of a great follow up read to this post he wrote: “What I Look For In A Social Analysis Tool”

Related Posts

57 Comments
  • Andy Beal
    Posted at 17:56h, 21 June

    I agree with just about everything you say Brian.

    You are much better off taking the $30-100k you could spend on high-end tools and use that money to hire someone to actually dig in, read, and react to the conversations. As you know, Trackur targets the sub-$100 a month market and users that don’t expect eye-candy to make their decisions for them.

    I, for one, agree that all vendors need to do better. There is not a day that goes by that I don’t question what more we could be doing for our users.

    Cheers!

    Andy Beal
    CEO, Trackur

  • Joe Hall
    Posted at 18:00h, 21 June

    Awesome post. I think the reason these problems, and others exist is because social media in large part has all the same complexities of the web + a bunch of other things thrown in.

    Folks always say that search is an engineering problem that strives to cover huge goals while managing scarce resources. I think social media monitoring is a lot like this. It requires access to data, (which at volume, requires strong business/relationship development), processing power to develop timely metrics, and massive storage (you are basically building your own index). I am sure there are other technical hurdles that I am missing, but these seem huge. And because of that, it really does require a vendor that can charge a 5 to 6 figure fee to keep it going…..because to do it right, is no small feet…..you are right though many aren’t doing it right.

  • Joe Hall
    Posted at 18:00h, 21 June

    Awesome post. I think the reason these problems, and others exist is because social media in large part has all the same complexities of the web + a bunch of other things thrown in.

    Folks always say that search is an engineering problem that strives to cover huge goals while managing scarce resources. I think social media monitoring is a lot like this. It requires access to data, (which at volume, requires strong business/relationship development), processing power to develop timely metrics, and massive storage (you are basically building your own index). I am sure there are other technical hurdles that I am missing, but these seem huge. And because of that, it really does require a vendor that can charge a 5 to 6 figure fee to keep it going…..because to do it right, is no small feet…..you are right though many aren’t doing it right.

  • Anonymous
    Posted at 18:22h, 21 June

    Thanks Joe,

    The technical ramifications needed to deliver a strong social monitoring solution is indeed very difficult. I for one do not take lightly the developers jobs it is to crawl, index and de-spam the webs mentions. Especially across social sources which can be finicky at best, with downtime, changing of APIs and de-throttling of accounts etc.

    With that said, however, I just get a gut feeling from my experiences with the various platforms, the people running them don’t take most of the things I mentioned into priority. Most of them focus on customer acquisition mainly b/c the market has been a crap shoot where folks are just trying to acquire as many new contracts as possible, looking for any means to reach an Exit.

    A lot of Venture capital is floating around in the market and I truly feel the people paying the price are the actual customers.

    Compare the Web Analytics field to Social Media Monitoring, I think they are very similar in the technical specifications needed to run and scale both businesses, they are difficult and cumbersome. Work with a firm like Coremetrics, Webtrends, Omniture or any of the other robust analytics platforms then compare that to the monitoring space. You will see the difference in how the two industries are run rather quickly.

  • BrianChappell
    Posted at 18:22h, 21 June

    Thanks Joe,

    The technical ramifications needed to deliver a strong social monitoring solution is indeed very difficult. I for one do not take lightly the developers jobs it is to crawl, index and de-spam the webs mentions. Especially across social sources which can be finicky at best, with downtime, changing of APIs and de-throttling of accounts etc.

    With that said, however, I just get a gut feeling from my experiences with the various platforms, the people running them don’t take most of the things I mentioned into priority. Most of them focus on customer acquisition mainly b/c the market has been a crap shoot where folks are just trying to acquire as many new contracts as possible, looking for any means to reach an Exit.

    A lot of Venture capital is floating around in the market and I truly feel the people paying the price are the actual customers.

    Compare the Web Analytics field to Social Media Monitoring, I think they are very similar in the technical specifications needed to run and scale both businesses, they are difficult and cumbersome. Work with a firm like Coremetrics, Webtrends, Omniture or any of the other robust analytics platforms then compare that to the monitoring space. You will see the difference in how the two industries are run rather quickly.

  • Jeff Kryger
    Posted at 20:20h, 21 June

    I agree with this post, especially the spam comment. I spend every day going through about 100+ new query results, and deleting about 85 of them. That being said, I feel that eventually certain tools will rise to the top (as there are literally dozens of options available now) and become the standard for SMM in the coming years, as the algorithms for search and the tools specifically improve

  • Jeff Kryger
    Posted at 20:20h, 21 June

    I agree with this post, especially the spam comment. I spend every day going through about 100+ new query results, and deleting about 85 of them. That being said, I feel that eventually certain tools will rise to the top (as there are literally dozens of options available now) and become the standard for SMM in the coming years, as the algorithms for search and the tools specifically improve

  • Technically Social
    Posted at 00:34h, 22 June

    Now that you put it this way, I agree. For huge companies and brands using these tools will have all the limitations you’ve talked about. My niche is small businesses, and typically the results are not quite as numerous, so these problems aren’t as apparent. 

