Over the past few weeks, a common thread in many of the more enthusiastic conversations about social media has been the role that people play compared with the technology.
These discussions have been focused on issues such as sentiment, particularly whether automated sentiment technology is accurate, and whether people can do a better job of assessing whether a conversation is positive, negative or neutral.
In particular, Jason Falls’ blog post on “Why You Shouldn’t Trusted Automated Sentiment Scoring” attracted a lot of attention.
There’s also been talk about social media monitoring and discovery versus curation done manually by people.
The emergence of these discussions is not a surprise. In fact, it’s a positive development because it reflects how the social media industry is evolving and maturing.
As social media monitoring usage becomes more widespread, there’s bound to more talk about the “what”, “who” and “why”, as well as the inevitable “man vs. machine” debate.
I think some of this discussion is motivated by the fact technology is taking over a role played by people for many years.
Before the Web and social media emerged to search and aggregate information, many companies used clipping services that used people to manually go through newspapers and magazines, as well as radio and television broadcasts. It was an intensive and expensive process.
Now, technology has pushed people into the background. Social media monitoring and analytics technology can handle a lot of the grunt work by quickly collecting, aggregating, processing and presenting millions of conversations – something that people are not capable of doing.
But – and here’s the big “but” – technology can’t completely replace people, and technology shouldn’t completely replace people.
Despite the advantage of social media monitoring and analytics services, people will continue to play a key role in taking the information collected, and then providing insight, advice, context and recommendations about what the data means and what should be done with it strategically and tactically.
For example, technology can make it easy to identify key influencers and opinion leaders but people need to step into the fray to engage with these people and develop relationships.
People can also play an active role in helping perfect the technology. For example, all the talk about automated sentiment fails to take into account that perfection can be elusive because of things such as sarcasm and nuance.
Still, I’d take a system any day that processes millions of conversations to offer information and intelligence about what’s happening within the social media landscape, AND gives provides the ability to easily adjust sentiment when needed.
The bottom line is that social media is a partnership between people and technology. They serve different roles and, in many ways, couldn’t operate as well without each other.
The “man vs. machine” discussions are healthy but it’s really a “man AND machine” world that we’re living in.
Tags: people, sentiment, Social Media, Technology
*Man* versus Machine?
What about Women?
You’re right, it should be people vs. machine.
As the singularity draws neigh (see R. Kurzweil) I think many, many more things will be replaced by technology. However, just like the rabbit trying to cross the finish line, you can break things down in halves and more halves as you edge closer to the finish line without really reaching it because there is always a half of a distance you must travel…
Good point — I wonder if there are stats on who uses which types of social media more — men or women — or if it’s about equal.
I agree that sentiment analysis has to be a combination of automated data gathering and human verification of the data found. Going just by even the best sentiment programs will give you misleading information. But for most social media work, the amount of data to be checked is just too high for any human to do the evaluation in real time.
What to do? Make sure you are using the best system available. Double check the findings. And make sure that you are measuring what you think you are measuring. Make sure that the terms you run the sentiment analysis on don’t have built in sentiment.
As for the men vs. women conversation in the comments: Does it matter?
Interesting conversation! Thanks for the idea
Saying that social media is a partnership between man and machine is too huge a leap.
My opinion is that ultimately, what we are seeking is to connect to one another using the shortest and quickest link possible — that’s why we will always seek to invent new ways to take care of the grunt work that gets in the way.
In the context of social media, technology will always be driven by the kind of interactions we wish to have with one another, and the intent behind them. For me, sentiment scoring isn’t a social action per se as it isn’t driven by a need for interaction but rather tries to qualitatively represent an event. There is no social intent behind it. The intent only resides in what I want to say and how I want to say it, when I want to say it. No machine can track this.
In the same way, curating content is an activity done by someone who wishes to create interaction with collected content. While a machine could collect the content if told how to do so, it couldn’t spontaneously collect it with a particular ‘intent’ in mind and thus give this collection a purpose.
If interaction is what drives ‘social technology’, I would say that it’s ultimately only about the partnership between people and the intent behind it, machine being just means to an end.
Eventually we won’t need any machines
Mark,
Thanks for the link to my post. I think that social media monitoring can definitely be accomplished by a machine to some degree. It currently works especially nice when you’re looking for specific and unique product names like Pepsi. But the nuances around language in general and word combinations specifically are too complicated to return great results without spending a lot of time programming a Boolean search string.
I’m upset that we aren’t there yet. But am also hoping that the machines never get so good that they become self aware…