Social Media Market Research: A Study of the Tropicana Repackaging

At the 2009 MRA First Outlook Conference, Owen Shapiro of LJS Associates and Tom Malkin of GeeYee, Inc. presented "The Impact of Social Media on Market Research". In order to better understand the strengths and weaknesses of social media research, LJS conducted phone and social media research into the Tropicana unrepackaging incident.

Earlier this year, Tropicana rolled out repackaging that many shoppers complained looked like a generic brand and made the product much harder to find on store shelves. The resulting pushback led to Tropicana to revert to its earlier packaging.

To study this, LJS surveyed 1,000 U.S. consumers by phone and online and then weighted the results to be demographically representative. Fully 20% of consumers surveyed had noticed the new Tropicana packaging; of these, 32% liked the packaging, 32% were neutral and 27% were negative. Only 1% of consumers had posted online about the packaging.

For social media research, GeeYee scraped 1,900 posts from over one million web pages analyzed; each post typically covered three or four topics. Here is the categorization of the most frequently mentioned topics:

• Price - Primarily discussions of coupons and sales
• Premium OJ - Discussions of comparatively higher cost
• Ingredients - For instance, the shift to Brazilian rather than Florida oranges
• Confusion in store - Fully 34% of posters discussed not being able to find the product in stores
• New package design - Discussions of this accelerated once the packaging was withdrawn
• Reversion to "classic" packaging - Posters commented about Tropicana's announcement or upon seeing the old packaging restored

Comparing surveys to social media is like comparing apples to orange juice (sorry):

• Social media's impact is disproportionate to the level of activity. The multiplier effect for each posting produces an impact often missed by traditional marketing metrics. Lost in a typical survey is how vocal those who disagree might end up becoming.
• The data is dirty. When it comes to social media, you can't distinguish between grassroots and "Astroturf" (sponsored marketing campaigns design to mimic grassroots efforts), between legitimate blogs and "splogs" (spam blogs), between memes being relayed because they resonate or because they reciprocate (blog-post exchanges).
• Brands with unique names are easier to research. It is easier to extract quotes about the iPhone than Coach. Researching Tropicana required removing references to the Tropicana casinos in Las Vegas and Atlantic City, a level of cleaning not needed in coding verbatim survey responses.
• Topics are not always organized to answer our questions. Since posters can discuss anything, you need to develop a much broader coding sheet than for open-ended survey questions, which are tightly focused. Sometimes social media may not be talking about your question at all.
• Social media is not representative, only directional. You can try to quantify it to better understand the direction but it will still be qualitative.

Owen concluded that social media is best used early in the market research process: to develop hypotheses to test with surveys or to use up front as a qualitative tool to more fully understand issues and to see what you might have overlooked.

 Experience Vovici    Request a Call from a Vovici Rep     Request a Demo with a Vovici Rep