    I’ll agree with Jeff, in that these tools are still adapting and a few will eventually rise to the top and resolve the issues that you’ve mentioned. I won’t ever agree with a company using bells and whistles to cover up sub-par results (if that is truly what’s happening) but hopefully companies using this ‘polish the turd’ technique will improve their products instead of improving their appearance. 

  • Ingenu, LLC
    Posted at 00:34h, 22 June

    Now that you put it this way, I agree. For huge companies and brands using these tools will have all the limitations you’ve talked about. My niche is small businesses, and typically the results are not quite as numerous, so these problems aren’t as apparent. 

    I’ll agree with Jeff, in that these tools are still adapting and a few will eventually rise to the top and resolve the issues that you’ve mentioned. I won’t ever agree with a company using bells and whistles to cover up sub-par results (if that is truly what’s happening) but hopefully companies using this ‘polish the turd’ technique will improve their products instead of improving their appearance. 

  • ingrid rubin
    Posted at 06:47h, 22 June

    HI Brian
    You make some valid points regarding what you need from an ORM tool, in essence the tool is a means to an end to derive data and information that allows you to make strategic decisions and provides you with an understanding or the market-place environment and the overall sentiment of your Brand out there.
    You will always need human intervention to support in making the strategic analysis and decisions, a tool will never do this for you.
    In regards to your comment around sentiment and the quality score, 65-70% for auto-sentimenting is all that you are going to get as no matter how advanced the filtering is on the applications, a system will never be able to pick up sentiment and total article tone, so will often get it wrong.
    I would love you to try out our product http://www.saidWot.com, we have been in the market for 3 years. Are well priced and have had much success. We have a simple user-interface which is intuitive. 
    Regards
    Ingrid Rubin 
    @ingridr   @saidwotorm:disqus 

  • ingrid rubin
    Posted at 06:47h, 22 June

    HI Brian
    You make some valid points regarding what you need from an ORM tool, in essence the tool is a means to an end to derive data and information that allows you to make strategic decisions and provides you with an understanding or the market-place environment and the overall sentiment of your Brand out there.
    You will always need human intervention to support in making the strategic analysis and decisions, a tool will never do this for you.
    In regards to your comment around sentiment and the quality score, 65-70% for auto-sentimenting is all that you are going to get as no matter how advanced the filtering is on the applications, a system will never be able to pick up sentiment and total article tone, so will often get it wrong.
    I would love you to try out our product http://www.saidWot.com, we have been in the market for 3 years. Are well priced and have had much success. We have a simple user-interface which is intuitive. 
    Regards
    Ingrid Rubin 
    @ingridr   @saidwotorm:disqus 

  • Elias Shams
    Posted at 08:25h, 22 June

    Brian, Excellent post. It’s time for such a hot
    space to enter into a new category.

    There is a need for a portal to provide a quick and
    intelligent decision for both the consumer and the enterprise about their
    online connections. A Platform to Help us to Distinguish Our Quality vs.
    Quantity Friends, Fans, Followers, and Companies.

    Facebook, Twitter, LinkedIn, Youtube, Flickr and others
    have been doing a decent job of providing additional marketing exposure and
    even in some cases, additional revenue. However, as more and more social
    networking sites pop up, the task of picking the right channel getting more
    challenging.

    This reminds me of the early 90’s when WWW was adopted
    broadly by the general public. Every company rushed to have a presence, to the
    point it became literally impossible to find the right information on the Web.
    That’s when yahoo and google where born and helped us find the most relevant
    information by just typing simple keywords.

    Then came Web 1.0 & 2.0 – Youtube, Flickr, myspace,
    Facebook, Twitter and countless others have turned everyday people into content
    producers, influencers and experts. We basically tripled down on the
    information overload How do you know which channels to select for deploying
    your social media strategy?    

    That is why I built awesomize.me to accomplish such a
    mission – the portal to all your existing social media channels.

    Elias
    CEO & Founder
    http://awesomize.me

  • Elias Shams
    Posted at 08:25h, 22 June

    Brian, Excellent post. It’s time for such a hot
    space to enter into a new category.

    There is a need for a portal to provide a quick and
    intelligent decision for both the consumer and the enterprise about their
    online connections. A Platform to Help us to Distinguish Our Quality vs.
    Quantity Friends, Fans, Followers, and Companies.

