“Context is king” – may very well sum up the problems with relying on some of the more popular social listening platforms, but it’s worth a deeper look to understand why they may create more work than they alleviate. Below we’ll discuss a few problems with so-called modern social listening platforms and why that human element is so necessary.
Dashboards are notorious for NOT actually including forums and message boards. (This omission is dangerous for building your assumptions in any industry – from consumer products to healthcare to automotive; there are message boards for everything; even B2B!) Yes, some can take a quick scan if you plug it in – which means you need to have a trained analyst FIND those forums first – but what they won’t do is take advantage of the various context clues. How long has a poster been a member of the message board? Who asks questions? Where in the decision journey are respondents when they answer?
Dashboards are also great at understanding sarcasm and ironic humor! That was sarcasm, BTW – something AI is still bad at; never mind that these dashboards don’t read emojiis or abbreviations well or at all. It turns out that human analysts are still the best at parsing behavior, but then that assumes that your investment in the tools includes someone to go back over the data, not just look at the pretty top lines.
For the record, working with AI can be very rewarding in social listening and is something I do advocate – if it comes at the right point in the process. Coding behavior first, to ensure accuracy and appropriateness (read: clean for trolls and spam) can result in an AI pass that is very helpful for detecting and categorizing emotions. We’ve experimented with Watson through agency partners that really understand the value of both behaviorists and the best use of AI. If you rely on AI on the front end, you’re putting an awful lot of trust in a bot to not only discover where your audiences live, but parse their casual language with peers well without assistance.
The money problem that comes with these problems are perhaps the most sneaky of the issues; we feel like we’re investing in something that is shortcutting an otherwise laborious process – it should be easy right, just let a bot scan chatter and tell us what’s going on? But as we’ve explored this week these dashboards are often missing whole channels of behavior and when they do look they often don’t understand the context, sarcasm, or vocabulary used.
Which begs the question: are most people hired to operate these dashboards busy acting on insights, re-analyzing, or having to spend time correcting? What are they actually trained to do? Are they acting or reporting up information that could be incomplete or without context? In our desire to feel efficient we have effectively turned over listening to something without ears. We wouldn’t trust a Ring doorbell that couldn’t distinguish between a bird and a UPS driver, why are we satisfied with dashboards that aren’t sure if a poop emoji is saying your product or service is crap, or “the shit?”
“Context is king” – even in how many dominoes bad or missing data can knock over. Or to use another analogy that came up in the conversations of this week: People want speed but assume that nothing is sacrificed in the process; Dashboards are more akin to fast food; if you’ve never had a gourmet meal, it can look appealing… until you realize what you’re missing. As a colleague said, “No dashboard in the world will tell you WHY something is happening; it only tells you THAT something is happening.”
A Dashboard investment includes the license, the FTEs assigned, and the accountability of the decisions made as a result of the analysis. With eyes wide open you can examine whether the time and money spent actually pulls reliable, helpful insights – which helps you create more successful solutions and communications. Knowing what you know now – would you let a dashboard assess a KPI for your strategic plan? It may be worth actually trying out a pilot deep dive to identify channels and behavior and let an actual behaviorist take a look – let a human being help recalibrate your insights and possibly your investments.
– Dean Browell, PhD
If you’re interested in learning more about your audiences, reach out and we can customize a research plan and approach that works for you: Dean @ discoverfeedback.com .