Keeping up with the latest terms used in social media analytics isn’t easy. There is a constant stream of new technology, data, and processes that brands use to uncover consumer insights from social media. That’s why we decided to create a glossary to help keep track of it all. By no means exhaustive, this glossary will continue to grow with new terms and updates to definitions.
For a high level introduction to social media analytics, check out our Elements of Social Media Analytics series.
Action recognition is part of image analysis, and is the ability for a computer program to identify specific categories of actions like running, dancing, eating, etc. For example, tracking which brands and products appear most often in social photos of “eating” could help you better understand consumer behavior and your industry.
Affinities refer to the interests of an audience how those interests compare to the general population or another specific audience.
API is an acronym for “Application Programming Interface”. APIs serve as a standard for different software programs to communicate with each other. Within the context of social media analytics, APIs enable:
- Analytics firms to procure data from content sources (Twitter, Facebook, etc.)
- Customers to export analytics into business intelligence tools (Domo, Tableau, and similar), proprietary dashboards, CRM solutions, or other custom applications
- Researchers to import/upload their own unstructured data into analytics tools, such as open-ended survey responses, customer service cases, or product/service reviews
More on APIs:
In social media analytics, audience analysis refers to researching the interests, preferences, demographic, location, and other aspects of a group. This could be a broad audience like “all Twitter users” or a much narrower group like “millennial female fans of The Bachelor in Massachusetts.”
More on audience analysis:
- What is Audience Analysis?
- Crimson Hexagon’s Audience Analysis
- The 3 Audiences of Back-to-School Shopping
A list of specific content sources (such as websites or particular authors) that can be designated by the user which will not be included in the collection of data.
In Boolean searching, an “and” operator between two words or other values (for example, “pear AND apple”) means one is searching for documents containing both of the words or values, not just one of them. An “or” operator between two words or other values (for example, “pear OR apple”) means one is searching for documents containing either of the words. In social media analysis, seasoned analysts may choose to use boolean, while other may use quick search tools that don’t require any understanding of boolean operators.
On social media, accounts that post automated promotional content or spam are referred to as bots. It’s important to filter out such posts and accounts for accurate social media analysis.
Brand health refers to a collection of metrics surrounding your brand’s online presence. While the specific metrics tracked often vary, the ultimate goal is to measure your brand’s awareness and sentiment in the eyes of consumers.
More on brand health:
Crisis monitoring for brands means constantly looking for signs of a potential brand crisis as well as monitoring the response to a current brand crisis. Social media analytics is incredibly useful for crisis monitoring as it allows brands to track consumer reactions in real-time on a large scale.
Competitive intelligence (CI) can be defined as the collection and analysis of information about products, customers, competitors, and any aspect of your business’s competitive environment. Social media analysis allows brands to gather competitive intelligence on what competitors are doing and how consumers view them in real-time.
Computer vision is the science that aims to give machines or computers the ability to understand and extract useful information from images. Image analytics is considered part of computer vision. For more, see image analysis.
Customer or consumer insights is an interpretation of trends in human behaviors. These insights are used by brands and agencies to increase the effectiveness of marketing, products and services, advertising, and more.
More on consumer insights:
Custom social media analysis categories allow brands to use machine learning to create their own categories for measuring social media conversation. For example, standard categories for sentiment analysis are positive, negative, and neutral. If a brand wanted to gauge consumer reaction to a PR crisis they could create categories like “outraged / expect better/ this issue is typical / sympathetic / neutral .” These categories specific to the situation would provide a more accurate and detailed look at consumer response to a brand crisis.
Any social media content that is shared outside of what can be measured by analytics—it has no known source For example, if you copy and pasted this blog post’s URL and shared it via email or text message, that would be dark social.
A virtual, shareable canvas with the ability to customize a collection of meaningful metrics in an easy, straightforward way.
More on dashboards:
In general, demographics refers to statistical data relating to the population and particular groups within it. When applied to social media data, demographics look at factors like the age, gender, location of specific groups or audience segments. Social media analytics can identify the demographics of groups discussing any brand, topic, or product.
Earned media refers to media exposure you’ve earned through word-of-mouth. On social media, this refers to people discussing your brand Whether it was the fantastic content you’ve distributed, the influence of your SEO efforts, the customer experience you’ve delivered, or a combination of all three, earned media refers to the recognition you receive as a result.
Engagement rate is a metric used to describe the amount of interaction a social post receives such as likes, shares, comments.
Emoji analysis is the ability to analyze emoji use the same way you would analyze text. From example, instead of just seeing the top words in the conversation about your brand or product, you can also see the top emojis used in the conversation.