    Facebook, Twitter, LinkedIn, Youtube, Flickr and others
    have been doing a decent job of providing additional marketing exposure and
    even in some cases, additional revenue. However, as more and more social
    networking sites pop up, the task of picking the right channel getting more
    challenging.

    This reminds me of the early 90’s when WWW was adopted
    broadly by the general public. Every company rushed to have a presence, to the
    point it became literally impossible to find the right information on the Web.
    That’s when yahoo and google where born and helped us find the most relevant
    information by just typing simple keywords.

    Then came Web 1.0 & 2.0 – Youtube, Flickr, myspace,
    Facebook, Twitter and countless others have turned everyday people into content
    producers, influencers and experts. We basically tripled down on the
    information overload How do you know which channels to select for deploying
    your social media strategy?    

    That is why I built awesomize.me to accomplish such a
    mission – the portal to all your existing social media channels.

    Elias
    CEO & Founder
    http://awesomize.me

  • Anonymous
    Posted at 11:59h, 22 June

    Great post! and you are so dead on.. But not only  is insight and analysis lacking, I find too many organizations pick the tool first before they even know what their social media goals are. What is needed is a clear consensus around goals, then an anlaysis of what data is available, THEN an decision about the tools. Too often I go into organizatoins that are using five different monitoring tools and not getting any acitonable data out of any of them

  • KDPaine
    Posted at 11:59h, 22 June

    Great post! and you are so dead on.. But not only  is insight and analysis lacking, I find too many organizations pick the tool first before they even know what their social media goals are. What is needed is a clear consensus around goals, then an anlaysis of what data is available, THEN an decision about the tools. Too often I go into organizatoins that are using five different monitoring tools and not getting any acitonable data out of any of them

  • James Ainsworth
    Posted at 14:21h, 22 June

    Interesting thoughts and clearly based on experience of a lot of varying premium solutions. I dont think there is a vendor that isnt striving to do better, social media monitoring is no different to any other business in this respect. 

    There is a strong case for the use of a premium tool as a time saver but not only for its efficiencies, but the practicality of feeding the social media data back into every relevant part of the business is something that is tricky to put a value on. The ability to work with vast data sets to make better informed business decisions is a real strength of social media monitoring.

    I dont think it benefits a vendor to take a sales approach that doesnt take into account the need of the user, it doesnt make business sense to do that and it may be that the need and the tool dont always match up. You have to be honest in these circumstances.

    James Ainsworth
    Community Manager – Alterian
    @AlterianJames:twitter 

  • AdamSinger
    Posted at 17:11h, 22 June

    Yes most solutions are way overpriced. TBH a simple solution like Trackur or even Google Alerts is plenty for most brands. If you are digitally savvy monitoring the web is not that hard. Not sure why everyone freaks out about it.

  • AdamSinger
    Posted at 17:11h, 22 June

    Yes most solutions are way overpriced. TBH a simple solution like Trackur or even Google Alerts is plenty for most brands. If you are digitally savvy monitoring the web is not that hard. Not sure why everyone freaks out about it.

  • Anonymous
    Posted at 18:03h, 22 June

    5 different monitoring tools sounds like a very lethargic – we get nothing done in a day – approach. We have seen the same scenario several times as well Katie. Some companies just simply hate hiring folks and would rather buy into a piece(s) of software that does the work for them.

  • BrianChappell
    Posted at 18:03h, 22 June

    5 different monitoring tools sounds like a very lethargic – we get nothing done in a day – approach. We have seen the same scenario several times as well Katie. Some companies just simply hate hiring folks and would rather buy into a piece(s) of software that does the work for them.

  • Andy Beal
    Posted at 19:28h, 22 June

    I totally agree about the need for goals. In fact, I’ve outlined 6 steps before even paying for a tool: http://www.toprankblog.com/2010/09/6-steps-social-media-monitoring/ (Brian, please feel free to delete that link, if you feel inappropriate).

  • Andy Beal
    Posted at 19:28h, 22 June

    I totally agree about the need for goals. In fact, I’ve outlined 6 steps before even paying for a tool: http://www.toprankblog.com/2010/09/6-steps-social-media-monitoring/ (Brian, please feel free to delete that link, if you feel inappropriate).

  • Michelle C
    Posted at 09:38h, 23 June

    Nice post, Brian. I’d say that the huge problem with social media monitoring tools is not with the tools, but what people expect of them. They haven’t yet advanced to the point where it’s plug and play – you plug it in and you get actionable insights out. 
    There must be some point of human analysis, for sentiment analysis and also to catch spam where our filters haven’t. As a disclosure I work with Synthesio, that is a high-end tool that combines service and technology, so that our clients do always get insights that they can use to move forward out of what we’re listening to. 
    Also, with regards to the percentage of accuracy for automatic social media monitoring, the fact is that no one has yet tested all automatic tools on the same set of data to see how they compare on an even playing field. Since we monitor in 30 languages, we simply haven’t yet found a solution better than native language analysts for understanding sentiment and business information on a granular level, meaning reports that highlight and summarize actionable data. 
    I’d be happy to put you in touch with someone at Synthesio if you’d like to take a look. In the meantime, I’d say spend your money hiring someone to analyze information, as Andy suggests, or choose a tool that offers qualitative and quantitative customized reports! 

