A Framework for Social Media Measurement Tools

By on in , , , , , , with 2 Comments

Fundamental marketing measurement best practices apply to social media as much as they apply to email marketing and web site analytics. It all begins with clear objectives and well-formed key performance indicators (KPIs). The metrics that are actually available are irrelevant when it comes to establishing clear objectives, but they do come into play when establishing KPIs and other measures.

In a discussion last week, I grabbed a dry erase marker and sketched out a quick diagram on an 8″x8″ square of nearby whiteboard to try to illustrate the landscape of social media measurement tools. A commute’s worth o’ pondering heading home that evening, followed by a similar commute back in the next morning, and I realized I might have actually gotten a reasonable-to-comprehend picture that showed how and wear the myriad social media measurement tools fit.

Here it is (yep — click on the image to view a larger version):

‘Splain Yourself, Lucy

The first key to this diagram is that it makes a distinction between “individual channel performance” and “overall brand results.” Think about the green box as being similar to a publicly traded company’s quarterly filing. It includes an income statement that shows total revenue, total expenses, and net income. Those are important measures, but they’re not directly actionable. If a company’s profitability tanks in any given quarter, the CEO can’t simply say, “We’re going to take action to increase profitability!”  Rather, she will have to articulate actions to be taken in each line of business, within specific product lines, regarding specific types of expenses, etc. to drive an increase profitability. At the same time, by publicly announcing that profitability is important (a key objective) and that it is suffering, line of business managers can assess their own domains (the blue boxes above) and look for ways to increase profitability. In practice, both approaches are needed, but the actions actually occur in the “blue box” area.

When it comes to marketing, and especially when it comes to the fragmented consumer world of social media, things are quite a bit murkier. This means performance measurement should occur at two levels — at the overall ecosystem (the green box above), which is akin to the quarterly financial reporting of a public company, and at the individual channel level, which is akin to the line of business manager evaluating his area’s finances. I use a Mississippi River analogy to try to explain that approach to marketers.

Okay. Got It. Now, What about These “Measurement Instruments?”

Long, long, LONG gone are the days when a “web analyst” simply lived an breathed a web analytics tool and looked within that tool for all answers to all questions. First, we realized that behavioral data needed to be considered along with attitudinal data and backend system data. Then, social media came along introduced a whole other set of wrinkles. Initially, social media was simply “people talking about your brand.” Online listening platforms came onto the scene to help us “listen” (but not necessarily “measure”). Soon, though, social media channels became a platform where brands could have a formally managed presence: a Facebook fan page, a Twitter account, a YouTube channel, etc. Once that happened, performance measurement of specific channels became as important as performance measurement of the brand’s web site.

When it comes to “managing social media,” brand actions occur within a specific channel, and each channel should be managed and measured to ensure it is as effective as possible. Unfortunately, each of the channels is unique when it comes to what can be measured and what should be measured. Facebook, for instance, is an inherently closed environment. No tool can simply “listen” to everything being said in Facebook, because much of users’ content is only available to members of their social graph within the environment, or interactions they have with a public fan page. Twitter, on the other hand, is largely public (with the exception of direct messages and users who have their profile set to “private”). The differing nature of these environments mean that they should be managed differently, that they should be measured differently, and that different measurement instruments are needed to effectively perform that measurement.

Online listening platforms are not a panacea, no matter how much they present themselves as such. Despite what may be implied in their demos and on their sites, both the Physics of Facebook and the Physics of Twitter apply — data access limited by privacy settings in the former and limited by API throttling in the latter. That doesn’t mean these tools don’t have their place, but they are generalist tools and should be seen primarily as generalist measurement platforms.

Your Diagram Is Missing…

I sketched the above diagram in under a minute and then drew it in a formal diagram in under 30 minutes the next morning. It’s not comprehensive by any means — neither with the three “social media channels” (the three channels listed are skewed heavily towards North America and towards consumer brands…because that’s where I spend the bulk of my measurement effort these days) nor with the specific measurement instruments. I’m aware of that. I wasn’t trying to make a totally comprehensive eye chart. Rather, I was trying to illustrate that there are multiple measurement instruments that need to be implemented depending on what and where measurement is occurring.

As one final point, you can actually wipe out the “measurement instrument” boxes and replace those with KPIs at each level. You can swap out the blue boxes with mobile channels (apps, mobile site, SMS/MMS, mobile advertising). I’m (clearly) somewhat tickled with the construct as a communication and planning tool. I’d love to field some critiques so I can evolve it!

Similar Posts:

2 Comments


  1. Hi Tim,

    this is actually very good. I pretty much came up with the same thing and I am desperately looking for an open source solution that is somehow implementing this. It would even be good just getting the basic and easy to come by numbers like followers or likes…

    There are some quite expensive solutions available, mostly offered as a service via a web app. Some with limited functionality, e.g. just monitoring one service are even available for free.

    I am looking for something that provides the main architecture (model and database storage abstraction, framework with basic GUI, cron interface), a plugin system for the social media services to be monitored (the getters or collectors) and a smart output engine where I can configure graphs, charts etc. (a dashboard).

    Also flexible dashboard solutions are available but tend to be expensive too.

    Further I was thinking of the architecture and the technological basis. There is quite some code flying around in php, so the usual LAMP stack seems feasible. Personally I would like to use python and therefore I am thinking of a django based project with some Javascript/AJAX frontend e.g using jquery and flot.

    The thing is that I am not a seasoned programmer, just a little hacker…..

    Let me know if you find any solution or even a good basis or starting point for a solution…

    Good luck with your idea,

    Peter

  2. Thanks, Peter!

    We’re definitely grappling with the same issue — using a wide array of inexpensive and imperfect point solutions vs. using a smaller set of much more expensive (and still imperfect) solutions vs. something in between vs. some combination combined with our own proprietary technology. The landscape is shifting super-fast, and there are lots of people jumping into the space with their solutions. With both the social platforms themselves evolving, AND new data collection/presentation technologies emerging (generic open source as well as traditional BI tools), it’s an exciting, yet daunting, mess!

    I would have called myself a “hacker”…but you’ve raised the bar on that label, it seems!

Leave your Comment


Notify me of followup comments via e-mail. You can also subscribe without commenting.

« »