By Linchi Kwok
The advance of technology and the widespread adoption of smartphones and handheld devices in recent years have enabled us to publish our experience about a product or service through online photo or video sharing and provide a review.
Online review websites have also updated their features, making it easier for consumers to attach pictures or videos to their reviews. As both consumers and businesses adapt to the new photo-sharing trend, it becomes crucial to expand our knowledge regarding user-generated photos’ (UGPs’) effect on online reviews.
An empirical study about user-generated photos
I worked with a research team in an interdisciplinary project to assess UGPs’ effects on the helpfulness of hotel reviews. We published our findings in the Journal of Hospitality and Tourism Research. The paper is entitled “Let photos speak: The effect of user-generated visual content on hotel review helpfulness.”
We drew from the media richness theory and advanced five hypotheses for statistical analysis, including:
- Online reviews attached with UGPs have significant positive effects on review helpfulness.
- Online reviews attached with UGPs containing more guestroom objects are rated as more helpful than those containing fewer guestroom objects.
- Online reviews attached with UGPs containing more F&B (food & beverage) objects are rated more helpful than those containing fewer F&B objects.
- UGPs’ positive effect on review helpfulness is stronger for lower-priced hotels than higher-priced hotels.
- UGPs’ positive effects on review helpfulness are stronger for negative reviews than positive reviews.
The data and the analysis
We used a Python-based web crawler to collect the data needed for hypothesis testing from Qunar.com, one of the largest online travel agent (OTA) sites in mainland China. The data covered 12,138 hotels in Beijing, with about 1.16 million valid reviews and over 464,000 photos published by close to 900,000 users between January 2014 and April 2018. In this sample, about 9.84% (114,000) reviews were attached with UGPs.
We adopted YOLOv3 (You Only Look Once version 3), a real-time object detection algorithm, to identify the guestroom and F&B objects in UGPs. The top 10 guestroom objects identified include bed (15.34%), chair (14.55%), TV (6.23%), sink, toilet, couch, clock, remote control, refrigerator, and laptop.
The top 10 F&B objects included bottle (8.45%), cup (4.39%), bowl (2.82%), dining table, microwave, spoon, wine glass, oven, cake, and knife. Compiling with other numerical data, such as the number of helpfulness votes a review received, star rating of a review, etc., we then tested the hypotheses in a series of linear regression models.
Our analyses confirmed the user-generated visual content’s positive effects on review helpfulness. Moreover, consumers rated UGPs with more product-specific images more helpful than those with fewer product-specific images (guestroom or F&B objects for a lodging product in this case). Such a positive effect becomes more salient for lower-priced hotels (than higher-priced hotels) and reviews with lower ratings (than reviews with higher ratings).
What do the research findings mean?
Besides this study’s theoretical contributions, the research findings provide a few specific practical implications for hotel managers, web admins managing online review platforms, and the consumers relying on online reviews for decision-making. Here, it is imperative to note that the following actionable suggestions are exclusive content only available in this viewpoint article but not in the original journal publication.
- Strategically respond to selected reviews with UGPs.
- Respond to most, if not all, reviews with more guestroom objects in UGPs.
- For hotels with various F&B offerings, respond to the reviews with more F&B objects in UGPs.
- It is unnecessary for hotels with limited F&B offerings to pay attention to the reviews with F&B objects in UGPs.
- For hotels of a lower price, make every attempt to answer reviews with UGPs.
- Make sure to respond to reviews of lower ratings and with UGPs.
Web admins in online review platforms
- List those consumer reviews with UGPs at the top, allowing easy access for potential customers.
- Promote the reviews with more guestroom objects in UGPs, regardless of how many F&B offerings a hotel has.
- Promote the reviews with UGPs containing more F&B objects for the hotels with various F&B offerings.
- Cross-list or promote the reviews with more F&B objects in UGPs for the hotels even when internet users search for restaurant reviews.
- It is unnecessary to promote reviews with F&B objects in UGPs for the hotels with minimal F&B offerings.
- Highlight the reviews with UGPs, especially for the hotels of a lower price.
- Display the ones with UGPs first when internet users want to check out the negative reviews.
Consumers using online review websites
- Pay attention to the information conveyed in the reviews with UGPs in general.
- Pay even more attention to the reviews with UGPs that show more guestroom objects.
- Quickly skip the online reviews with UGPs showing more F&B objects if a hotel has limited F&B offerings or if the traveler does not plan to use the F&B services offered in the hotel.
- Pay attention to the reviews with UGPs that show more F&B objects only when the traveler also wants to use the hotel’s F&B services.
- In a search for hotels of a lower price, make sure to check out the reviews with UGPs carefully.
- When browsing through a hotel’s reviews for decision-making, pay special attention to those reviews of lower ratings and with UGPs.
Although our work is not without limitations, our analysis with an integrated analytical model that incorporates both econometric analysis and image-processing techniques yielded additional insights about user-generated visual content’s effect on online reviews. Once again, this study shows that a picture is indeed worth a thousand words.