    Best,
    Michelle @Synthesio:twitter 

  • Seth Grimes
    Posted at 09:57h, 23 June

    Brian, there are strong insights throughout your article. I’ll comment on the area I know best, sentiment analysis.  Your observations are valid, but they *also* stem from tool deficiencies.

    On two points —

    1) “Sentiment is subject to various interpretation, making it hard to
    train  a robot to decipher a long blog post that might have multiple
    tonalities to it.”

    That’s a fault of the tools, not the concept.  Better tools will resolve sentiment at the “feature” level, not just score a blog, article, or tweet as a whole.  Feature-level resolution means distinguishing each of the entities (persons, companies), topics (“hotel room” “room service” “location”), and concepts (“cleanliness” “value”) — and further, separating out comparative mentions of two different brands.

    I wrote about this stuff and much more, earlier this year, in an article, What I Look For In A Social Analysis Tool, http://www.informationweek.com/news/software/bi/229301079 , and we’ll discuss it at a conference I’m organizing, the Sentiment Analysis Symposium, November 9 in San Francisco, http://sentimentsymposium.com

    2) “The other problem I see with many of these automated sentiment
    solutions is that users who end up relying on the scores tend to lose
    sight of the purpose in the first place: reading what your customers
    actually say. If you automate sentiment, you are missing all the golden
    insights that might be derived from your customers.”

    Shallow social monitoring tools that paint an aggregate picture are misused by organizations that should, instead, be seeking more-capable customer-experience management tools from vendors such as Allegiance, Attensity, Clarabridge, and Medallia.

    Seth, http://twitter.com/sethgrimes

  • Seth Grimes
    Posted at 09:57h, 23 June

    Brian, there are strong insights throughout your article. I’ll comment on the area I know best, sentiment analysis.  Your observations are valid, but they *also* stem from tool deficiencies.

    On two points —

    1) “Sentiment is subject to various interpretation, making it hard to
    train  a robot to decipher a long blog post that might have multiple
    tonalities to it.”

    That’s a fault of the tools, not the concept.  Better tools will resolve sentiment at the “feature” level, not just score a blog, article, or tweet as a whole.  Feature-level resolution means distinguishing each of the entities (persons, companies), topics (“hotel room” “room service” “location”), and concepts (“cleanliness” “value”) — and further, separating out comparative mentions of two different brands.

    I wrote about this stuff and much more, earlier this year, in an article, What I Look For In A Social Analysis Tool, http://www.informationweek.com/news/software/bi/229301079 , and we’ll discuss it at a conference I’m organizing, the Sentiment Analysis Symposium, November 9 in San Francisco, http://sentimentsymposium.com

    2) “The other problem I see with many of these automated sentiment
    solutions is that users who end up relying on the scores tend to lose
    sight of the purpose in the first place: reading what your customers
    actually say. If you automate sentiment, you are missing all the golden
    insights that might be derived from your customers.”

    Shallow social monitoring tools that paint an aggregate picture are misused by organizations that should, instead, be seeking more-capable customer-experience management tools from vendors such as Allegiance, Attensity, Clarabridge, and Medallia.

    Seth, http://twitter.com/sethgrimes

  • leeodden
    Posted at 17:55h, 23 June

    I think that’s a fine link Andy. 🙂

  • Lee Odden
    Posted at 17:55h, 23 June

    I think that’s a fine link Andy. 🙂

  • leeodden
    Posted at 17:56h, 23 June

    Trackur makes sense to start with but Google Alerts is crap – in my opinion.

  • Lee Odden
    Posted at 17:56h, 23 June

    Trackur makes sense to start with but Google Alerts is crap – in my opinion.

  • Anonymous
    Posted at 18:06h, 23 June

    “the fact is that no one has yet tested all automatic tools on the same set of data to see how they compare on an even playing field.”

    Here is a lengthy analysis comparing many of the top monitoring tools. Very insightful IMO.
    http://shared.freshminds.co.uk/smm10/whitepaper.pdf

  • BrianChappell
    Posted at 18:06h, 23 June

    “the fact is that no one has yet tested all automatic tools on the same set of data to see how they compare on an even playing field.”

