This is a guest post by @LizaSperling, a social business professional in San Francisco, CA.
If your eyes glaze over when considering the slew of social media metrics (SMM) companies are tracking, you are not alone. You could dedicate all of your time to chasing metrics, but you don’t have to. Just start here.
Measure what can be measured: Some metrics cannot be calculated. The data is not consistently available or is horribly inaccurate. Don’t waste your time chasing down the holy grail of metrics.
Choose metrics that meet your business objectives: If you are using social media to respond to customer service inquiries, you will rely on different metrics than someone who is using social media for product insight. Metrics should evaluate your business objectives.
Consider both the quantitative and the qualitative metrics: Qualitative metrics like sentiment and influence are admittedly imperfect, but they should not be ignored. Social media, like any form of communication, is part science, part art.
Start somewhere: Metrics require a baseline, so even those businesses that are just beginning to embrace social media should measure their efforts. If you hope to make social media an accepted part of any business, you must have metrics to evaluate the value of the investment.
We’ll help breakdown what some of these metrics actually mean and offer you a few examples of tools that do a particularly great job at measuring that metric. Want a list of these tools for reference? See my Toolkit on oneforty.
1. Volume is simply the absolute number of times a brand or keyword is mentioned over a given period of time.
Tool: Alterian SM2
2. Share of voice is the comparison of the volume of mentions. For example the number of times Netflix is mentioned versus Hulu. Brands often measure their share of voice relative to competitors, their products versus competitors’ products and overall industry mentions. You can drill down to evaluate share of voice on Twitter versus Facebook or to track your share of voice during a campaign or product launch.
Tool: Lithium SMM / Scout Labs
3. Response time is how long it takes for a company to respond to an online mention. This is most relevant for customer service-oriented social media strategies with a heavy reliance on Twitter. Many companies also focus on response or resolution rate to gauge not only how long it takes to respond to a mention, but also how long it takes to resolve an issue associated with the mention. For example, if a customer tweets about an AT&T billing issue, AT&T may respond in 20 minutes asking for more information, but it may take an hour to resolve the customer’s billing issue and “close the ticket”. Tools with workflow features and ticketing/assignment capabilities are optimal to track response-related activities (both internally and externally) for each mention.
4. Frequently used words are the words used in conjunction with your brand. How are people talking about your brand and products? Chances are that your customers will refer to your brand differently on Twitter than forums. Is your brand frequently mentioned when your competitor is mentioned? Knowing exactly the words used, on which channels and the degree of frequency of each word will ensure that you are monitoring the right keywords, crafting content suited to each medium, catching emerging memes, and speaking customer-centric language.
Sysomos’ MAP tool offers the Buzzgraph, a way to track frequently used words around your brand. This sample is a Buzzgraph for Starbucks.
Tool: Sysomos (Buzzgraph)
5. Reachis the number of people who see a particular keyword or phrase. This value is an attempt to measure social media using traditional media metrics and terms like impressions and uniques. Admittedly, social media and traditional media do not lend themselves to an “apples to apples” comparison. One’s followers on Twitter do not necessarily indicate how many people are actually reading every tweet. Nevertheless, Tweetreach does a good job simplifying a calculation that is inherently murky and defines reach, or the total number of users who received a tweet, vs. exposure, or the total number of times tweets were received. The report also breaks out how users engage (retweeting or @ replies) and which users contribute the most reach.
6. Sentiment is the feeling or emotion of a particular mention. It is a qualitative measurement that will never be 100% correct – sarcasm, irony and slang make reading someone’s mind difficult in person, and even more so online. So why bother? Sentiment puts quantitative metrics like volumeand share of voice in context. AT&T may have more volume than T-Mobile, but if the mentions are largely negative, AT&T may consider their share of voice an unimportant metric.
A sentiment algorithm with at least 70% accuracy will save you the time required to read, interpret and hand score every mention. Be sure that you choose a tool that: 1) Determines sentiment on a key word specific vs. document specific level, i.e. identifies the sentiment of the keyword in the context of the document rather than the sentiment of the entire document; 2) Offers you and the ability to override the machine score; and 3) Is a learning algorithm that improves over time based on users’ corrections. Not every mention is positive, negative or neutral, so you may also consider tools that extract quotes that indicate wishes, caveats and comparison.
7. Influenceis who or what is driving actions, including purchasing decisions, brand awareness, adoption of behavior, etc…Like sentiment, influence is imperfect, but it is nevertheless a key to determining if your efforts are driving action or falling on deaf ears.
Tools: Klout and Twitalyzer
Again, for a complete list of these social media metrics tools check out my Toolkit on oneforty to guide you as you explore. And give us your take in the comments: What tools do you like to get these metrics? What analytics are particularly important for your brand? We’d love to hear from you.
Originally published Aug 16, 2011 2:11:00 PM, updated July 28 2017