Determining Consumer Sentiment: Trends and Common Mistakes

Most marketers (89%) say they still rely heavily on manual analysis to determine consumer sentiment and do not yet feel comfortable leaving the task completely to software, according to recent research from Synapsify.
The report was based on data from an online survey of 70 social media community managers and marketing analysts who are regularly tasked with determining consumer opinion based on written content, such as Web comments, social media posts, tweets, product reviews, and emails.
Despite the availability analytics tools to process such data, most of the professionals surveyed report that they still conduct manual research: 33% say they always do so, 30% more often than not, and 26% sometimes.
More than half of respondents (59%) say they still do some manual analysis because they believe the sentiment word clouds used by software programs are incomplete, 21% do so because they do not have enough budget for technology tools, and 11% because they don't trust the accuracy of their current tools.

Below, additional key findings from the report.
Difficulties
Respondents say their biggest challenges in determining consumer sentiment are separating useful signals from noise and having too much data to sort through.


Ayaz Nanji is an independent digital strategist and a co-founder of ICW Content, a marketing agency specializing in content creation for brands and businesses. He is also a research writer for MarketingProfs. He has worked for Google/YouTube, the Travel Channel, AOL, and the New York Times.
LinkedIn: Ayaz Nanji
Twitter: @ayaznanji
Rate this
Overall rating
Add a Comment
Comments
Interesting data. Thanks for sharing Ayaz.
I think the biggest problem here is the accuracy of sentiment algorithms. At best...you're looking at 67% accuracy. Current algorithms just aren't accurate enough to detect things like sarcasm.
Smart marketers use sentiment to flag things, and then manually check them out to see if it's actually accurate.
Josh
This Is Excellent....Marketeers need to understand the technological devices and how to use them for their business analysis. The survey drawn are very surprising. Analysis of the business properly will only help to know the requirement for the business and how to maintain them for future.
Josh & Kimmy - Thanks so much for your comments.
Josh - Agreed on the algorithms, though it is pretty amazing to see the progress they've made over time. I'm really curious to see how well they do 5 years from now.
Kimmy - Good point. A big part of business analysis is indeed understanding how to use tech tools properly.
One of the biggest untapped sources of ground-level insight is the customer service team. Often these folks aren't trained to capture marketing intel or trigger an upsell, but they're often in the best position to do so.
But if you ask 4 or 5 customer service reps, "What are the things customers don't like about the product?" There's always a common thread. These individuals are usually people friendly, which makes them susceptible to quality feedback, even when it's negative.
LinkedIn %3E%3E http://www.linkedin.com/in/jamesclouser/