211 million pieces of user-generated content are created online every minute. People are talking, and they have a lot to say that’s important to you.
But people don’t just sound off on social media.
There are dozens of different mediums people turn to to provide feedback and details about their experiences online.
These include places like blogs, forums, articles, news stories, article comments, and review sites. All of these are in addition to the available platforms like Twitter, Instagram, Facebook, Pinterest and more. That means, there are tons of different ways people use to discuss topics that are relevant to you. And these user conversations are a treasure trove of data available to your team.
There’s real, actionable, accurate data to be gleaned from all of these other resources in order to give you key insights about your customers, potential customers, loyal fans, competitors and customers of your competitors and influencers.
Here is how and why your team needs to be utilizing conversational data from all over the web to better understand your audiences:
The Problem With Only Using a Limited Number of Sources
While the web is teaming with people who utilize social media frequently, this number only scratches the surface when it comes to useful data on audiences and online users.
It’s important that your team incorporates various types of data from online conversations, while also filtering out those that may not be as relevant.
Let’s go over a few reasons why limiting the number of sources could cause errors in your market research analysis.
1. People Talk Everywhere, Not Just On Social
For one, many of these users are passive consumers of social content. We’ll touch on this more later, but not everyone on social talks. Meanwhile, many of the 3.2 billion people who log on to the internet everyday are conversing on the web elsewhere. This means there is an opportunity to analyze user-created content for insights all over the web.
Let’s say Netflix wants to see how one of their shows is performing based on user feedback. You see here that people are talking about this specific series on social media, as you would assume is the case. But in our analysis during this 3-month span, only 55% of user-created conversations about it took place on social (shown here as microblogs):
Meanwhile, another 40% of conversations took place on blogs and forums. That’s an additional 40% of data that was missed out on.
2. The Sample Size Isn’t Representative
It’s crucial that you get a representative sample of your audiences when analyzing conversational data. The larger the the volume of data, the more reliable it will be. That’s because, as you add to the amount of data points, a true pattern will emerge from data, giving you a data set that is representative of actual population.
Of course, you may remember this key element of data analysis from your high school statistics class. But too many marketers still rely on analytics platforms that help them monitor the online community, but don’t give large enough samples to get a good read.
For example, many tools out there will give you a sentiment analysis of a brand or its campaign, but only provide a measly sample of 200 users. How reliable is that?
Instead, you need technology that can analyze and document all the many different sources on the web. If you think about all the many ways people can express themselves online from forums to review sites, not just social, this gives you a pretty good, reliable score of your brand health.
Limiting your team to just specific social media sources will keep you from learning the whole story.
3. The Quality Of User Data Matters
Quantitative analysis will help you better understand the saliency and relevance of the topics and searches you’re conducting, but the quality of what you’re searching matters as well. Someone’s 30-character tweet about an experience with a product may be useful, but a 1,200 word blog post is going to be a much better source to analyze.
Consider this: if we remove all sources except social media in our social listening tool for the Starbucks brand over the last 6 months you’ll get results that look like this:
Conversely, we can include other channels that contain higher quality user content (and it helps us up the sample size too!)
Here we have more precise insights and you’ll see we’ve increased our sample size and improved the different sources to get a better read on the audience personas and how they feel about the brand:
You see some differences in the above query overview, compared to the analysis of just social networks:
- The median age is 23, instead of 22
- There is a more female voice than male
- The audiences make over a thousand dollars less
- The content's reach was 325 million, compared to 262 million
- The brand's overall sentiment from audiences was way higher, 54% compared to 40% positive
With so many trolls and bots on social media these days, it certainly helps to add better quality conversational content into your analysis. Social listening tools that rely on more than just Twitter and Reddit will help you more accurately gauge how well your efforts or products perform.
The bottom line? Your analysis should be flexible enough to filter out results that aren’t representative or high quality data and give you the most relevant insights to your search.
Why Discovering and Removing Channel Bias Is Key
Different social channels have different user behaviors. Some are more conducive for conversations than others.
Note that more social media users consume content, rather than creating it. As to where people who are present on other platforms, such as messengers, forums and blogs, may be more apt to converse.
This is not to say people do not talk on social. On the contrary, millions of user-created posts happen every day on channels like Twitter, Facebook, Instagram, Pinterest and the like:
It’s up to you to figure out which conversations are most relevant to your brand and when and how analyze them in order to do so with precision.
Channel bias can exist thanks to previously mentioned elements like trolls and bots and similarly less reliable user content. But it can also be the result of your own impact. Maybe if you advertise on Instagram, you may have more data than on Twitter that is reliable, especially if you’re not running ads on Twitter. Often, it’s simply the medium that may cause channel bias within social media analytics. It is important for your research that you understand if your channel data skews negative or positive.
Twitter can, in fact, be a reliable source for brands to analyze. However the bite-size nature and virality and rapid spread of content on the channel may also be conducive to creating too much conversation that you don’t want to measure.
Each social media channel will have a bias on feelings. According to our analysis, Twitter is inherently more negative.
Also, geography should play into the equation. When we think about bias, cultural lifestyles could come into play, especially on a global, more accessible channel like Twitter. Therefore country of origin should be considered. Some people may be more difficult to please or harsher in their feedback than others, which could skew your data.
To account for these biases, our platform analyzes user-created content from all over the web, from long-form to social and everything in between. We also enable researchers to filter search results by channel, interest, geography, and age-range, among other things so you can fine tune your search based on your target audiences.
Again, it’s up to your team to properly refine your research to eliminate channel bias. But social listening tools like Infegy Atlas can surely help you do it. And do it well.
Now more than ever, people are talking about brands, products and experiences online. This will only continue to increase as we become a more digital-savvy universe with new innovative technologies.
It will therefore be imperative that you identify and analyze those conversations to learn more about your audiences, customers and influencers. Regardless of which side of the brand you are: marketing, advertising, media, market research or strategy, you can’t continue to rely on less than accurate data due to it only coming from limited sources.
This will mean investing in better tools and resources that extend the sample sizes, dig deeper into more relevant mediums and eliminate channel bias from your research.
In all, you have to listen where your audience is talking.
Want to find out where that is? Social listening tools like Infegy Atlas help you understand who your target audiences are and where they hangout online. Get a free demo today.