I just finished testing two low cost (<$50.00/mo) social media monitoring tools. They were both easy to use with clean user interfaces. Both had some nice features, especially around reaching back out to people making comments in social media. However, running these two services side by side brought home some issues with these kinds of offerings – specifically in the area of coverage, sentiment analysis, and analytics. Note that I am not naming names because I believe the issues I ran into are not unique to these specific tools, but to the low cost social media monitoring market, in general. Some of these issues will also apply to higher priced offerings!
What I did:
I ran an analysis using the term “advanced analytics” including the term itself as well as companies (as additional topics) in the space. I made sure to be as specific and clear as I could be, since I knew topics were keyword based. I let the services run side by side for several weeks, interacting with the tools on a regular basis.
What I noticed:
1. Topic specification. Tools will vary in how the end user can input what he or she is looking for. Some will let you try to refine your keyword search (and these are keyword based, they won’t let you discover topics per se), other won’t. Some will only allow search across all languages, others will allow the user to specify the language. Find out how you can be sure that you are getting what you are searching for. For example, does the tool allow you be very specific about words that should and should not be included (i.e. SAS is the name of an analytics company and also an airline)? Since these low cost tools often don’t provide a way to build a thesaurus, you need to be careful.
2. Completeness of Coverage. The coverage varied tremendously between the services. Nor was the coverage the same for the same day for the same company name that I was tracking (and I was pretty specific, although see #1 above). In fact it seemed to vary by at least an order of magnitude. I even compared this manually in twitter streams. When I asked, I was told by one company that if they weren’t picking up everything, it must be a bug and it should be reported (!?). The other company told me all of my content came to me in one big fire hose, because there had been a problem with it, before (!?). In both cases, there still seemed to be a problem with the completeness of content. The amount of content just didn’t add up between the two services. In fact, one company told me that since I was on a trial, I wasn’t getting all of the content – yet even with the firehose effect, the numbers didn’t make sense. Oh. And don’t forget to ask if the service can pull in message boards, and which message boards (i.e. public vs. private). For an analysis, all of these content issues might mean that completeness of buzz might be misrepresented which can lead to problems.
3. Duplicate Handling. What about the amount of buzz? I thought that part of my content counting discrepancy might be due to how the company was dealing with duplicates. So beware. Some companies count duplicates (such as retweets) as buzz and some do not. However, be sure to ask when duplicate content is considered duplicate and when it is not. One company told me that retweets are not counted in buzz, but are included in the tag cloud (!?).
4. Sentiment analysis. The reality is that most of the low cost tools are not that good at analyzing sentiment. Even though the company will tell you they are 75% accurate the reality is more like 50%. Case in point: on one offering, one job listing was rated positive and another job posting listed as negative. In looking at the two postings, it wasn’t clear why (shouldn’t a job post be neutral anyway) Note, however, that many of these tools provide a means to change the sentiment from +/-/neutral (if they don’t then don’t buy it). So, if sentiment is a big deal to you then be prepared to wade through the content and change sentiment, if need be. Also ask how the company does sentiment analysis and find out at what level it does the analysis (article, sentence, phrase)
5. Analysis. Be prepared to ask a lot of questions about how the company is doing its analysis. For example, sometimes I could not map the total buzz to other analytics numbers (was it duplicates handling or something else). Additionally, some social media monitoring tools will break down buzz by gender. How is this determined? Some companies determine gender based on a name algorithm, while others use profile information from facebook or twitter (obviously not a complete view of buzz by gender, since not all information sources are twitter and facebook like). Additionally, some of the tools will only show a percentage (a no-no in this case), while others may show the number and the percent. Ditto with geolocation information. If the data is incomplete (and isn’t representative of the whole) then there could be a problem with using it for analytical purposes.
What this means
Certainly, the lure of low cost social media platforms is strong, especially for small and medium businesses. However, I would caution people to do their homework and ask the right questions, before purchasing even a low cost product. I would also suggest testing a few products (running them side by side for the same time period, even if you have to pay for it for a month or so) to compare the tools in terms of coverage, sentiment, and analysis.
The reality is that you can end up with an analysis that is completely wrong if you don’t ask the right questions of the service provider. The amount of buzz might not be what you think it is, how your company compares to another company might be wrong based on how you specified the company name, the sentiment might be entirely wrong if you don’t check it, and the analysis may be misleading unless you understand how it was put together.
4 thoughts on “Five basic questions to ask before leaping into low cost social media monitoring”
Viralheat is a <$50 tool that offers a great blend of price, features and ease of use. It also offers sentiment analysis, using natural language processing (NLP), compared to keyword analysis that some vendors use. I go over the benefits during my bi-weekly webinar. Join me on 1/27: http://bit.ly/fL6eI5.
Great post, Fern. I’m Gene from ViralHeat and we are a “low-cost” SMM service. Our plans run at $9.99/29.99/89.99 and we offer pretty in-depth anaytics and monitoring tools across Twitter, Facebook, Buzz, YouTube, real-time webb (incl blogs and forums), and others. Give us a peek and let me know if you have any questions!
You review is awesome. You touched main areas which needs to addressed by an social media monitoring tools.
I would take this as opportunity to introduce Factualz here. I would be happy if you play around with Factualz and send feedbacks. We value your feedback.
Factualz provides live social media dashboards with actionable insights for marketing and competitive analysis. No software to download, configuration or setup. Subscribe to Factualz over internet and access it from anywhere. Signup now for a 2 weeks free trial at http://www.factualz.com
• Listen to all the conversation on any brand and competition – With our unique “business context aware” engine, Factualz gives you 95% relevant buzz from 16 media types, including news, blogs, videos, bookmarks, and micro blogs.
• Comprehensively follow all the social marketing channels of your competition including Facebook, Twitter, YouTube and Google Buzz.
• Automatic sentiment, relevancy and reach analysis of all content with respect to brand’s business context
• Automatic tagging, trend analysis and identification of top domains to channelize your web marketing and PPC programs
• 20+ customizable dashboards and reports to compare and benchmark with competitors
• Profiles of all the people who interacted with the brand and their social ranking and demographic analysis with auto classification of fans and critics
• Collaborate with your team to analyze content, add tasks and route them for closures
• Discover new business and marketing opportunities by using customizable topic and keyword segmentation
• Identify new business and marketing opportunities
• Measure your social and web marketing effectiveness and ROI
• Improve your web marketing initiatives such as PPC, targeted ads, SEO
• Comprehensively track and understand your competitors marketing campaigns and strategy
• Keep track of your brand’s online reputation
For more information please visit http://www.factualz.com.
[…] Filed under: advanced analytics, Business Analytics, data analysis, social media, Text Analytics, text mining | Tagged: Fern Halper, keyword search, low-cost social media tools, monitoring buzz, social media buzz, social media monitoring | 3 Comments » […]