I’m a change management and strategy analyst at my job, but our company is small so I wear many hats.
My first project at my job was doing social media monitoring for SQL Azure, a cloud database product that’s part of Windows Azure. It was a whirlwind of chaos and feeling like an inadequate moron. In fact, most nights I went to bed feeling like I had nailed my tongue to the wall and set the room on fire.
But. Never waste a developmental situation as my professor Dr. Paul Yost would say.
So in the spirit of learning and development I decided to reflect on what I’d learned. And lo and behold I did manage to learn a few things….
- You Can Get Lost—social media analytics requires an ability to do a deep dive into data but also requires that you don’t get lost in the data. This is easier said than done. Example: you are looking for influencers on your brand, so you start to look by channel (a channel being a social medium like twitter, blogs, forums etc.), but you realize that most bloggers talking about your brand are part of a forum community, this takes you to the forum community, then top forum posters, then forum poster’s twitter handles then….. It’s the proverbial case of playing in the woods and getting lost, because you didn’t mark your way or remember what you were doing in the woods to begin with!
- It’s a moving target—the social media conversation about your brand is always changing. This requires a strategic understanding and extreme tactical precision. Like a good sniper (I know a war analogy…go figure) you need to know where your target is going, how the target is moving, how the wind, elevation, dew point, distance etc. Will affect your shot. Because every shot you miss you have to recalibrate everything…oh and you don’t have a spotter. War analogies not your thing? Let me explain it this way:
- You have to understand, really understand, the BUSINESS PROBLEM you are trying to solve. As as a social media maven (internal or external) this is really just good customer service: what is the real need??? Remember no one wants social media analytics because they want pretty pie charts…they want to know who is talking about their brand, where they are doing this talking and whether or not they love, hate, or are in between their brand, how they can promote their brand, acquire new customer, and identify brand evangelists.
- Once you know the problem or question you are trying to answer, dive into the data to see what will give you that answer. But don’t dive in blind, that’s a time drain. Create a hypothesis of kind of data that might provide this answer and THEN dive into the data. For example: where the conversation is taking place can be found using a topic analysis by channel search.
- Forget What You Learned In Statistics— Period. Statistics is about isolating variables, controlling variance, designing studies to rule out plausible alternative hypothesis etc. Social media analytics is iterative. You might do 5 different “experiments” before lunch. With social media analysis it’s not as much about proving a theory it’s about testing narratives and story arcs. Think of it as research on crack, instead of research being 3-5 years behind, it’s done in real time. You start with an assumed story, grab the data you think might fit the story, if the data fit the story great, that’s it. If not, re-write the story and look at the data again. OR grab all the data you can and then sift through it to see what emerges. Data is data. It’s always trying to speak to you and social media analysts are better at listening than most statisticians. If that was abstract let’s me say it this way: Social media analytics is just real time consumer behavior research. It’s iterative, fast and has little time for traditional statistical and research procedures and methods.
There you have it. Three things I learned about social media.
Questions, comments, rants or praise? Let me know.
Guilelessly,
-The Behavioral Statistician