The Big List Blog

| All |

The human side of social media analytics — Live from the Measurement Pre-Conference

Coverage of this session by Bridgette Cude of SocialMedia.org. Connect with her by following her on Twitter.

9:35 — SocialMedia.org’s Kurt Vanderah introduces eMetrics Summit and Digital Analytics Association’s Founder, Jim Sterne.

9:36 — Jim: This is not tech and it’s not software and it’s not math, it’s about people.

9:37 — Jim explains that’s why he wants to tell his story. We used to have a single communication, and story. “That was owned by Don Draper — until guys like Dilbert came along.”

9:38 — Jim: The Don Drapers of the world got it and saw it, but he had to talk to guys like Dilbert. Don Draper needs to understand his audience (through focus groups). Dilbert says this isn’t enough.

9:39 — Dilbert: We know what your audience is looking at, where they’re searching. There’s a whole bunch of tools if you’re willing to spend the money.

9:40 — Don Draper: I want to be able to talk to everyone wherever they are in their journey.

9:41 — Dilbert: There’s this thing called “big data.” We have more data than we’ve ever seen, it’s coming in much faster. And now we have to do these demographics, psychographics, and behavioral data. We have this raw material — but we need serious tools.

9:42 — John explains: ETL: Extract, Transform, and Load — in plain words: we have big analytics, but it’s not enough — we can map it, but there’s still a bottleneck when we analyze it.

9:43 — Map Reduce or Hadoop (the yellow elephant): All of these tools together make a giant ecosystem. But we want to know who is responding to the message. We need help on the human, creative side.

9:44 — A data detective: His job is to figure out which question we’re asking. “Now that I have all of this information, what do you want me to ask?” For example, does it raise awareness, improve attitude, influence, inspire?

9:45 — John says “I want you to recognize who these analysts are and what they need.”

  1. Understand Raw Material
  2. Understand the Tools
  3. Understand the Problem to be Solved (This is where the magic happens)Define it in language everyone understands. Be specific. Everyone has to agrees.

9:46 — John: Analysts care about “The Art of Social Media:” Start with the end goal, figure out how them got here, be careful with statistics. Use the data for illumination. Beware of cognitive bias. Human loves patterns — but they’re not always as they seem (for example: Ice cream sales and drownings are correlated.)

9:47 — “All models are wrong, some models are useful.” – George Box

  • Have knowledge of the tech, the math, and the business
  • We have to hire the multilingual problem solver
  • Be intuitive, a dreamer, and a lateral thinker.

9:48 — Here’s how an analyst should communicate: “Let me tell you a story.” People immediately understand the format and feel comfortable. Relate it in terms of the bottom line. Bonus points: figure out the individual’s goals.

9:49 — Most important: Have an opinion and be specific.

9:50 — Get happy with fuzzy. We’re not accountants. These are likelihoods.

9:51 — Big data is not just a hype term.

9:52 — It takes a village, you need a team. 9:53

9:54 — Data requires creativity to have insights. You have to have specific goals.

9:55 — It’s about asking the best questions. All models are wrong. Some are useful. Insights only work when you create the clearly. Get OK with fuzzy. Have an informed opinion based on the data. That’s the value the human brings to analytics.

Q & A:

Q: How do you hire for intuition?

A: John: We’re looking for people willing to take a leap. Do the lateral thinking “What Color Is the Bear Test.” You can get test questions like this on Amazon. It’s a gift and a talent, not learned or trained.

Q: How would you identify a culture where intuitiveness is celebrated and not dismissed?

A: John: Show examples of quick wins. Find someone willing to listen and test. They’ll ask you more, you’ll find out more, and it will grow organically. John jokes, if you’re in corporate culture, you’re never going to find that, so find another job.

Q: How do you take the pressure off with getting fuzzy?

A: John: use phrases like it is likely, the data suggests, it feels like.. “Statstics means never having to say you’re certain.”

Q: How many models is too many models? And how long are the good for? What does that conversation look like?

A: John: The classic consultant answer: It depends. Use it until you notice it’s usefulness starting to drop off. It’s time dependent, and campaign dependent.

December 9, 2013 0 comments

Comments are closed.