    Here is a lengthy analysis comparing many of the top monitoring tools. Very insightful IMO.
    http://shared.freshminds.co.uk/smm10/whitepaper.pdf

  • Jim Durbin
    Posted at 18:20h, 23 June

    Good post, but I’d like to point out one small thing. 

    Buying tools is sexy, and makes you a lot of money.  Hiring people is hard, both in getting the invoice and setting up payment.  Managers have three different kinds of buckets – personnel, projects, and capital.  Hiring a consultant comes out of projects, while tech can come out of projects or capital.  Add in the cost of a consultant, and you then get HR, the VMS if you have one, and the issue of what to do if the consultant leaves or doesn’t work out. 

    This means consultants can get 5K here, 10K here, but when you get to 60k, and above, it’s signature time at the level above you.  Quite frankly, technology vendors are set up to sell to the budget process better than consultants.  Which means expensive technologies make more sense than a consultant or a cheap technology.  

    I was working with megacorp several years ago, and was surprised when they told me they had a 20,000 a month tool for monitoring.  No one actually looked at the results, but they had it, and spent the money on it, and that was in someone’s budget, which was protected because it was a long-term contract.  The “reasoning ” behind it didn’t matter, the budget process did. 

    As I struggled in early years (2006, 2007) to sell social consulting, it seemed the big deals were all in tools.  Hiring an individual for a project didn’t fit into how they spent money.  They preferred either the Ogilvy/Edelman/Fleishman approach, or a tool they could bring on.  It wasn’t until I adapted to the budget process and clearly defined my offering that my sales took off, and that was in no small part because big agencies were losing contracts to boutique firms during the recession.

    Vendors know this.  It’s a lot easier to spend money on technology than people, and a lot safer for the career of the manager. It’s also less hassle.  Tools you buy can sit around unused.  Consultants never give you that luxury.  

    While we can agree that these tools should do all you ask and more, the reality is they adapt to the market they can sell to.  There are plenty of tools that do just as you ask.  They have no market share, and can’t get footholds into the major companies.  It’s the same just about everywhere, with every technology.     

     

  • Anonymous
    Posted at 18:31h, 23 June

    Thanks for sharing your thoughts Seth. I am glad someone is taking a stab at organizing a conversation around “Sentiment Analysis” RE: the conference you are organizing. Many of the conferences I have spoken at personally and attended have not really addressed the subject matter.

    I also added your article to my original post. Great read.

  • BrianChappell
    Posted at 18:31h, 23 June

    Thanks for sharing your thoughts Seth. I am glad someone is taking a stab at organizing a conversation around “Sentiment Analysis” RE: the conference you are organizing. Many of the conferences I have spoken at personally and attended have not really addressed the subject matter.

    I also added your article to my original post. Great read.

  • AdamSinger
    Posted at 18:44h, 23 June

    You’re right Lee — if you are a big brand it is unmanagle with Alerts or even Trackur and a solution like Vocus is necessary. For a small scrappy brands or just website/blog owner with a unique name, Google Alerts (with some operator queries) works OK. Not perfect, but OK that you won’t miss the important stuff.

  • AdamSinger
    Posted at 18:44h, 23 June

    You’re right Lee — if you are a big brand it is unmanagle with Alerts or even Trackur and a solution like Vocus is necessary. For a small scrappy brands or just website/blog owner with a unique name, Google Alerts (with some operator queries) works OK. Not perfect, but OK that you won’t miss the important stuff.

  • Reginald Jackson
    Posted at 07:18h, 24 June

    Thanks for that information. I’m one of those guys that’s bad about getting any monitoring tool I can get my hands on if it doesn’t cost too much.

  • Reginald Jackson
    Posted at 07:18h, 24 June

    Thanks for that information. I’m one of those guys that’s bad about getting any monitoring tool I can get my hands on if it doesn’t cost too much.

  • 40deuce
    Posted at 20:51h, 24 June

    Great post Brian.
    I won’t argue one bit that there is always room for improvement in this area, and I can say that for my company this is something we’re always striving to do. We’re constantly trying to improve everything we do to be of more use and better for all of our clients.

    There’s two points in here that I would argue with though:
    First, I do think that being able to read through all posts will show the most information to gain insights from, but sometimes there’s just not time or manpower to do that. This is where some of the tools we provide really come in handy. I know that when I’m analyzing large amounts of data at one time it’s quite handy to use some text analytic tools. Some of the tools that Sysomos and other companies provide allow you to see the main points of conversation that reoccur throughout large sets of data so that you can get a good understanding of what kinds of conversation are going on without having to read through hundreds or even thousands (depending on your brand) individually. If something really stands out you can then go inspect those posts more closely as a smaller set of data.
    The second point I’d argue is over sentiment. I won’t even dare to argue that automated sentiment is 100% percent accurate. However, again when dealing with a large set of data, these tools can prove to be quite handy. While they won’t give you a 100% accurate reading of sentiment, it can help you to gain a good idea of how the conversation is generally going.

