Understanding The Social Data Landscape
Social media ROI is a hot topic. As you may know, one of our strengths is rigorously measuring the return on marketing investment. Social media falls under the overall “marketing umbrella.” When we’re working with a client on measuring marketing ROI, one of the first things we do with them is sit down and figure out what data they have. Sometimes a client wants a specific question answered but they don’t have the right data to answer it. When measuring marketing ROI don’t just search for a single number that sums up your entire marketing effort. Instead, it’s better to ask specific questions like:
- Am I spending enough or too much?
- Where should I be spending my marketing dollars?
The bad news is that most social media data is classified as “unstructured data” – meaning that it’s highly unorganized and hard to automatically extract meaningful insights. Let’s spend some time taking a closer look at which social metrics you should be tracking (organized by structured or unstructured), and some of the challenges, advantages, and opportunities for each of them.
Structured social media data
Guardian recently used a useful analogy to explain the difference between content (unstructured) and metadata (structured), where content is “the letter” and metadata is “the envelope” that holds the letter. Structured data can provide the who, where and when, but not the what, how or why.
Structured social media data provides directional information (what’s working/what’s not) and is great for reporting your “social pulse”. The big hang-up on social listening information is that it’s very surface level. By itself, structured social media data is simply reporting what’s happening. It doesn’t provide you with information on why things are happening and how to improve them in the future.
Another downfall is that social “meta-data” may or may not be accessible, depending on the platform. Linkedin, for example, is notoriously restrictive of their data. Platforms like Facebook and Twitter continue to change how people access their data fire hose. Restrictions on API access may change the impact of structured data in the future.
Below is a list of some of the most commonly tracked structured social media metrics.
Structured social media metrics
Reach (audience size)
- Audience growth rate
- Brand “likes” or “followers”
- Unlikes (people who stop following you)
- Demographics of “followers”
- Demographics of people who comment or “like” posts
Engagement (audience activity)
- Average engagement rate
- Post clicks
- Post likes
- Post shares
- Frequency of posts over time
- Post times (time of day)
- Web traffic referrals by social media channel
- Leads by social media channel
- Lead conversion rate by social media channel
- Assisted social conversions (buyer had some engagement with you on social media along the sales cycle)
- Do not confuse this metric with “last click (direct) conversions.” As we’ve mentioned before, last click attribution can be a slippery slope because it over-values an action in the sales cycle. For example, if someone clicks on an online ad and makes a purchase, should the online ad get all the credit for the sale? Most likely it was a combination of activities that led the buyer to click on the advertisement, which, in turn, led to the sale. Assisted social conversions, although not an exact measure of social influence on a sale, are a nice way of measuring if social media is making a difference overall.
Unstructured social media data
More than 90% of social data is unstructured. Unstructured social data, a.k.a. content, is free text, frequently plagued by grammar, spelling, and “TXT speak”, making it particularly problematic to analyze. Unstructured data is the “meat and potatoes” of social media data. It may be hard to digest but it’s packed full of nutrients (insights).
Text analysis tools available today are imperfect at best, particularly when it comes to measuring sentiment. It is still hard to identify things like sarcasm. Achieving the most accurate insights from unstructured data requires a lot of manual labor, which can get expensive quickly.
Examples of unstructured social media data
- Status updates/tweets
- Brand sentiment
The opportunity lies in understanding both structured and unstructured data; how they work together and impact each other. Structured data is great for reporting the results. Unstructured data is great for identifying the “why” and how you can change behavior going forward. It’s a continuous cycle of data analysis and testing. To be successful, both structured and unstructured data need to be in the equation.
Are you interested in how to measure the ROI of social media, along with the ROI of your marketing as a whole? Subscribe to our blog to read our upcoming posts on this topic.