For me, the slightly more pressing issue is
"how do you convince a client of the value of engaging in a social media campaign"
- particularly a client who isn't familiar with social media, [or one that thinks Facebook will lead to your website getting hacked.]
I guess it's the same kinda thing as measurement, but usually put in a more blunt way i.e.
"how does social media help me sell more stuff?"
I reckon if you can answer this question in 30 seconds with geeking-out massively, then you are onto a winner, and are a better man than I am. [When I say 'answer', I also mean prove].
This is as far as I've got:
social media = engagement = favourability = likelihood to buy
It's not scientific and at the moment I can't prove the correlation. But, I'm getting a bit farther with the middle bits [engagement = favourability].
The tool I've been having most success with is Radian 6. Using simple boolean search [like Google] it allows you to measure stuff like:
- Volume of buzz - how many posts in different media from blogs to forums and MSM [mainstream media] via what they call a River of News.
- Engagement level - by measuring on topic posts and comments.
- Buzz topics - what's being talked about in relation to your brand.
The nifty thing about Radian 6 is it allows you create powerpoint friendly images of the analysis. Here's one I prepared earlier that shows the volume of buzz around Asda and three search terms:
Plus, you can produce word clouds of all the things being said about a brand:
I guess the only draw back is that it isn't great at filtering by country. If you set it to pick up only UK ones, then it just picks up .co.uk domains and UK based IP addresses, which excludes lots of UK blogs and forums.
Still, it allows you to show that your client is being talked about and what's being said. Plus, you can click through on each post to get a favourability score. All of which allows you to talk sensibly about a brand's level of engagement.
I'm sure I've more to learn but Radian 6 seems a good starter for ten.
Posts
11 comments:
I've been using (testing) Radian6 for about 6 months now. My experience is that it is a good start for tracking conversations and looking at traafic for certain key words.
However, if you delve into the results, you'll find a lot of unrelated information. Another problem is that bloggers are very savvy and can create 'fake blogs' that get picked up too. So a lot of manual work is needed to filter out the rubbish.
I've found Radian6 to be a good tool to conduct online perception audits or track buzz - but it doesn't really 'measure' social media as such.
I do like your formula and look forward to see how you justify each element...
Robin and Yeelim, thanks for the feedback on the platform. What I've found to work to reduce fake blogs is tweaking at the topic profile configuration level. Gradually over time the ones that slip through can get trapped out but either using the blog URL or unique words associated with those potential conversations. It's something to consider at least.
As for the UK only filtering we also try to assess the location based on the "about me" section of a blog, Twitter etc... Again, not always perfect but it's another way to pull out the UK ones. As you pointed out unfortunately it's not an exact science at this point.
We've added a few more enhancements this past weekend that you may want to try out to help delve into who the posting individual is and what the conversation is surrounding, say, a spike on a topic trend graph is all about. Based on the formula, trying to build relationships is key when going down the path of determining an ROI for investing time in social media activities and both these should hoping help in making connections more efficiently and making the right connections.
Great discussion guys and know that all of the feedback is always much appreciated as we continue to evolve the platform and add new functionality.
BTW, we just put out a youtube video on the new features this morning here - http://www.youtube.com/watch?v=9YJ9tOu2EqY - if you are interested.
Yeelim and David - thanks for the comments. Really useful.
Yeelim - my experiences are similar to yours, although I haven't come across the fake blogs yet. I get good results for UK specific brands, but mixed ones for international ones as you can't easily filter out non UK posts.
David - thanks for the tips on tightening up the search profile, will give it a whirl.
I've been using Radian6 for a while too - found it's the best of the bunch. And getting better with each iteration...
That, and a combination of the free tools available (I think you pointed me to quarkbase) tends to do the job....
Robin, great piece, we'll have a natter about this tomorrow I am sure!
david, I saw your response and was pleased, although not surprised that you picked up on the conversation - if you guys aren't monitoring your brand name, who else will?! Full marks on that one!
I saw some of the new features this week and think that it is a great feature to be able to quickly recognise different media.
Different sources of mentions is no doubt useful, but to be able to distinguish the media from within these sources is especially effective, particularly given the full multimedia nature of social media.
Have to agree with Robin and Yeelim on the location thing - very frustrating, but i and think i understand enough of how the systems work to know why!
The one MAJOR downside to the toolset is what seems to be the manual marking of sentiment - something which the Live Buzz tool I use frequently and conduct most of my client work with measures this automatically and, after a "four-week training period" is 80% accurate.
Volume of mentions, comments, unique comments etc. are no doubt incredibly useful, but without knowing the sentiment of those comments, or being able to aggregate them into general satisfied, dissatisfied categories etc, is a major weakness of the radian offering.
That said, don't get me wrong, as a quick and easy way to manage buzz monitoring, it couldn't be better, and makes a big difference, but there are still a few things missing that the other system has.
Thanks for getting the conversation going on this Robin.
Chris - Thanks for the encouragement. With each release each month we try to further fine tune things as well as offer new and useful features to take monitoring and engagement to a new level - based on customer feedback.
Paul - thanks for the kudos on listening to our brand. I definitely love to be able to use our own platform to stay in the loop and to reach out to those in the community.
And thanks for the vote for sentiment scoring. We've so far avoided implementing it because of the accuracy issues of most of the engines out there that we've seen. Not to say we haven't been looking into it and studying what may work though. Your experience with the 80% scoring sounds encouraging for sure.
Keep the feedback coming. It's what's gotten us to this point and what drives us to keep putting out new features. Thanks guys.
What I've found is that most clients use things Google Alerts or paid for media monitoring tools to track presence in MSM, but it's accurate tracking buzz in blogs and forums that is what is in demand. The more accurate we can make the search, particularly location and relevance, the more successful campaigns are. It's unfortunate that a few inaccurate or irrelevant results can make clients dismiss the whole analysis. The more advanced tend to understand that it's a learning process.
I'm due to attend another training course soon to be updated on the latest enhancements.
I think it will be hard for Radian6 folk to develop a system of 100% accuracy on relevance and location given the nature of social media. I understand the challenges, but it's hard to explain this to clients who may know a little less about how the internet ticks.
I often tout Radian6 as one of the most sophisticated systems available - though not perfect - yet!
Right now, I use a combination of Radian6, other free online tools, and some manual work to present a better picture.
Yeelim
I agree - its going to be difficult to get 100% accuracy on relevance and I reckon we'll also need some human input. I just wish there was a way of confidently filtering by country. One day maybe:)
looks like I need to check out Radian 6 having invested loads of time looking at different solutions and with a few clients eager to get some monitoring set up. Great discussion here btw
Thanks Jonathan
Would be good to get your thoughts on the whole thing, Radian 6 in particular. Feel free to throw in any other systems we all should be looking at [sorry David :)]
Post a Comment