7 Deadly Sins of Sentiment Analysis

When it comes to monitoring and measuring social media there are plenty of tools to choose from ranging from the entirely free to the incredibly pricey. Generally, you will get what you pay for on the lower end of the spectrum, but the value really tends to vary once you cross over the $500 per month threshold.

As firehose data access becomes increasingly commoditized, it is the depth of analysis that truly differentiates these solutions. Terms like natural language processing (NLP), text analytics, and artificial intelligence (AI) have become all the rage. However, it is sentiment analysis that has received the most hype and it is frightening how often it is abused.

The following will examine some critical mistakes to avoid when analyzing sentiment…

LUST: Sentiment analysis can be extremely tempting to use, but any good analyst will recognize that there are plenty of occasions where it can be highly irrelevant. Understand that you must put everything in context. It may be extra sexy to crank out a fancy visualization breaking down positive, negative, and neutral data. However, there is a time and place when sentiment can be valuable. Recognizing this will require self-control.

VANITY: Whether you’re a well-established Fortune 50 juggernaut or a promising lean startup, being arrogant can have severe consequences. Just because your sentiment may be trending positively for a designated period of time doesn’t mean you are doing everything right. It is easy to become overly confident and fail to take existing threats seriously. Having your own desired sentiment metrics that are clearly aligned with your business and strategic objectives will allow you to stay on track, the more specific to your business the better.

ENVY:  Sentiment analysis can be particularly useful when examining your competitive landscape. Seeing how you stack up against your competition is helpful to an extent. However, becoming obsessively jealous of such benchmarking can cause you to drift away from the predefined goals of your campaigns. The leading players in your market may have more appealing metrics when it comes to sentiment, but this may simply be due to the fact that you have different strategic objectives.

WRATH: At a high level, sentiment analysis allows you to identify individuals that feel negatively about some aspect of your business. This will range in severity and often times your business is indisputably at fault. Keep in mind that many of these individuals are simply seeking attention or looking for an altercation. Regardless, the worst thing you can do is fuel the fire by being aggressive or angry. Make a solid effort to show overwhelming compassion and kindness while attempting to move the conversation to a more private channel in a timely fashion.

GREED: Sentiment analysis can offer actionable insights for multiple business units other than just public relations and marketing. This is especially true with both the product development and customer service departments. Being stingy with your findings and failing to dispense relevant insights across your entire business is a missed opportunity. Having a proper workflow seamlessly integrated into your analysis will allow your business to make the most of its data. Time-sensitive opportunities should be capitalized on and risks should be mitigated whenever possible.

SLOTH: With sentiment analysis becoming an increasingly automated process, it is relatively easy to get lazy and make unrealistic assumptions. Accuracy will vary significantly based on the sentiment engine of the solution, but none of them will be absolutely perfect. In fact, many will require frequent human interaction to fine-tune the algorithms. This is not something to be overlooked and having some zeal can make all the difference. Failing to consistently verify the precision will cause the margin of error to fluctuate erratically resulting in useless data or potentially bad business decisions.

GLUTTONY: Sentiment analysis must be utilized in moderation. Overuse can lead to misguided assessments that can negatively impact your bottom line. Make sure that you are taking a balanced approach when analyzing your data and avoid being heavily reliant on sentiment. Keep in mind that sentiment is often entirely irrelevant. Just because you have the ability to leverage it doesn’t make it practical.

(Note: Artwork by Chris Hill)

About kramereric

Serial Entrepreneur, Digital Enthusiast, Strategic Consultant, @erkramer
This entry was posted in Social Media Monitoring and tagged , , , , , . Bookmark the permalink.

1 Response to 7 Deadly Sins of Sentiment Analysis

  1. Pingback: Social Media Monitoring: Return On Ignorance |

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s