    Mostly what I’m saying is that these tools aren’t for everyone and every brand, but if you are one that needs to deal with large sets of data all at the same time, these tools can come in quite handy. I do think that humans should be going over and looking at all the data they get from tools to be absolutely certain, but some of these tools (my competitors included) are very good with helping people to sort and go through all this data.

    Cheers,
    Sheldon, community manager for Sysomos

  • 40deuce
    Posted at 20:51h, 24 June

    Great post Brian.
    I won’t argue one bit that there is always room for improvement in this area, and I can say that for my company this is something we’re always striving to do. We’re constantly trying to improve everything we do to be of more use and better for all of our clients.

    There’s two points in here that I would argue with though:
    First, I do think that being able to read through all posts will show the most information to gain insights from, but sometimes there’s just not time or manpower to do that. This is where some of the tools we provide really come in handy. I know that when I’m analyzing large amounts of data at one time it’s quite handy to use some text analytic tools. Some of the tools that Sysomos and other companies provide allow you to see the main points of conversation that reoccur throughout large sets of data so that you can get a good understanding of what kinds of conversation are going on without having to read through hundreds or even thousands (depending on your brand) individually. If something really stands out you can then go inspect those posts more closely as a smaller set of data.
    The second point I’d argue is over sentiment. I won’t even dare to argue that automated sentiment is 100% percent accurate. However, again when dealing with a large set of data, these tools can prove to be quite handy. While they won’t give you a 100% accurate reading of sentiment, it can help you to gain a good idea of how the conversation is generally going.

    Mostly what I’m saying is that these tools aren’t for everyone and every brand, but if you are one that needs to deal with large sets of data all at the same time, these tools can come in quite handy. I do think that humans should be going over and looking at all the data they get from tools to be absolutely certain, but some of these tools (my competitors included) are very good with helping people to sort and go through all this data.

    Cheers,
    Sheldon, community manager for Sysomos

  • Gabrielle Jeans
    Posted at 20:59h, 24 June

    I agree with you Andy. Companies are better off hiring a Social Media Manager than automate everything on Social Media because that defeats the purpose of socializing

  • Real Estate Trainer
    Posted at 20:59h, 24 June

    I agree with you Andy. Companies are better off hiring a Social Media Manager than automate everything on Social Media because that defeats the purpose of socializing

  • Robert Madison
    Posted at 21:52h, 24 June

    Brian, I welcome you to contact me at your convenience.  I work for Spiral 16, and bias aside, think we have some things to offer that you might be interested in.  Especially concerning spam filtering & automated sentiment analysis.

    For $30K/yr, hell…we’ll find Jimmy Hoffa! 😉

    robert.madison@spiral16:twitter .com

  • Robert Madison
    Posted at 21:52h, 24 June

    Brian, I welcome you to contact me at your convenience.  I work for Spiral 16, and bias aside, think we have some things to offer that you might be interested in.  Especially concerning spam filtering & automated sentiment analysis.

    For $30K/yr, hell…we’ll find Jimmy Hoffa! 😉

    robert.madison@spiral16:twitter .com

  • Robert Madison
    Posted at 22:02h, 24 June

    BTW, good post!

  • Robert Madison
    Posted at 22:02h, 24 June

    BTW, good post!

  • Joe Tierney
    Posted at 15:37h, 07 July

    Brian,

    This statement is totally nuts:
     
    “I am talking about reading every single post, every single Tweet and every single mention of your company, or at least making a concerted effort at doing so.” 

    Your two tips are also horrible. No wonder you’re seeing problems. If you’re looking to read each post you don’t need any application, there are plenty of search engines available for free. The science of natural language processing isn’t going away and its development is driven by market demands not vice versa. 

    Reading every single post is the exact type of work these tools should alleviate. Every comment, every tweet, every mention is not created equal and there is definitely a point of diminishing returns once you reach a particular volume of content. Similar to traditional market research we are searching for meaning and for understanding – this is the whole point of listening in the first place. The power of these applications comes in their ability to aggregate terabytes of data for on-demand analysis. I’m not sure which platforms cost 6-figures per month but that is WAY overpriced. 

    I work with clients to help them leverage Social Radar. Customers typically fall into 3 categories. 

    1. Social Radar is a poor fit for their use case for a variety of reasons. We flat out tell them this as soon as it becomes evident which usually occurs in the first several conversations. 

