Algorithms are what makes the world go round, at least, that is what appears to be the conventional wisdom, and, which starts off my post today.  And you know what, there’s algorithms for everything, including Art and Incidental Comics shows how The Algorithm can help make it easier to create art.

In theory, using the right Algorithms should make life a lot easier, but often, companies and individuals choose the wrong Algorithms and make life a lot more complicated (and that’s where you end up with 10 people to change a light bulb - one to screw the bulb in and out, the other 9 to turn the ladder).

Still, what we should strive for is choosing the right platforms along with the right algorithms and the right people – that is a winning combination  (and what Social Media Analytics is all about …..).

Figured that too often, businesses that I have seen pick the wrong algorithms and platforms, often to have a reason to over hire or over charge clients, move the ladder around, almost robotically.   But I want to take this notion of algorithms further by suggesting they can be automatic/unconscious or consciously chosen, as discussed in a recent Marketing Pilgrim post by Joe Hall.

In this case, algorithms chosen by others can trap us, and that seems to be what Douglas Ruskhoff is saying in this post where many of our business decisions, are a result of the environment we operate in.  The implication is that, in many cases, organizations are innovating and seeking to grow in, what I take to be, the wrong way, instead of thinking about new products and services, they should be looking at changing the environment that created the dysfunctional results.

I believe, with the right algorithms, you don’t many people, just the right teammates, but with the wrong algo, no matter how many people you bring in, or how many systems are added, or how many processes are made up or forgotten, nothing really moves forward, or is sustainable, or even scalable.

There were also some good posts by MetricsMan on Social Media Listening along with the iCARD ECG, which is a pretty interesting device that could make it much easier to get a cardiogram in the near future.  Plus, the story about Neuromarketers Using MRIs to Predict the Next Teenie-Bopper Sensation and the Lytro self focusing camera, which seems like the next new thing,  a camera you can focus after the fact (the picture is taken) and any way you want.   The stories I read were great – but I found I had actually noted less of interest to me than usual – and often I have double this many stories to share for a busy two week period.

I am planning a more indepth review of Social Radar - but I just wanted to quickly throw out a few things here today. I did a query of Sade in “Kiss of Life” which is a song I’m listening to tonight.

But one of my favorite things about Social Radar is the Topics Breakdown (see below).

To be honest, I haven’t see anyone else come up with quite this type of breakdown – unless you want to count Crimson Hexagon, which does something sorta like this, but is an entirely different type of platform that is based mostly on machine learning, which Social Radar, is not.

Was tied up all this week on a variety of projects and simply didn’t have time to post – but at least my book is now being printed (and I’m grinning and smiling – the final manuscript came out very well and seems to encapsulate what I set out to accomplish, when I first decided to write Social Media Analytics almost a year ago).

Happy to see my Radian6 Influencer account just now has access to the Radian6 Insights Platform and the Radian6 Summary Dashboard (though the Radian6 iPhone/iPad app does not appear to be available at the iTunes Store yet).  I refrained from writing about the Insights Platform in detail until I had access to my own profiles in my own account I could play with – and now I do.

One of the topic profiles I set up in my Influencer Account at Radian6 is for my own blog, WebMetricsGuru.com; it had a tie in with Google Analytics Integration as well as topics gleamed from Google Webmastertools, Google Analytics Keyword Referrals and what I know about my own activities at conferences and upcoming book.   I first had to decide how much conversation to filter out (as my initial profile keywords used the keywords that drove traffic to my site, but may have also included conversations that had nothing to do with me).

The first diagram above is a word map, much improved from the original Radian6 word cloud showing the prevailing conversations around my profile.  I will point out ]word clouds, in general, may only use a subset of verbatim to build a map (certainly, Giles Palmer, CEO of Brandwatch thinks this is the case – building a word cloud of several thousand verbatim may be too processor intensive, Giles felt, and I agree – but we don’t really know how many verbatim are being used.  Crimson Hexagon, in building it’s word diagrams, only used the first 250 or so, verbatim, so my guess is Giles is spot on here).

I don’t have access to all the current Radian6 Insights, just Basic Demographics, and Radian6 Insights – the rest are additional charges and I would have to ask for that to be turned on to test Klout, OpenAmplify and OpenCalais Insights against my own profiles.   However, even the Free Insights bring Radian6 up to about the same capabilities as Sysomos in terms of Demographics, which had been missing up till now.

