Metrics For Tracking Social Media Engagement
23

credit: luc legay
Social media is supposed to be a conversation. How do you make a conversation tangible with standard measurements? We would liken it to the term “engagement,†which defines how deep a visitor was on a site regardless of purchase. Typically engagement is measured by looking at several metrics. We can do the same thing for measuring your conversation rating. For our purposes engagement means the three-Rs: reference, response and retweet.
Ideally what you’d like to do is read, listen and watch every mention and interaction. However, without some sort of third party tool that will be next to impossible. There are several companies out there that can provide this sort of tracking. VMS is a well known and established company that has transitioned itself from a solely traditional media tracker to more robust tracking than anything you will need starting off. However, if you want to dive in head first, you should look at their software. An alternative would be a newer and lesser-known company called StatsIT. They offer a social media dashboard that is pretty compelling. Finally, Coremetrics, Omniture and Google Analytics all offer some solutions that are integrated with their existing software, however these solutions are all things that you can design yourself. You can build a Twitter API that will scrape any mention of your brand terms and will send the data to your database or analytics provider of choice. RowFeeder offers a service that populates a real time spreadsheet of your keyword (tell them Carlos sent you).
You’ll need to start with the metrics that you think are important to you. You could go as basic as appending marketing links to the tweets with URL’s and then measuring clicks, visitors, shoppers, and orders for each link. However, you can only measure your online statistics this way. The metrics you might be more concerned with, or rather will be more successful at utilizing immediately are the statistics that you can glean from Twitter itself. Our metrics are intended to save you the time and cost of a developer in the short run and help you improve your engagement immediately.
Before jumping into the metrics we recommend, there are some points you should keep in mind. You should take a holistic and a detailed approach to each of these metrics. What this means is that you should have a top level mark for all of your social media, then a secondary level for each individual account and then you should consider grouping these by brand term, category, or region. Finally, you should be able to dive into each account and break down your measurement by individual data line.
Several metrics that you could use are the following:
- Sentiment
- Reach
- Experience Rating
- Quality Score
Sentiment
As mentioned above, the search mechanism can be used to gauge how people feel about your brand. However, this is a fairly nebulous measurement and is not entirely precise. A live person looking through the tweets best determines sentiment, but Twitter Search can do a rudimentary filtering. Filter out invalid (non-actionable) tweets and if possible move them to the more appropriate bucket. We suggest that you establish a plan for responding to both positive and negative sentiment. You’ll also want to tailor these to your specific request; keep your eye out for tweets that have questions. Tweets that contain questions are the low hanging fruit for improving your sentiment values. Example of sentiment monitoring for Twitter: Twitrratr.com
Reach
Reach is the potential network of people who could be influenced by your social media. Think of a pyramid. Say your Twitter account has 100 followers, and in turn each of those followers has 100 followers of their own. Your reach would be 100 x 100, or a 10,000-person network (second-degree followers). You can then set up a query to show how many impressions per tweet you got and then divide that by your potential reach to determine a benchmark for success.
If you are doing a good job engaging your followers you should see a growing reach value. Once you have a large following (more than a 1,000 followers) you should consider refining your reach to only the people that have directly engaged in the recent past. Twitter will catalog either 9-days or 100-pages of tweets–which ever comes first. Assume that your real reach is roughly the value based on the people who engage (reference, reply or retweet) in a 24-hour period. So, if you generally have 100 users engage in a day either an expansion of breadth (engagement number) or depth (second-degree followers) should be considered a success.
The reason that reach should be considered with breadth in mind as you grow is because Twitter as a channel has a high loss rate. Only about 40 percent of Twitter users are still active after one month of sign-up. This doesn’t mean that no one comes back. If you have a large following reach is best measured by activity—not raw numbers.
Experience Rating
Experience rating is how you compare specific users based on the influence and quality of their interaction with your brand or account. So if @jhandy says that he likes your product but he only has a following of 10 and has never made a comment before, his rating would likely be 1. However if @welch says that he likes your product and he has a following of 10,000 and has made 100 comments on your product, he would likely have a rating of 100. You can aggregate your RT’s and mentions and then base them on the average rating to determine whether what is being mentioned is really impacting anything. If you find that your growth is slowing your should consider engaging your middle to high experience rated followers.
Quality Score
Finally, quality score is used to measure the number of mentions or comments you receive on your tweet. If you have one tweet and get 30 responses then you’d have a 30:1 ratio. You should be looking at your quality score on the small level of individual tweets and on the larger level of total engagement for the day versus total tweets for the day.
There will be a point where your followers and reach will slow. This is the time that quality score will be of particular interest. If you do not have a high quality score when your growth slows you should be investing in improving your engagement with your mid-rated users.
These are just some of the methods that you can use to measure your social media conversations and engagement. Creating your performance indicators will vary based on your goals and particular vertical.



January 24th, 2011 at 6:32 pm
Thanks for all the explanations of the metrics. I think when using these metrics, it’s important to take some of these metrics and establish goals for them up front, and associate with how they would help your final goals. I particularly like how you discussed having different marks for each individual account, brand, etc. These kind of segmentation is so important to really understand the actual impact of these relatively new metrics.
I also like your choice of names for these metrics, and how the “true” reach changes as you progress through.
I think it would be cool to do some comparison of how some of these metrics might differ when on different social platforms. For example, the magnitude of your reach on FB (Likes of your page) is likely different from your Twitter following. The responses that improve your quality score may vary between the different platforms as well.