    2. Social Radar is a good fit but they spend all there time “reading every single post, every single Tweet and every single mention of your company” and “pay less attention to the reports”. They treat the application as an expensive search engine and fail to realize a strong ROI. They are unable to incorporate the analysis into their current or potential offerings to customers. Rather than focusing on the meaning of the data (the field), they spend their time fixated on individual posts (the grass). If this company was working on traditional market research and analysis they would be fixed on statements like, “I have 10,000 survey forms” or “5 people didn’t answer question 6” rather than “we conducted our survey, this is what we found, this is what it means, this is what we should do and this is how we’re going to do it.” 

    3. They understand their objective is to set, implement, support and continuously tweak strategy and operations either for themselves or their customers. They look for new opportunities or they justify their current direction. They see ROI figures well in excess of 1000%, they’re winning new clients as well as expanding their projects within their current clients and most importantly they’re hiring. 

    Leveraging these applications is not easy. It takes time and teams do have to work through many frustrations in process, implementation and inherent short comings of market analysis in general. 

    Like any tool, the real value comes from how you use it. 

  • Joe Tierney
    Posted at 15:37h, 07 July

    Brian,

    This statement is totally nuts:
     
    “I am talking about reading every single post, every single Tweet and every single mention of your company, or at least making a concerted effort at doing so.” 

    Your two tips are also horrible. No wonder you’re seeing problems. If you’re looking to read each post you don’t need any application, there are plenty of search engines available for free. The science of natural language processing isn’t going away and its development is driven by market demands not vice versa. 

    Reading every single post is the exact type of work these tools should alleviate. Every comment, every tweet, every mention is not created equal and there is definitely a point of diminishing returns once you reach a particular volume of content. Similar to traditional market research we are searching for meaning and for understanding – this is the whole point of listening in the first place. The power of these applications comes in their ability to aggregate terabytes of data for on-demand analysis. I’m not sure which platforms cost 6-figures per month but that is WAY overpriced. 

    I work with clients to help them leverage Social Radar. Customers typically fall into 3 categories. 

    1. Social Radar is a poor fit for their use case for a variety of reasons. We flat out tell them this as soon as it becomes evident which usually occurs in the first several conversations. 

    2. Social Radar is a good fit but they spend all there time “reading every single post, every single Tweet and every single mention of your company” and “pay less attention to the reports”. They treat the application as an expensive search engine and fail to realize a strong ROI. They are unable to incorporate the analysis into their current or potential offerings to customers. Rather than focusing on the meaning of the data (the field), they spend their time fixated on individual posts (the grass). If this company was working on traditional market research and analysis they would be fixed on statements like, “I have 10,000 survey forms” or “5 people didn’t answer question 6” rather than “we conducted our survey, this is what we found, this is what it means, this is what we should do and this is how we’re going to do it.” 

    3. They understand their objective is to set, implement, support and continuously tweak strategy and operations either for themselves or their customers. They look for new opportunities or they justify their current direction. They see ROI figures well in excess of 1000%, they’re winning new clients as well as expanding their projects within their current clients and most importantly they’re hiring. 

    Leveraging these applications is not easy. It takes time and teams do have to work through many frustrations in process, implementation and inherent short comings of market analysis in general. 

    Like any tool, the real value comes from how you use it. 

  • Joe Tierney
    Posted at 15:37h, 07 July

    Brian,

    This statement is totally nuts:
     
    “I am talking about reading every single post, every single Tweet and every single mention of your company, or at least making a concerted effort at doing so.” 

    Your two tips are also horrible. No wonder you’re seeing problems. If you’re looking to read each post you don’t need any application, there are plenty of search engines available for free. The science of natural language processing isn’t going away and its development is driven by market demands not vice versa. 

    Reading every single post is the exact type of work these tools should alleviate. Every comment, every tweet, every mention is not created equal and there is definitely a point of diminishing returns once you reach a particular volume of content. Similar to traditional market research we are searching for meaning and for understanding – this is the whole point of listening in the first place. The power of these applications comes in their ability to aggregate terabytes of data for on-demand analysis. I’m not sure which platforms cost 6-figures per month but that is WAY overpriced. 

    I work with clients to help them leverage Social Radar. Customers typically fall into 3 categories. 

    1. Social Radar is a poor fit for their use case for a variety of reasons. We flat out tell them this as soon as it becomes evident which usually occurs in the first several conversations. 

    2. Social Radar is a good fit but they spend all there time “reading every single post, every single Tweet and every single mention of your company” and “pay less attention to the reports”. They treat the application as an expensive search engine and fail to realize a strong ROI. They are unable to incorporate the analysis into their current or potential offerings to customers. Rather than focusing on the meaning of the data (the field), they spend their time fixated on individual posts (the grass). If this company was working on traditional market research and analysis they would be fixed on statements like, “I have 10,000 survey forms” or “5 people didn’t answer question 6” rather than “we conducted our survey, this is what we found, this is what it means, this is what we should do and this is how we’re going to do it.” 

