10 key take-aways on Social Media Analytics (Dec 9 NYU)

Which social media tools help managers to best predict brand attitudes? Do hotels advertise more or less faced with online reviews? Why do you buy red dresses online but return them the next day? These were just some of the questions addressed on the second day of the NYU conference on digital, mobile marketing and social media analytics (http://www.stern.nyu.edu/experience-stern/about/departments-centers-initiatives/academic-departments/marketing/events/conferences/big-data-conference-2017)

Screen Shot 2017-12-10 at 2.50.50 PM

  • Hotels advertise less after online ratings hit a 0.5 cutoff threshold and more when they have either low ratings (below 3) or very high (above 4.5)
  • Disclosure of being sponsored does not reduce Facebook ad effectiveness, instead the higher source recall increases brand attitude and purchase intention.
  • Combining model-based and design based experimental setup can cut the standard error in half, giving more chance to find significant effects at a lower cost
  • Loan defaulters mention the following words more in their loan application: God, god bless, payday, please, thank you
  • Hotel app adopters may spend less at your hotel because the app encourages them to search more and reduces their loyalty
  • Solicited reviews give higher ratings to the hotel, but are less trusted by consumers and confuse their decision making, as they rank different-quality hotels closer together.
  • You buy red dresses in small sizes online, but return them at great loss to the company, while you are wearing your black and blue outfits offline.
  • Both novelty and usefulness of your idea increase its Kickstarter success, but with negative synergy
  • As we do not get private information on Facebook users, we should target prospects by observing which other brand they comment on in a similar way to our current customers.
  • Likes, language negatives (‘hate’) and Support Vector Machine positives (‘not bad’) are the best brand attitude predictors among sentiment extraction tools from social media.

Screen Shot 2017-12-10 at 2.54.39 PM

This last insight is from my work with the amazing Raoul Kubler and Anatoli Colicev – send me an email to get the full presentation and paper! kpauwels@northeastern.edu

Advertisements

Leave your comment here

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s