More on emoji analysis:
Emotion analysis goes beyond positive, negative, and neutral categorization. Posts can be automatically categorized predetermined human emotions. For example, the most common emotions according to leading psychologists and anthropologists are anger, fear, disgust, joy, surprise, and sadness.
More on emotion analysis:
In social image analytics, face analysis is used to identify the age and gender of the faces within a photo to track demographics. This differs greatly from other uses facial recognition, which aims to identify the specific person within an image.
An online forum, or message board, is a discussion site where people can hold conversations in the form of posted messages. Reddit is one of the most popular forums, with subcategories of discussions on a wide variety of topics.
More on forums:
A data “firehose” is a steady stream of all publicly available data from a source in real time. The stream is constant, delivering new, updated data as it happens. For example, Crimson Hexagon receives all public Twitter posts in a constant stream, thus this stream of data can be referred to as the “Twitter firehose.”
Shorthand for “geographic filter”, or a method of filtering your social analysis by location. For example, you may want to see posts about your new product, but only in New York.
The location data users can choose to attach to a social post.
More on geotags:
A feature on many social media platforms that allows users to share their content with geographically defined audiences. Instead of sending a generic message for the whole world to see, you can refine the messaging and language of your content to better connect with people in specific cities, countries, and regions. You can also filter your audience by language.
In social media analytics, historical data refers to the archive of social posts from the past that are available for analysis. Historical data is crucial for trend analysis and understanding how consumer opinion has changed over time.
More on historical data:
Influencers are social media users or accounts that drive and influence a conversation. In social media analytics, identifying influencers adds context to analysis and helps brands understand the reasons behind the influence and how to harness it.
More on influencers:
Image analysis (also know as “computer vision”) is the ability of computers and software to recognize attributes within an image. In social media analysis, image analysis adds another layer of context on top of text analysis. The same methods of categorization used in text analysis also apply to image analysis. Image analysis allows you to gain insights from the logos, objects, scenes and action in images shared across social media.
Klout is a website and mobile app that uses social media analytics to rate its users according to online social influence via the “Klout Score”, which is a numerical value between 1 and 100.
Logo recognition detects brand logos within images. Logo recognition technology can be applied to social media images to allow brands to identify how their logo and products are appearing online, with or without text mentions of the brand. This is useful for identifying sponsorship ROI, how customers are using products, as well as misuse of brand logos.
Machine learning gives a social media analytics tool the ability to learn exactly what you’re looking for in your analysis and categorize results based on example posts instead of just keywords and rules.
More on machine learning:
Metadata means data that provides information about other data. In terms of social media data, the metadata about a post could include background information like the time of day, location, user info, device type, etc.
More on metadata:
Moments of Consumption
Moments of consumption are the instances when consumers are using or consuming a specific product and the context around how and why. For example, someone drinking a soda on the beach with friends would be a “moment of consumption” from the soda brand’s perspective. Image analysis allows brands to measure and better understand these moments of consumption through images shared across social media.
In social media analytics, a monitor refers to group of reports based on a specific set of keywords, content sources, date range, filters, training data, etc. For example, your brand may choose to create a specific monitor to measure the conversation around a product launch.
Crimson Hexagon’s platform has two types of monitors:
- Buzz Monitor
A Buzz Monitor is not based on training documents. Results are based on a single set of keywords, content sources, date range, and filters.
- Opinion Monitor
An Opinion Monitor is based on a combination of training documents and a single set of keywords, content sources, date range, filters, etc.
Natural language processing
NLP is used to analyze text, allowing machines to understand how humans speak. This human-computer interaction enables real-world applications like automatic text summarization, sentiment analysis, topic extraction, named entity recognition, parts-of-speech tagging, relationship extraction, stemming, and more. NLP is commonly used for text mining, machine translation, and automated question answering.
Net Sentiment the net value the opinions and feelings consumers express in an online conversation, ultimately determined to be positive, negative, or neutral. With social media analytics, you can measure the sentiment about any brand, product, or topic and track how it changes over time.
More on net sentiment:
Object recognition is the ability for a computer program to identify objects within images. Objects could include anything from broad categories like “building” or “person” to specific products like “smartphone” or “coffee.”
Describes the number of unique people who view your content without paid promotion. The distinction between organic and paid reach is, of course, that the former is free. People come across this content through the feeds, streams, posts, pages of their contacts—usually friends, family, colleagues, trusted brands, and cats/dogs.
Owned media is content you’re in full control of, such as your company website, your blog, and your social media accounts.
