Every day, someone asks me why they should pay for social listening technology. [Full disclosure: I server as a senior strategist for Converseon and Revealed Context.] Often, the answer is that they shouldn’t.
I mean, if all you want to do is detect a crisis leaping out of control, free tools can do that. If you are willing to sit a monkey in front of a keyboard to stare down the stream and extract what is important, free tools will work just fine.
You need to pay when that isn’t good enough.
Typically, I find the tipping point comes when you move past looking at individual tweets or blog posts and you suddenly want to aggregate the data. If you want to say that our positive sentiment is up 4% over last month, suddenly the free tools can’t cut the mustard.
Here’s why:
- You need to know the sentiment is correct. People do a good job of detecting sarcasm and otherwise interpreting whether something is positive or negative, but free tools don’t. Look for extensive machine learning and text analytics in the tool you select to be confident about its calculation of sentiment.
- You need to be fishing in the right lake. If all you do is type in a word or two and look at all the conversation that contains those words, it’s like drawing conclusions from a survey where you haven’t controlled the data sample. If you don’t know whether it is the right conversation, it doesn’t matter what the sentiment is. For example, if Sprint just looks at every tweet with that word in it, the conversation includes so many off-topic comments that the sentiment is meaningless. Look for technology that uses relevance feedback to isolate the right conversations before you aggregate the numbers.
If you are just looking at individual conversations to find trending stories or to pick out sales leads, free tools might be fine for you. But if you are doing market research, reputation management, campaign effectiveness, product development, or anything that requires you to draw conclusions about groups of conversations, you can’t skimp on the tool.
After all, how much less do you want to pay to get the wrong answer?