The Word Cloud is one of my favorite part of the Insights Platform – as it can produce some pretty interesting results.   I looked at the domains mentioned in the “Media” part of the Word Cloud that I then did another word cloud on the first result, t.co.

It’s more meaningful to me to see the word clouds because I know that they are about – whereas, exploring a subject that I am just learning about, it would be much harder to figure out what it all means.  That’s a basic issue with Social Media Analytics, which is why I advocated for people with deep industry knowledge doing listening reports rather than a team on interchangeable analysts – often the typical practice in most agencies.

Also did a breakdown of all the twitter hastags found in “Media” part of the word cloud above.  Tried the “% Change” feature but I don’t have the data in the profile to fully support comparing the last 3 months to the preceding – so it doesn’t tell me much.  However, by digging into the Gender of those who speak of “Media” in the original word cloud I got a decent breakout of 86 insights around male/female (shown below).  However, the total number of verbatim or mentions for the word cloud around “media” was 3705, meaning Radian6 Insights had data for 2.5% of it.  I bet Sysomos is about the same here with a pretty small sample, but there are some who think 2.5% is too small a sample to make meaningful insights around.  What you’ll probably end up with is using the Insights to get an idea of what the audience is for a topic, without assurances the information is actually accurate, or not – as the state of these platforms, while improving, is still not good enough for basic market research yet.  But if you just want to get some information to base your theories off of, or your working for a MarCOM firm and just want to tell a story, this data is probably good enough.

And how good is the gender determination? All of the River Of News for “Male” (66) was Male – I should know, I was in about a third of them, plus most others who I also personally know, and therefore knew the results to be accurate.  On the other hand, of the 20 verbatim that were said to be female, 2 were not, or 10% were Male, which I could probably live with.  In fact, 90% accuracy isn’t bad at all, all things considered.

FEMALE RESULTS:

I think the problem may be that people have an expectation provide Market Research capabilities; it’s clear to me why that is – Market Research is something people will pay a lot of money for.

One more insight that worked for me was “Sources” which are the actual domains or websites where conversations originate on.

Almost half the insights here are coming from my own domain, my own blog posts from the last 3 months is what that is made up with.  I suppose sources are among the most useful insights as it revels what is feeding a particular message in the word cloud, and that can be very helpful.  However, recall we’re only getting 97 insights around sources and we’ve got 3705 verbatim or mentions, meaning Radian6 Insights gave us sources on only 3% of the content  – probably not enough to draw any conclusions on -which is not enough to go to the bank with.

Then again, Location data is even thinner and derived all from Twitter, as far as I can tell, and fell  under 2% (don’t laugh, Sysomos isn’t much better here) with 55 verbatim out of 3705 total.   If you had a couple of million verbatim, 2% might be enough to do some basic modeling around, but with a couple of thousand records, it’s not enough.  But how accurate is the results for location?  The 17 verbatim for New York City are all me, so that’s correct.  However, I’m often in Providence, RI, would it also know to pick me up there or not?   Well, I that wasn’t an option I had a chance to test with this data set.

Usernames are also interesting, and of the 3705 verbatim Radian6 was able to find the usernames of 85 Twitter mentions, about 2%, see above.   I would say this is as close to “influencers” as I can get without buying the Klout module. It’s also interesting that most of this information is just textual mining in the verbatim (largely available on other platforms such as Sysomos and possibly, Social Radar, however, the data is gathered from Insights partners, and has the potential to be layered open in a way the other platforms can not).

Retweets are another interesting report – nothing revolutionary here but Radian6 provides the Insight, free of charge.

But the Re-tweets can be segmented by another dimension, Location, though it’s not clear to me the use of it.

But none of this tells a story – and honestly, it helps to know what the story is that you want to tell before pulling any data.  I’m reminded of what it’s like to go to Art School, there are plenty of pretty tubes of paint and drawing utensils I played with and saw others devour, but tubes of paint don’t make a story, and i didn’t have one here, either.

And it gets back to what I wrote about in Social Media Analytics – that you need to be an industry expert in whatever is being monitored for, otherwise it’s hard to understand the data your looking at.

I guess that’s it for tonight – I could have written about the Summary Dashboard, but that will be left for another day.