    3. They understand their objective is to set, implement, support and continuously tweak strategy and operations either for themselves or their customers. They look for new opportunities or they justify their current direction. They see ROI figures well in excess of 1000%, they’re winning new clients as well as expanding their projects within their current clients and most importantly they’re hiring. 

    Leveraging these applications is not easy. It takes time and teams do have to work through many frustrations in process, implementation and inherent short comings of market analysis in general. 

    Like any tool, the real value comes from how you use it. 

  • Anonymous
    Posted at 19:46h, 07 July

    Joe,

    Thanks for the comment. I take it you work at Social Radar? 

    I chatted with your president several year(s) ago. It was one of the first tools we looked at. It was in fact pretty cool I thought, could allow you to deep dive into the data, it was fast and easy to use, but the big problem, once again, was the pricing model. Way to steep for most folks and only fit the Walmart-esk, in-house, situation.

    Also I wouldnt necessarily call your product offering a “social media monitoring tool’ its more a social analysis tool. Two completely different use cases so please don’t confuse my analysis above with ‘social media analysis’ tools.

    I still stand behind my point with regards to Social media Monitoring tools. Get out their and read what people are saying and stop relying on these super expensive tools to give you reports with answers. 

  • BrianChappell
    Posted at 19:46h, 07 July

    Joe,

    Thanks for the comment. I take it you work at Social Radar? 

    I chatted with your president several year(s) ago. It was one of the first tools we looked at. It was in fact pretty cool I thought, could allow you to deep dive into the data, it was fast and easy to use, but the big problem, once again, was the pricing model. Way to steep for most folks and only fit the Walmart-esk, in-house, situation.

    Also I wouldnt necessarily call your product offering a “social media monitoring tool’ its more a social analysis tool. Two completely different use cases so please don’t confuse my analysis above with ‘social media analysis’ tools.

    I still stand behind my point with regards to Social media Monitoring tools. Get out their and read what people are saying and stop relying on these super expensive tools to give you reports with answers. 

  • Stuart Davidson
    Posted at 11:28h, 13 July

    Wow, very nice article and on a topic I haven’t read much into. Some interesting points raised and I definitely agree that social media monitoring tools need to improve. I also like the methods you describe in how you would like to see them improved too. Good job!

  • Stuart Davidson
    Posted at 11:28h, 13 July

    Wow, very nice article and on a topic I haven’t read much into. Some interesting points raised and I definitely agree that social media monitoring tools need to improve. I also like the methods you describe in how you would like to see them improved too. Good job!

  • Anonymous
    Posted at 12:29h, 12 August

    It is right that instead of using tools all should write the fresh data and do conversation and provide information which can be very useful. Use of tools is not bad thing but do not use same as it but little update that tool and create new idea and with that use them so this way you can get the best result.
    manual social bookmarking

  • Internet Marketing Company
    Posted at 12:29h, 12 August

    It is right that instead of using tools all should write the fresh data and do conversation and provide information which can be very useful. Use of tools is not bad thing but do not use same as it but little update that tool and create new idea and with that use them so this way you can get the best result.
    manual social bookmarking

  • Sergei Dolukhanov
    Posted at 16:22h, 06 September

    Get’em Brian! BAM. You tell them how social media monitoring isn’t enough, and that cool graphs and charts can only get you so far. Finally, tell them how social media monitoring simply doesn’t address one of the most important questions for any enterprise user; ‘how is this tool going to correlate social data with key business performance metrics? How is this tool going to increase or show change in my quarterly revenue? Where is the actionable insight?

    It just gets to a point where ‘social listening’ alone isn’t enough; companies need actionable metrics to tie to their existing business processes. That’s where social (media) business intelligence comes in; social + business intelligence = IMPACT. It’s the crossroads of social data and your business.

    Thanks for the post man. Keep it up.

    – Sergei Dolukhanov@sdolukhanov:twitter 

  • Sergei Dolukhanov
    Posted at 16:22h, 06 September

    Get’em Brian! BAM. You tell them how social media monitoring isn’t enough, and that cool graphs and charts can only get you so far. Finally, tell them how social media monitoring simply doesn’t address one of the most important questions for any enterprise user; ‘how is this tool going to correlate social data with key business performance metrics? How is this tool going to increase or show change in my quarterly revenue? Where is the actionable insight?

    It just gets to a point where ‘social listening’ alone isn’t enough; companies need actionable metrics to tie to their existing business processes. That’s where social (media) business intelligence comes in; social + business intelligence = IMPACT. It’s the crossroads of social data and your business.

    Thanks for the post man. Keep it up.

    – Sergei Dolukhanov@sdolukhanov:twitter 

Post A Comment