On DigitalCan Photo-Driven Crowdsourcing Provide Brand Insights?
Irfan Kamal April 19th, 2009
A new experimental site, Photocracy.org, is a crowdsourcing initiative to explore whether images are useful in providing brand identity information. The people behind the site, Princeton University student Josh Weinstein and Sociology Professor Matthew Salganik, are currently collecting information on perceptions of 3 countries – the United States, China and Japan – using votes on images of those countries. Visitors from each country are asked to choose which of 2 pictures is more representative of the country. Since the site’s launch a few weeks ago, over 100,000 votes have been cast for comparisons like this one:
Photocracy.org US comparison images - cowboy on horse vs. Obama
Why is this useful? To find out, I asked the people behind the project some questions.
What specifically are you hoping to accomplish from this project?
Matthew: This is a study of cross-national perceptions, and we are hoping to find the photos that best represent Japan, China, and the United States according to voters from Japan, China, and the United States. [We], as well as visitors to our site, will be able to see how Americans view their country, as well as how it is viewed by people from Japan and China. Likewise, we can see how people from Japan and China view their country as compared to how it is viewed by others. If the project works well with these three countries, we hope to expand it to cover additional countries in order to get a richer understanding of global public opinion.
What are the broader potential implications and uses for this type of work?
Josh: The potential implications for crowd-sourced pairwise comparisons is limitless. Both in textual … and pictorial form (as in Photocracy.org), pair-wise comparisons are an excellent and easy way to sort lists of ideas. Further, the availability of user-generated options adds a further bonus so that, potentially, ideas users submit can reach the top.
The visual basis of the methodology adds another level of potential. For starters, by using pictures instead of text, translation issues are effectively eliminated. Most importantly, this methodology is entertaining and simple for the user.
Another approach would’ve been to ask people to describe the different countries in words, or using a set of written questions. Is there something about using photos that is more effective for brand identity research – in comparison to using written questionnaires?
Matthew and Josh: The choice of pictures over text was made for two reasons. First, offering visual aids instead of written words bridges the language gap. For example, if an American uploaded that she thought that Chinese were “aggressive”, how exactly would we translate that into Chinese? There are probably several words with similar definitions, but different connotations (for example, in English there are “warlike”, “violent”, “assertive”, etc). This translation problem plagues cross-national research, but we avoid it by using photos. On our site, people interact directly with items uploaded from other countries with the researchers standing in between them. A second reason for using photos is that interacting with pictures is more interesting than interacting with text. This site relies on volunteers to come and vote and we thought that using photos would be more fun. As we enter an increasingly visual world (the internet and television instead of newspapers) – iconography and understanding the perceptions thereof becomes evermore important.
See results on how people perceive the United States, China and Japan – and cast your own votes – at Photocracy.org .
- Innovation , Social Media , Uncategorized
- Comments(1)
I was really surprised by some of the images, but more so by my personal responses. I tended to be influenced by somewhat cliche and romanticized images of the countries,always opting for the obvious when things got vague. Maybe that’s the point! Very cool site.