The Big List Blog

| All |

Social media intelligence: Brand tracking using social media data — Live from the Measurement Pre-Conference

Coverage of this session by Kristen Platt of Connect with her by following her on Twitter.

11:55 —’s Megan Uithoven introduces University of Maryland’s Associate Professor and Director of MS in Marketing Analytics, Wendy Moe.

11:56 — Wendy defines social media marketing/intelligence: Looking all the social data online and using it learn insights about our customers. Then you take those insights and integrate them into your social media strategy.

11:57 — Start with social media monitoring: following influencers, monitoring social channels, pulling reviews, etc. But this is sometimes biased with what your particular end-goal is.

11:58 — Determine what data is important to your social strategy to measure. We don’t have a solid understanding of how these metrics link to customer behavior.

11:59 — There is little correlation between social sentiment and offline surveys (.008).

12:00 — What’s wrong with current measuring practices: scalability, analyst bias, venue effects, selection effects that favor buzz-worthy topics social dynamics that favor the extreme, and what exactly does the number of mentions or average sentiment mean?

12:02 — Wendy says you first need to understand the drive behind the customers’ behavior to post online.

Then, discover the implications for observed metrics and trends.

Integrate social with traditional sources of marketing tactics.

12:04 — Wendy: Break down the behavior and figure out the online opinion:

  • Who? (posters vs. lurkers)
  • What? (expressed sentiment vs. underlying opinion)
  • Where? (venue differences lead to venue effects)

12:05 — Wendy: Posting behavior can be broken down into two components:

  1. Opinion formation (a function of satisfaction)
  2. Opinion expression (subject to a variety of biases and dynamics like expert effects)

12:06 — Wendy asks, “How do dynamics affect what we observe? Answer: Overtime we see a variance of opinions.”

12:09 — Wendy shares what influences expressed sentiment: Venue, venue-specific dynamics, and message topics.

12:11 — General Brand Impression (GBI) better correlates with offline brand tracking surveys when you use a one-month lag (because social is so timely, you need to give it some time). GBI has the best potential as a lead indicator.

12:13 — Wendy shares some big takeaways:

  • Significant social dynamics exist. Encourage a variety of opinions to include the moderate majority. This encourages discussions and insulates impact on sales.
  • Social media behavior varies across venue formats. Monitor multiple sources of social media data.

Q & A:

Q: Is there a model for tracking mentions?

A: We coded it for product characteristic, model, and venue got mentioned. What is left then, is the general brand impression once those are all filtered.

Q: Is engagement a factor in the poor correlation between Twitter impressions and offline surveys?

A: We don’t see a correlation even with the lag built in. The involvement with social media is going to vary from market to market.

Q: Is there a bigger model we could take a look at about social dynamics?

A: Yes, you can find that info in my book, as well as in my research papers that can all be found on my website.

December 9, 2013 0 comments

Comments are closed.