Given that social analytics is an automated process, the assumption might be that accuracy is a given. But any social analyst will tell you there are several ways presumed accuracy is thrown for a loop. And it’s a really big deal.
Why “close enough” isn’t
Why does accuracy matter? Because you aren’t polling social audiences simply to get a glimpse of your own popularity. You’re looking for data to support every brand initiative planned for the upcoming quarter, year, etc. When you’re spending real time and real money on real strategies, you need real insights to justify that spend. And they’ve got to be precise.
It’s not enough to know sales are up, or engagement is falling, or sentiment is constantly fluctuating – you need to know why. Without that information you can’t fix anything or move forward.
Machines need to be more human
What skews accuracy? Humanity. Humans are sarcastic, snarky, tongue-in-cheeky, dry-witted, etc. That means their words often convey sentiment opposite to their literal interpretation. Taking words like love and hate and their various iterations at face value is enough to throw accuracy out the window.
It’s imperative to have social analytics tools that understand real human language, in all its complicated glory, if you want reliable insights.
What else throws results off track? Social conventions like emoji and GIFs, which don’t require text to convey their message. Users of Facebook Messenger send 5 billion emoji per day. That’s just one app.
Emoji are often used as emotional punctuation within posts, but some social posts contain only emoji, with no text. If all your data is based on keywords and keyword-related sentiment, your data won’t be accurate. You need to hear from all members of your audience in all forms to properly assess brand perception in the world.
When time is of the essence
Data analysts are great when you have complicated queries to run, but they shouldn’t be the only members of your team focused on accuracy. And you shouldn’t need to rely on them for it.
Social trends crop up quickly, as do consumer issues. Your social listening software must deliver spot-on insights in real time – because you don’t always have the luxury of having an analyst pore over your data before taking action.
And sometimes you need to know whether to act at all. Agency Camp + King were running the campaign for the Sacramento Kings new logo reveal when an early version of the logo was leaked. Fans reacted on social media, criticizing the new logo and lamenting the change.
Sometimes it’s important to respond to negative commentary on social media, but sometimes it’s best to leave things alone. This was one of those times – but Camp + King could not have known that without accurate information about what had Kings fans so upset. Fans didn’t love the black and white logo, and without context about the rebranding could not get excited about it.
Since the official logo reveal – in full color – along with the event celebrating the opening of the new Golden 1 Center would speak to both issues, the agency didn’t have to take to social media to defend their client.
Had they done so, they might have made things worse, dampening expected fan enthusiasm on the day of the grand opening. Instead, they used NetBase’s precision sentiment analysis tools to guide their decision making – and it paid off.
Don’t settle for inadequate social analytics tools
Social analytics tools are evolving to include the entire organization. Analysts still have their place, but accuracy is everyone’s domain. Whether you have a quick decision to make, or are planning for the future, you need insights you can count on. Be sure your tools are able to provide them.
We’d love to show you how NetBase finds precise insights with our patented NLP engine and other tools. Reach out for a customized demo!
Image from WOCinTech Chat