Paid media refers to all paid advertising, sponsorship, and promotion of your brand. On social media, this is any social advertisements or posts that you pay to promote.
Posts per million
A way of measuring the relative frequency of posts about a given topic, essentially answering the question: “Out of one million social posts, how many are about a given topic?” Using PPM as a metric is also an important way to control for the changing volume of social media posts overall.
The study and classification of people according to their attitudes, aspirations, and other psychological criteria. In social media analytics, psychographics can refer to the overall preferences and opinions specific audience segments or consumers in general. The affinities, or interests, of audiences would be considered psychographics.
Purchase intent is a measurement of the probability that a consumer will purchase a service or product. In social media analytics, purchase intent can be calculated with custom categories based on posts that express varying levels of intent to purchase a product.
More on purchase intent:
A social media analytics report contains a select group of insights and metrics from your monitors. A report is typically static, in that it is a snapshot of a given period of time. Dashboards provide the same data as reports, but can track metrics dynamically in real-time.
Scene recognition is part of image analysis and refers to the ability for a computer program to identify specific categories of locations like restaurant, park, beach, etc. For example, knowing that your beer brand’s logo appears most often on beaches gives you valuable context on how consumers are using your product.
On the most basic level, sentiment analysis, or “opinion mining,” strives to answer the questions “what do people think?” or “how do people feel?” about a particular topic. To get more technical, sentiment analysis is the measurement of positive language and negative language.
More on sentiment analysis:
Share of eye
Share of eye on social media refers to a brand’s share of visibility social media images. If share of voice is based on text mentions of a brand, share of eye is based on visual mentions in images such as a brand’s logo, branding, or products in comparison with competitors. With image analytics, brands can now accurately measure share of eye.
Share of voice
Share of voice on social media refers to a brand’s share of social media discussion in comparison with competitors. Traditionally, share of voice referred to a brand’s share of advertising and visibility, but now includes sources like social media and organic search in addition to all forms of paid advertising. Social media is the only share of voice metric that measures the voice of the consumer.
Social media listening
Social listening is the process of monitoring social media for specific information such as mentions of your brand, competitors, product, or other topics relevant to your business. Listening refers to the identification and collection of specific data, but the next step is analyzing that information for relevant insights.
Social media analytics
Social media analytics refers to extracting valuable insights from the vast amounts of social media data to enable informed and insightful decision making.
More on social media analytics:
Social media intelligence
Social media intelligence refers to the collective tools and solutions that allow organizations to monitor and analyze social channels and conversations and synthesize social data into meaningful trends and insights.
Structured data refers to the kind of data that is organized and displayed in a database with rows and columns, making it straightforward to work with. Examples of structured data include everything from social shares and likes to sales numbers and business addresses. These types of data are easy to sort and categorize.
In machine learning-powered social media analysis, training data consists of example posts use to teach the algorithm how to automatically categorize posts. For example, if you want to teach the algorithm to identify posts that convey “purchase intent” you would provide examples of social posts that convey purchase intent as training data to serve as examples. Typically the more training data you provide, the more accurately the algorithm can identify similar posts.
Text analytics refers to analyzing text to uncover meaningful insights and other valuable information. The process of achieving this goal falls broadly into 2 types of approaches:
- Linguistic rules and keywords approach
- Statistical or machine learning approach
Unstructured data is variable and complex, making it much more difficult to sort, categorize and analyze than structured data. Examples of unstructured data include text in emails, images, and any form of human language in a conversational format (like social media posts).
More on unstructured data:
User Generated Content (UGC)
User-generated content is any content (videos, photos, quotes, etc.) created by consumers. Marketers often tap into their social audience to collect this type of content to support a campaign or initiative.
More on user generated content:
Voice of the Customer (VoC)
The voice of the customer refers to the process of understanding customer’s expectations, preferences and aversions. The voice of the customer can be captured in a variety of ways including social media analysis, interviews, surveys, focus groups, and more. Social media analytics provides brands with a fast and efficient way to understand the voice of the customer on a large scale.
More on voice of the customer:
Volumetrics simply means measurement by volume. In social media analytics, volume is one of the simplest metrics to track. More qualitative metrics like sentiment, emotion, purchase intent, etc can be tracked with an advanced social media analytics platform like Crimson Hexagon.
More on volumetrics:
A list of specific content sources (such as websites or particular authors) that can be designated by the user which will only be included in the collection of data.
Word clouds, also known as tag clouds or weighted lists, are a visual representation of text, where the frequency of a word determines its size in the word cloud. This is a great tool for identifying words that are repeated or most common.