-
Website
http://www.kenburbary.com/ -
Original page
http://www.kenburbary.com/2009/07/the-dirty-little-secret-of-social-media-monitoring/ -
Subscribe
All Comments -
Community
-
-
Top Commenters
-
Popular Threads
All Comments
New Blog Post: The Dirty Little Secret of Social Media Monitoring - [link to post]
- Posted using Chat Catcher
RT @kenburbary: The Dirty Little Secret of Social Media Monitoring: [link to post] #socialmedia #PR
- Posted using Chat Catcher
RT @kenburbary The Dirty Little Secret of Social Media Monitoring | Web Business by Ken Burbary [link to post]
- Posted using Chat Catcher
Further, I'd argue there is always an inherent margin of error in monitoring. It almost doesn't matter what tool and terms you utilize. Google News, Factiva, Lexis-Nexis, Radian6, Techrigy, whatever. There are always posts/articles/tweets that will elude the spiders. Perhaps they will come in a day or two, but expecting 100% capture from any service is just unrealistic in my view.
I can't help but feel like I'm picking a cell phone provider/plan. Who has the best coverage for blogs? forums? twitter? etc... These could be very important details due to the strategies/actionable outcomes that are derived from the insights monitoring reveals.
Anyway, I look forward to the update from R6 and Techrigy.
RT @kenburbary: New Blog Post: The Dirty Little Secret of Social Media Monitoring - [link to post] [David]
- Posted using Chat Catcher
@davidalston @ambercadabra @jimmyrey @cbensen - Can you solve this monitoring riddle? [link to post]
- Posted using Chat Catcher
RT @kenburbary: New Blog Post: The Dirty Little Secret of Social Media Monitoring - [link to post] [they spider different sites]
- Posted using Chat Catcher
Great post. I think your experiment solidifies how important it is to not rely solely on only one form of social media monitoring. There are definitely differences between the main market leaders right now. The space is huge, and the technology, channels and discussion is constantly changing. To expect that one tool can get everything you need would be underestimating the space. When I structure a social media monitoring program, I couple Radian6 with several outside search tools (Google, BlogPulse and others) to be sure we’re picking up everything we possibly can and giving the client the most accurate, relevant information.
-Josh
RT @kenburbary: New Blog Post: The Dirty Little Secret of Social Media Monitoring - [link to post]
- Posted using Chat Catcher
Most measurement comps use the same data, Radian6 and Techrigy SM2 different [link to post] (via @kenburbary)
- Posted using Chat Catcher
RT @adamcohen: I'd like to see @AmberCadabra and @cbensen arm wrassle over @kenburbary's latest post [link to post]. In good fun | Me2!
- Posted using Chat Catcher
Interested to see the conversation that comes from this (from @KenBurbary): [link to post] (sketchy shortened URL, I know)
- Posted using Chat Catcher
Jim Reynolds AKA Jimmyrey
Director Of Sales
Techrigy
You could export the results for the blog posts from ea tool & compare them (if you have some time)
We have been asked this question & here are my answers:
1. our spam filter is different than other products. I doubt that there will ever be a 'standardized' filter set across tools
(when I was in library world there were various tools that we could use to meet the CIPA filtering requirements, but again, each accomplished the goal in a different way)
2. there isn't a standardized set of 'sources' either. Our method of collection is different from other tools. Our customers can add specific sources and use tools like Yahoo pipes to add the add'l sources that they require.
3. specific use of the tool - what filters are being applied in SM2? that may affect the data; what URL's are marked as spam by the user?
We devote a specific amt of dev't time to finding new sources. And we're continually evaluating what is spam or not. The social web is everchanging.
This helps better understand some of the specifics of how SM2 is handling the searches. That, coupled with the support I'm getting from @jimmyrey are painting a better picture that I will elaborate on in an update to this post after the dust is settled. Oh, and for the purposes of our search, we were not filtering any sources.
I'd like to see @AmberCadabra and @cbensen arm wrassle over @kenburbary's latest post [link to post]. In good fun. Yes I said wrassle.
- Posted using Chat Catcher
RT @adamcohen: I'd like to see @AmberCadabra and @cbensen arm wrassle over @kenburbary's latest post [link to post]. In good fun.
- Posted using Chat Catcher
The Dirty Little Secret of Social Media Monitoring - [link to post] (via @kenburbary)
- Posted using Chat Catcher
@kenburbary We'll definitely drop by and add a comment. Stay tuned...
- Posted using Chat Catcher
but what does it all mean?! @bivings RT @kenburbary: The Dirty Little Secret of Social Media Monitoring - [link to post]
- Posted using Chat Catcher
Terrific discussion started by @kenburbary on differences among results from various SM monitoring tools: [link to post]
- Posted using Chat Catcher
Interesting data RT @dariasteigman: Terrific disc. by @kenburbary on different results from various SM monitoring tools: [link to post]
- Posted using Chat Catcher
Why it's important to check how well your Social Media Monitoring tool is tracking different sources [link to post] (via @miketrap)
- Posted using Chat Catcher
I'm definitely checking back to see how each side responds. :)
Jenny
While there's no "perfect" sentiment analysis algorithm yet, there are some that are *much* better than counting emoticons! If you're going to market a tool with a claim of "Twitter sentiment analysis", I think you should disclose the algorithm.
As others have pointed out spam filters can play a major role in changing these numbers aswell.
