Effect of Product Images on Consumers ’ Evaluations in Online Shopping

We investigated how presentation order of product images influences consumers’ imagery processing and ultimately their purchase intentions. The between-subject experimental design was conducted with 2 (order of images) × 4(the number of images) × 2 (dress style replicates). The results showed that in one-image condition, participants were more likely to go through mental imagery processing when they saw an image evoking high mental imagery processing last rather than first, showing the recency effect. In the four-image condition, participants were more likely to be engaged in mental imagery processing when they were exposed to four images evoking high mental imagery processing before the four simple product images than for the reversed order, showing the primacy effect. This research provides theoretical and practical insights for both researchers and online retailers on the effectiveness of ordering product images according to the number of product images to elicit outcomes that are more positive.

primacy effect when a long series of information was presented. When many pieces of information are given to individuals, they can become tired of processing, and/or they become less sensitive to the information that enters later and may make only minor adjustments to their first set anchor (Hogarth & Einhorn, 1992). Further, Schnotz (2005) argued that the visual working memory has limited capacity to process and store visual information. Thus, as the amount of available pictorial information to process is increased, memory capacity is limited to process later information, resulting in primacy effect (Waugh & Norman, 1965).
H2. When the number of images are increased, consumers are more engaged in mental imagery processing, and consequently have higher purchase intentions when images evoking high mental imagery processing are presented first and simple product images are presented later than in the reversed order (i.e., primacy effect).

Method
The between-subject experimental design was conducted with 2 (order of images: image/s evoking high mental imagery before vs. after the simple product image/s) × 4 (the number of images: one, two, three, and four images) × 2 (dress style replicates), resulting in 16 conditions. For the image evoking high mental imagery, pictures with a relevant consumption background were chosen where a model was wearing a dress and accessories. For the image evoking low mental imagery, a simple product image of the dress on a mannequin against a plain white background was used. The four images included front, back, side, and detail/close-up views and were placed vertically. After reviewing the information, respondents were asked to answer questions about mental imagery processing (Yoo & Kim, 2014) and purchase intentions (Dodds, Monroe, & Grewal, 1991).

Results
In three-way ANOVA result, no main effect of dress styles and interaction effects between dress styles and other variables on mental imagery processing were found; thus, the data of two dress styles were merged. After merging the data, a two-way ANOVA was conducted. No significant main effects of the order of product images and the number of images on mental imagery processing were found. However, the interaction effect between image presentation order and number of images on mental imagery was significant (F (3, 299) = 3.82, p ≤ .01). Further, the planned contrast results showed that in one-image condition, participants were more likely to go through mental imagery processing when they saw an image evoking high mental imagery processing last rather than first, showing the recency effect (t = -2.30, p ≤ .05, MHigh imagery image first = 4.48, MHigh imagery image last = 5.17). In the four-image condition, participants were more likely to be engaged in mental imagery processing when they were exposed to four images evoking high mental imagery processing before the four simple product images than for the reversed order, showing the primacy effect (t = 2.26, p ≤ .05, MHigh imagery image first = 5.17, MHigh imagery image last = 4.50). For the two-image and three-image conditions, there were no significant image order effect. The PROCESS analysis showed that the interaction effect of the order and the number of image on purchase intentions was through mental imagery processing (-.54 < CI < -.14). When the amount of image information was high (i.e., four images), presenting images evoking high mental imagery processing first increased mental imagery processing more than presenting the images last, which increased purchase intentions (-.86 < CI < -.09). When the amount of image information was low (i.e., one image), presenting images evoking high mental imagery processing last rather than first increased mental imagery processing more, which increased purchase intentions (.18 < CI < .93). Therefore, the results support H1 and H2.

Conclusion
This research provides theoretical and practical insights for both researchers and online retailers on the effectiveness of ordering product images according to the number of product images to elicit outcomes that are more positive. By examining the effect of the amount of image information on image presentation order effect, which has been limited or missing in previous research, this study adds to the literature on image presentation order within the context of information volume, and demonstrated a moderating role of number of images that can influence the image presentation order. Further, based on the result of this study, online retailers should place an image evoking high imagery after rather than before a simple product image, when only two images are presented. In contrast, when many product images (e.g., more than six images) are presented, placing images evoking high imagery first rather than last is expected to increase intention to purchase and, thus, a potential for more profitable outcomes.