Business is having a love affair with big data. In the last few months it seems that every conference and webinar I come across has “big data” on the agenda.
Big data is the allure of more. More information. More access to behavior. Opportunities for more sophisticated analysis. The thinking goes likes this: If we know every move people make then we not only know everything, but we can predict everything. As business people we love it because the information is finite, scalable and measurable.
It’s ironic that big data has such an allure in an age when some of the most important information happening online is coming out of conversations. And the way to analyze online conversations is to read them, participate in them and try to understand them without a formula.
Conversations are not data. Anyone who has ever tried to analyze them through sentiment analysis knows this. If you ever watched Star Trek, you can imagine clearly that even an advanced, science fiction character like Data, misunderstood human conversation.
Computers don’t understand context, sarcasm, emotion and tone. They are likely to misunderstand a line like “I just love the way (insert brand name) waits 3 days before answering my email.” Formulas also can’t see subtle patterns and long term trends in the ways people respond to certain topics.
The fascination with big data is allowing marketers–particularly the bigger brands—to set aside the need to listen to and participate in conversations. They grapple with the concept of making meaning out of conversations since you can’t plug them into an algorithm and come out with useful guidance.
There is only one effective way to understand conversations, which are the most human form of information, and that is to listen to them and use a human thought process to analyze them. This may be blasphemy to the quants, but the human mind can intuit things that a computer can’t. There are some things that you just can’t take a measuring stick to.
Understanding conversations means you have to read Facebook comments and blog comments, often responding and then gauging the response.
Here are the objections I’ve heard from big brands: It doesn’t scale. We don’t have time for that.
The problem with this excuse is that you can’t really leave your Facebook page or blog on auto pilot. Someone trained and talented at qualitative analysis should be reading those conversations and hopefully responding when necessary and deleting spam. Nothing says “I’m not listening” like Facebook and blog comments filled with spam. The person reading those comments should be sharing what she learns weekly. Sometimes the learning comes from the content of the messages and sometimes it comes from recognizing that historically, certain types of posts get more comments than others. You’ll only understand why because you’ve read them.
The curious mind of a great community manager will ask herself questions to try to understand the conversation. How was the post worded to elicit the responses we received? How rich in content are the comments for different types of questions? What types of posts give us the best information about our readers?
One thing is true–understanding content on this level doesn’t “scale.” You have to put more time into it and use people that have the qualitative analytical skills to understand what’s happening. What doesn’t ring true is that there isn’t enough time or money. That statement reveals that it isn’t given importance and is too complicated or “fuzzy” to make an effort to understand.
If you don’t put the effort to understanding conversations, there is nothing genuine about them. Not only do you learn more by reading and responding, but when people know you’re listening, they share more.
On the Lion Brand Facebook page we delete spam comments and respond when its clear that our community needs information that isn’t available from others in the community.
What we learn by being part of the conversation
Why people do or do not like a pattern or product we shared. This helps us with product development. People aren’t robots and beyond hitting the “like” button, they’ll tell you why they like or do not like your offerings. This information is not the same as the information you get from doing surveys. It’s organic and self motivated.
We offer yarn, related products, and patterns to use them with. Every once in a while someone will suggest something different than what we are offering and when the community starts raising their hands and indicating that they’ve heard a great idea (not from us) we take notice.
The topics people bring up naturally, in response to our posts help us know what associations they have with subjects we bring up. After running a comic about knitting needles used as weapons, we discovered that people had been stopped from bringing their needles into courtrooms. This gave us an idea for a post about when and where people are stopped from doing their hobby—a very emotional topic.
Knowing our customer
You understand your customer on a whole different level when you are in conversation with them and care enough to learn to communicate better with them. We learned what topics turn our customers off (politics!), where they are uncomfortable with technology so we can respect their concerns by providing more information about how to access content, which venues are appropriate for edgier content and which will get you in trouble; the fact that we have really different kinds of people on different social platforms and how to speak to them in their own language.
Sometimes it’s not about numbers
We recently learned something from listening to our community that surprised us. We didn’t hear it from a lot of people but the fact that it came up a few times and the intensity with which it was said told us that we had to do something about a wrong impression people had. We intuited that it could be more widespread than we were hearing about. It’s the kind of information that I can’t share publicly but we do know just what to do about it and we would never have figured it out had we not paid attention to a few voices without counting the “ocurrences” of a word.
That’s the kind of information you’ll hear if you realize that humans do not communicate only in ways that can be analyzed by a machine.