The world of social media is awash in a sea of data: followers, friends, likes, links, purchases, engagement. Anyone trying to make sense of (and justify) their social media efforts has to dive into this sea. Fortunately, there are a plethora of tools tools available (free and otherwise) that slice and dice this data in different ways. But how much do they actually help? In the case of many of the free tools out there, the basic principles of information design are taking a backseat, with incomprehensible charts, mysterious metrics and marginal usability ruling the day and obscuring the valuable information present.
Below are a few examples of information design in free social media analytics tools that could use some improvement. It’s not an impossible task; a few simple refinements would make a huge difference. Instead, sometimes we’re just left scratching our heads, or even worse, drawing the wrong conclusions.
NOTE: Most of these free tools are new and still evolving, with developers contributing their time and efforts to bring a little order to the chaos. It’s easy to criticize and forget the innumerable hours of hard work that go into creating and improving free web apps. So, if any of you developers are reading this, keep up the good work!
Problem 1: Poorly labeled axes on graphs cause confusion
Every self-respecting graph must provide clearly labeled axes. Without them, data is left naked, and the viewer is left guessing about what’s being shown. Unfortunately, unlabeled or poorly labeled axes seem to rule the day with many social media graphs.
Problem 2: Non-linear data scales misrepresent information
In scientific data presentation, one encounters data scaled in a number of different ways (e.g., linear, logarithmic). Most web analytics tools (for social media or otherwise), use linear scales. Regardless, the scale should be labeled and used correctly.
In the particularly egregious example below, TwitterGrader shows followers over time and labels the axes, but the axes are a total mess (not linear, not logarithmic). As a result, what looks at first like a decreased rate of follower adoption is nothing of the sort.
Problem 3: Mysterious metrics, poorly displayed
It seems like many social media analytics tools use different metrics, none of which are standardized. These metrics are presented in dashboards with bold numbers, bright colors, arrows and checkboxes and thumbs up or down. But what do they actually mean and is good information design used to help clarify? Invariably, definitions are not shown inline, so one has to click around to find what a given metric means. Additionally, scales are not provided for context; raw numbers are shown, and it’s left to the user to figure out the maximum and minimum for any given value, not to mention whether or not a given number is “good” or “bad.”
Klout is an example of a powerful tool that’s grown a lot, but that still has some confusing ways of presenting metrics and data.
In Twitalyzer’s defense, their latest interface has inserted a popup window that tries to clarify the metrics and visual presentation of information. The alternative would be to design things clearly in the first place, so that excessive instructional content isn’t needed. Additionally, when basic things like legends or descriptors are hidden one or two clicks away, it takes people away from what they’re there to do: analyze data.
Problem 4: Lots of data is stale, but it’s not obvious
A lot of the free social media tools out there only scan for data when told. As a result, much of what you’ll see is out of date, and not just by days or weeks, but often months. Unfortunately, information in these interfaces doesn’t always make it clear over what time period a data set applies, or the last time a given profile was scanned. A little information design would go a long way to fixing this problem.
Conclusion: Caveat emptor
All of the examples shown above are for free tools, and in this case, you usually get what you pay for. Lots of data, some useful functionality, but some pretty rough edges. More sophisticated enterprise tools (like Radian6, Scout Labs, Sysomos or Viral Heat), don’t seem to suffer from these information design oversights, and it’s likely that the best free tools will improve significantly over time.
In the meantime, if you’re using free social media analytics tools, take the time to understand exactly how data is being presented, and the metrics being used. Free tools are great, and many are pretty powerful, but careless interpretation could transform great insights to lies or propaganda.
What are your thoughts? Are there free social media tools out there that you think get it right from an information design perspective?
NOTE: This entry was previously published on Sequence.com.