Online reviews: how to read between the lines
I’m currently planning a trip to Japan and seem to be spending an alarming amount of time on TripAdvisor. As any TripAdvisor addict will attest, trying to choose between hotels can soon result in a state of analysis paralysis.
Everyone has their own criteria when it comes to choosing hotels (a pillow menu isn’t really a deal-breaker for me), however, conventional wisdom dictates that our decisions should be based on whether the positive reviews broadly outweigh the negatives – right?
According to some recent research published in the Journal of Consumer Research, not exactly. While the overall sentiment of reviews is obviously important, our decisions are also being subtly influenced by how specific these comments are.
The study, quoted in Art Markman’s Ulterior Motives blog, highlights that when an experience reinforces a deeply held attitude, we tend to respond by making a general analysis of the brand or product’s overall quality. Conversely, when our experience contradicts our existing view we focus on specific product attributes:
For example, in one study, people try out either Bic pens (which people tend to think of as cheap and low quality) or Parker pens (which people tend to think of as expensive and high quality). The pens were manipulated so that they did not work. Then, they had to pick a statement that best described their feeling about the pen. For the Bic pen, people selected abstract statements (like “This Bic pen is bad”). For the Parker pen, people selected more specific statements (like “This Parker pen would not write”).
Art Markman July 5, 2010
So far, so straightforward – it’s common sense that our existing prejudices would play a role in the way we review products. More interesting in the context of social media is how this phenomenon affects the readers of these comments.
The next phase of the study focused on the influence comments had the purchasing intention of others who read them. It found that positive comments were more likely to motivate others to buy when they were abstract in nature (e.g ‘this pen is great’) as opposed to specific (e.g ‘this pen writes well’). Similarly, negative comments were less damaging to purchasing intention when they were specific in nature (eg ‘this pen writes badly’) than when they were more abstract (eg ‘this pen is awful’).
Just as reviewers do not consciously choose abstract or specific language based on their desire to influence, most of us do not consciously place more emphasis on reviews that are abstract as opposed to specific. However, even if we are oblivious to these nuances, the research suggests that the phrasing of these reviews can be highly influential.
This has implications for way we analyse attitudes to our brands online. Automated sentiment analysis tools – already compromised by their inability to detect sarcasm – fail to take into account the way language choices influence the reader. These findings suggest that by taking a deeper look at the way people phrase their reviews we might be able to gain some interesting insights into their broader attitudes about the brand or category and tailor our response accordingly.
The research is also relevant to the way we encourage people to talk about our brands online. More progressive companies are starting to aggregate social media mentions and reviews on their own sites – even when they these are critical about their products. While brands shouldn’t attempt to manipulate the content of these reviews, this research suggests that the way we frame the debate is worth some thought, particularly given that most decision making is based on minimising risk rather than optimising choice.