Executive Summary

After analyzing 3080 observations from 35 advertisements and 88 participants, we determined that L’Oréal’s ad performance performed poorly when tracking the attention-grabbing effectiveness of advertisements and memory retention of magazine viewers. To improve these factors, L’Oréal should focus on placing ads at the end of the magazine, increasing pictorial size, and placing their ads on the right-hand side of the newspaper.

Introduction and Background

Our team was asked to study pictorial and brand information based on eye-tracking technology. Tracking the eye movements of consumers is beneficial because it lets researchers know which aspects of an ad consumers focus on the most; researchers can then determine the effectiveness of an ad. Since vision focus is quite narrow, we want things to be scanned by the center of the retina to be clear to the beholder. We are given an eye-movement test data set that contains 3080 observations, including 88 participants who are asked to recall 35 different brand ads shown in a targeted magazine conducted by Dutch eye-tracking research company. We believe that loreal currently has a disadvantage for target consumers to memorize the brand compared to average.

L’Oréal is the world’s largest cosmetic company that has developed activities in the field concentrating on skin care, sun project, make-up, perfume. To keep the leadership, massive advertising to targeted consumers would be especially beneficial to them. Dependent variables include attention (brand and pictorial) and memory (recall accuracy) meanwhile independent variables include the size of ad elements, page position (right vs left), and page number.

Data and Methodology

The data is collected by simulating people browsing through magazines at home or in a waiting room. We analyzed eight variables in total: page number, page position, brand fixation, pictorial fixation, brand size, picture size, recall accuracy, recall time (Figure 1).

Two types of generalized linear models are used. Poisson model is used for events where the outcomes are counts where binary logit model is used when to predict a dichotomous dependent variable based on one or more continuous or count independent variables. In SAS Studio, we used Poisson regression models to measure the effect of ad’s design and placement to fixation count. In the brand fixation model, brand size and page position were selected to identify how the brand element size and ad location affect the attention to the brand element. Binary variable L’Oréal was introduced to compare its effectiveness to other brands after ad exposure. In the pictorial fixation count model, steps were repeated, only changing the dependent variable to pictorial count. Binary logit model was applied for recall accuracy. The first model showed how ad location, brand and pictorial fixation count may influence odds of recall accuracy. The second model explored effects of design and placement of an ad to odds of recall accuracy with 5 independent variables: brand size, picture size, page position, page number, and L’Oréal.

Key Findings

We found page number is not a statistically significant factor in brand fixation count model (p=0.46) (Figure2). Thus, page number is removed from the model. After removal (Figure3), brand fixation count is expected to increase by 6.7% when brand size increases by 1 square inches, brand fixation count increases 82.6% when repositioning the ads at the right-hand side (Figure 4). Variable “L’Oréal” has a negative effect on brand fixation count, indicating 67% less brand fixation after exposure (Figure 5). In the pictorial fixation count model, pictorial fixation count increases 1.8% when pictorial size increases l square inches. Pictorial fixation count increases by 11.3% when the picture is placed on the right-hand side. The pictorial fixation count goes up by 0.15% as page number increases 1 (Figure 6). The L’Oréal brand also has a negative influence on pictorial fixation count, with 25% decrease after viewing the ad. (Figure 7).

Looking at the binary logit model for recall accuracy (Figure 8), the odds of recall accuracy (p_i/(1-p_i )) increase 9.1% for additional 1 count of brand fixation and 2.6% increase for additional 1 count of pictorial fixation. The odds of recall accuracy go up by 0.8% for 1 unit increase of page number and increases by 20.8% when ads are placed at the right-hand side (Figure 9). The model shows that brand fixation has a greater influence than pictorial fixation in increasing odds of recall accuracy. To further investigate brand and picture size’s effects and location on recall accuracy, we introduced a new binary logit model with brand size, picture size, page position and page number as independent variables (Figure 10). The odds of recall accuracy will increase by 3.6% with 1 square inch increase of brand size meanwhile will decrease 0.6% with 1 square inch increase of picture size (Figure 11) Page position and page number share the same positive correlation as the first binary model: placing ads on the right-hand side or at last pages is more advantageous (Figure 12). Exposure to L’Oréal will negatively affect the odds of recall accuracy with 72% decrease in this model (Figure 13).

Conclusions and Recommendations

In conclusion, L’Oréal has negative coefficients in brand fixation, pictorial fixation count and recall accuracy models, indicating L’Oréal ad performance is not as good as other brands in terms of attention grabbing and memory retaining. In order to gain more attention from viewers, we recommend increasing ad surface size and placing it on the right side of the magazine. Page numbers have positive impacts as it increases for pictorial fixation count and recall accuracy, so placing an ad at last pages (recency effect) will be more effective. When surface size is limited, Surface size for brand name needs to be prioritized over picture since brand name has larger impacts on both memory retention and attention grabbing. We also suggest allocating more budget when bidding for the right side of the magazine given its higher viewer attention and accurate memory.

L’Oréal’s future sales depend on actual purchasing behavior and overall purchasing amount. Our current study does not link them to recall accuracy and fixation count, further study needs to be conducted to investigate their correlation.