Today, digital data reporting, analysis, and application is a core part of most digital teams’ weekly or daily workflows. But as time becomes scarcer in the coming weeks, teams will have to consider what kind of social media data actually makes sense to report on—and what they can actually learn from it.
I often see campaigns and organizations limit their social media data gathering and reporting to metrics like followers, likes, and video views—the default basics in any stock social media report. Even though these aren’t unimportant, they’re often called “vanity metrics” because they don’t offer meaningful context about what matters—and what you can do about it.
Instead of focusing on vanity metrics alone, I recommend focusing on meaningful metrics—and a variety of them—to more thoroughly understand how audiences are — and aren’t — engaging with your campaign’s content.
This isn’t to say that vanity metrics like followers are totally meaningless. It’s meaningful to know, for example, that a certain candidate’s following increased at triple her average weekly rate following a standout debate performance or launch of a viral video that garnered millions of views. But it’s also important to recognize that metrics like these don’t tell the whole story—and without the important context added by more meaningful and actionable data points, can even be misleading.
Let’s walk through a few common examples.
Totals ⇒ Changes
Let’s say you’re a first-time candidate who launched an Instagram profile for her campaign from scratch in March 2019, and now has a following of about 55,000. If the incumbent has 100,000 followers, it may seem like his or her account has a much larger audience. But if, for example, 45,000 of those are neglected accounts or bots, is there really a difference?
If anything, the first-time candidate may even be better positioned on Instagram due to the relative recency with which the same number of users expressed an interest in what she has to say by following her.
In general, the older that an account is, the more time it’s had to accumulate “followers” from accounts that don’t represent an actual following—such as bots or accounts that have since been abandoned or neglected. If engagement is measured as a percentage of followers — as it often is by social media reporting software — this can also cause older accounts to appear to have relatively lower engagement rates on their content.
How can we make this meaningful? Instead of static totals like a total number of followers, focus on monitoring and comparing changes—and targeting growth.
Followers ⇒ Engagement
Because virtually all social networks are now heavily algorithmic, a profile “like” or “follow” may be best thought of as an indicator that a person is interested in that account’s content. But it’s not a guarantee that the account will reach, let alone engage, that individual. Especially with the ability to follow-but-mute on Twitter and Instagram, or “like” but unfollow on Facebook, followers and even follower growth are no longer the sure-fire indicators of building an attentive audience that they once were.
Instead of monitoring a metric like profile followers, I recommend looking at content engagements—metrics like the number of likes, comments, replies, shares, and/or retweets that various pieces of content are receiving. In some cases, engagement can also include clicks and video views, but more on those two later.
Engagement doesn’t have to be account- or post-specific. It can also look like the volume of mentions of certain usernames and/or hashtags. For a second layer of insight into both, consider looking at the quality of those engagements, especially the most prominent ones—is it mostly “likes,” “loves,” and supportive comments, or is it mostly “sad,” “angry,” and leaving you googling “profanity filter”? Is it somewhere in between? You can learn more about the quality of how your audience is responding to your content—not just the quantity—with careful application of social listening.
Clicks ⇒ Conversions
Here’s another tempting one often put front and center by auto-generated reports: “clicks.”
It’s important to first distinguish between “post clicks” and “link clicks,” which are often grouped together by social media reports with language like “clicks (all).” Post clicks are simply clicks anywhere on the post itself—to expand a photo or video, for example, or to read the full text, view the comments, open the post in a new tab, click through to the profile that posted it, click any of the engagement buttons — like, comment, share — or click on a link in the post.
Post clicks are a signal that someone wants to take a closer look at the post, indicating interest. It’s a form of engagement that goes just above reach. But post clicks shouldn’t be conflated with or over-counted as clicks on a link in the post, however tempting that may be.
This becomes especially important when reporting on posts with links to a call-to-action. 100,000 “clicks (all)” on a post about signing up to volunteer takes on a very different meaning if only 1,000 of those were clicks on the link itself—and a new meaning altogether if 0 of those clicks actually converted to signups.
Instead of clicks, focus on conversions—total and new (unique) emails signed up, and if applicable, donors before vs. after signup. If reviewing a fundraising post, look at the number of donations and the amount raised.
Video “Views” ⇒ Watch Times
It can be tempting to read a data point like 1.5M views, and assume that means 1,500,000 people watched a video all the way through. I often see campaigns and organizations limit their organic video reporting to views—and when stakeholders read those reports, some assume exactly that.
But think about how you watch videos online. Do you watch every video that starts autoplaying in your feed all the way through?
Facebook and Instagram count a view as three seconds. Twitter, per standards developed by the Media Rating Council and Interactive Advertising Bureau, counts a view as two seconds—with at least 50 percent of the video visible on screen. Platforms like Instagram Stories, Snapchat, and TikTok use views as a synonym for impressions—that is, a view is any video that starts playing for any length of time. In any case, a view isn’t very long.
1.5M “views” means something very different if for a video with an average watch time of three seconds versus a video with an average watch time of 15 seconds—and both can mean different things entirely if one is a 30-second ad while the other is an hour-long livestream.
I recommend looking beyond video views to report, assess, and apply additional metrics about videos, like total and especially average watch times. You can also look at views broken out as a percentage of the video, or the percentage of viewers who watched through 10 or 30 seconds. For live videos, it can also be instructive to look at statistics like peak live viewers—and to keep in mind that videos that are live streamed tend to have longer average watch times than those that aren’t.
Something else to consider as you edit future videos: How long, on average, are most people watching your videos? Do certain kinds of video openings tend to keep them engaged longer than others?
Over the next few weeks, as you work through your reporting, keep in mind that all social media data is inherently at the mercy of the social networks themselves, and their ever-changing features and APIs. If history is any guide, the accuracy of that data could probably still use some work, too.
Anastasia develops and directs innovative digital strategies for leading progressive campaigns and organizations. Before serving as Elizabeth Warren's Social Media Director, Anastasia launched and led the Social Media Department at Trilogy Interactive, where she worked with dozens of high-profile clients on refining and growing their online presence. Follow her on Twitter at @golovashkina.