What is key if you are looking at using a vendor for social media monitoring that you compare the results across several at a time to see what works best for you.
We would recommend you add us to the list of vendors to compare!
Thanks for this great subject matter!
Leon
www.sentimentmetrics.com
This has long been a problem in the media monitoring industry in general as no two companies approach it in the same way and it is nearly impossible for them to replicate each others data sets exactly. For the social media monitoring companies, they often struggle to identify where the influential and relevant "conversations" are taking place. There is also the issue with how much of any given site is spidered and as well as how often. These factors, amongst other things, can affect the end results in a profound way.
- Your application should determine how important content integrity, accuracy or representativeness are.... our clients use SM2 for everything from sales lead gen to qualitative market research, and many others
- All data/content have caveats, you need to understand your business intelligence sources caveats so you can make the most informed decision
- I have seen the best and most scientific data gathering conducted only to be twisted into whatever conclusion the analyst wants, it happens....
- If your business decision is a big one, use multiple sources of intell mixed in with your gut and experience
Techrigy has spent 3+ years understanding how to collect, store and help users understand SM content, it is a very complex technology challenge and it's still an early market. We have documentation, training vidoes and other tools to help our users best understand the content they gather. Happy to share these with you all.
Jim Schwab
VP Sales & Marketing
I think it's important for any marketer looking at purchasing any type of marketing systems -- whether monitoring, marketing automation, content management, customer relationship management, etc. -- to keep in mind.
I know as marketers we tend to want to focus on the bottom line, but the details matter. And this means we need to understand 1.) how different platforms approach the task they are focused on managing/automating/etc. and 2.) what the underlying data is that they are using.
We need to be smarter buyers of marketing technology ... and we need to think more strategically about leveraging and purchasing marketing technology.
This -- in my mind -- is as much about a 'call to arms' for marketers as it is about vendor capabilities. We need to balance our strategic and operating mindsets as marketers with a system mindset as we build out our systems for complex, integrated marketing management.
I know this is a topic that Scott Brinker has touched on via his Chief Marketing Technologist blog (http://www.chiefmartec.com/), and I think it's a great call-out here.
Great dialogue -- and great start at raising the tech IQ among us marketers.
The other comparison that is far more important is the total amount of meta-data collected for each result. This meta data enables the wide range of analysis that the tools offer. Sheer results are meaningless without the ability to discover demographic, sentiment, emotive, popularity/influence, etc. of those results.
1. The focus is on two of the many, many monitoring / analytic products and services out there. They all do things differently just like every responder to this post does things differently from their competitors. If each of you were given the same question by a client, you'd all likely have a different answer because you all do things differently to remain competitive.
2. No search engine (Social Media Monitoring or otherwise) returns the same responses nor holds the same data. Why didn't you also ask Microsoft, Yahoo and Google to respond?
The real issue is to know enough about your chosen tool provider(s) to ask the right questions to ensure you get the informationyou needed for your specific project. This is one key reason why the KDPaine's, Nielsen's, Cymfony's etc of the world attack the monitoring packages on a regular basis. In my opinion, too many people rely on the output of these systems without truly understanding their customer requirements, the system capabilties and to a certain extent, its shortcomings.
Trial and error? Research on this is limited at best right now, given how young the space is. We're all learning what works/what doesn't together.
Frankly, I doubt you will be able to get a bottom line answer as to what these differences are attributed to. As noted by a number of participants in this discussion, the same issue applies to web analytics tools and search engines. So, what we usually do is rely on not just one, but a number of (paid and free) tools for social media monitoring and each time we go over the results we are getting with respect to the specific goal at hand. The social media monitoring tools are of great help (and will be of even more help as these technologies mature), but their use still requires significant oversight to ensure useful analysis and effective action taking.
You make great points, that I concur with. As much as I would like to see less complexity in choosing/operationalizing a monitoring tool, the reality is that we need to rely on many sources. I'd enjoy a conversation with you to learn a bit more about the differences you've observed.
I'll be publishing the results, observations and thoughts about both tools in a new post soon.
Finding the most relevant and interesting "chatter" whilst weeding out the spam will always be different one provider to the next, which solution(s) can provide the information you need in a timely and clean fashion?
There is so much information available to us right now that filtering out the noise and finding the juicy bits is as big a challenge as getting complete coverage.
Zak
www.moreover.com
More importantly, both services allow you to import RSS feeds, so it is possible to combine the two into a de-duped, aggregate report (with some fussing with it). Personally, I like to have the final product come through Techrigy because they allow me to download the full content of all posts into a database that I can analyze on my own machine.
Plug R6 into SM2, then add in a well developed Yahoo! Pipes feed as well, set your Filters wisely, and you've got the industry's best listening platform....
Paul Dyer
http://www.dyersituations.com
Interesting article. Looking forward to hearing the updates, since my agency went through the same experience of inconsistent results. We had been using Radian6, but after seeing several presentations, including TNS Cymfony and WoolLabs Web Dig, we settled with the latter. I won't go into why we didn't choose the others (frankly, the differences made my head spin), but we did go with Wool's product because of the sheer depth of data. Real data. It isn't the prettiest GUI, but they said they are working on it. If you haven't heard of them, it is worth checking it out. We started with them for a 3 month trial, and have now signed on for an on-going engagement.
Good luck!
Randy