Psst! Wanna Buy a Rolex?
Measuring The Effect of Low-End Marketing on Brand Equity
Brand equity can be a company’s most valuable asset, especially for a company with luxury brands. Our goal in this demonstrator study was to see if controlled exposure to cheap Swiss watches had any impact on average-consumer opinions about the luxury Swiss-watch brand Rolex. Our approach involved an experiment conducted on smartphones. We measured brand attitude and price perception for the Rolex brand with and without exposure to a treatment consisting of advertising copy about cheap Swiss watches. We found that exposure to the treatment had a measurable and surprising effect: respondents exposed to the treatment had lower estimations of the price of a Rolex watch and also better perceptions of the brand. One possible explanation suggested by this data is that exposure to cheap Swiss watches actually made the Rolex brand seem more accessible to average consumers, thus accounting for their more positive attitudes towards the brand. Beyond the details of this particular study, our overarching purpose was to show how a dedicated panel of mobile game players can be used in controlled experiments to measure the effects of marketing messaging.
Our study was motivated by a common dilemma for companies with luxury brands. They often would like to extend the market for their products by offering low-end, cheaper versions of their high-end, expensive offerings. However, in doing so they risk damage to their brand equity by appearing to cheapen the brand and reduce its exclusivity (Stegemann, 2006). To make an informed decision, brand managers need to be able to measure the effect on consumer perception of low-end marketing messaging. This is difficult to do in a rigorous, quantitative way: how can you control exposure to marketing messages for a large nationally representative sample of respondents and query them for their opinions? In this case study we show how a mobile game app with a nationally representative sample of dedicated players can be used to quantify the effects of low-end marketing on consumer opinions about a luxury brand.
Our study collected data on perceptions, opinions, and associations about Rolex watches from age and gender re-weighted samples in the US and UK over a period of three weeks. The respondents were players of our game app The Pryz Manor. This is an app for Android devices that was specially designed to reveal underlying attitudes and awareness about brands and products. Participants download our game from the Google Play store and play at their leisure: over 4 million games have been played to date (April 2015). Players can win prizes for the quality and quantity of their game play. They are aware that the purpose of the app is to gather market data. 80% of game play is just for fun; the remaining 20% of market-research game play is seamlessly interwoven into the same game mechanics and structure as the fun content. The games are described below.
Game mechanic: Slice of Life is a choice-based game that reveals true preferences. The player selects a preferred item from a set of 2-6 items, and then predicts how all of the players will split on the choice. The player who predicts most accurately wins his preferred item. Gameplay variables recorded: Individual preferences among the items presented and estimates of item popularity. Behavioral dimensions uncovered: The game reveals true, behavioral preference since players choose the item they really want because they might actually win it.
Use in this study: The game was only used as a way to expose players in our US and UK panels to advertising copy for cheap Swiss watches.
Game mechanic: Name Dropper is a real-time, two-player word-guessing game. One respondent gives clues to another player to help him guess a topic. The clues cost different amounts in game time, which incentivizes the players to choose truly differentiating clues. When used to gather market data, the topics are typically brands, and the clue words are attributes indicative of brand awareness and attitude. Gameplay variables recorded: Relative selection of the awareness and attitude attributes; order of attribute selection; speed-to-guess of the topic; and guessing rank order within a competitive set of brands. Behavioral dimensions uncovered: Brand attitude; brand awareness; top-of-mind response; conceptual map. Use in this study: The relative selection percentages of phrases in the game indicate brand attitude and awareness for Rolex.
Game mechanic: Speed Stampede is a real-time four-player game involving the matching of images with words. In a first phase each respondent chooses a text descriptor that best describes an image. Images can be photos, images of products, brand logos, etc. Then in a second phase all players race to work out which of the images in a large set were matched with the four descriptors from the first phase. Gameplay variables recorded: Players’ selected phrases or adjectives in response to a visual stimulus; time to make the selection. Behavioral dimensions uncovered: Awareness of marketing taglines, spokespeople, etc.; brand knowledge; brand attitude; and attitudes to new product concepts or ad copy. Use in this study: The visual stimulus was an image of a Rolex watch, and the text descriptors were a mix of positive and negative attitudinal phrases that could be applied to the Rolex brand.
Game mechanic: The Salon is a modular survey where questions on different topics are posed each day. The questions can be open-ended or multiple choice, and can include rich media. As with all our games and activities, there is a mix of fun questions and market-research questions. Gameplay variables recorded: The question answers. Behavioral dimensions revealed: Standard survey metrics. Use in this study: We asked one question about pricing for Rolex watches.
In addition to using the games for inquiring about brands, we also use the games to collect demographic and psychographic data. This data is combined with US and UK census data for the players’ locations to generate an accurate and comprehensive segmentation of our player panels.
Recall that our goal in this study was to see if exposure to cheap Swiss watches had any impact on opinions about Rolex watches. The treatment was a Slice of Life game in which players were offered a choice between a cheap Swiss watch worth $120 (£80) and a cash prize of $40 (£25) – see the game description above. Players in our US and UK nat-rep panels were exposed to this treatment. (Although not immediately relevant to our study goals, of the 1,522 players given the treatment, 52% chose the cash, with the remainder splitting between the men’s watch (25%) and the women’s watch (23%).
Half of the players in the study were probed about their attitudes to Rolex watches from January 11-21 2015; the treatment was presented to all players on January 22 2015; and then the other half were probed from January 23- February 2 2015.
Table 1. Average responses to a question in The Salon about the estimated price of a Rolex watch.
So how did the treatment change attitude and pricing perception for Rolex watches? We used a question in The Salon to have a randomly chosen subset of players estimate the price of a ‘genuine Rolex watch’ – see above in the description of The Salon activity. Exposure to the treatment had a pronounced effect on both panels (see Table 1): the US mean estimated price dropped by 28% and the UK mean estimated price dropped by 38%. Median prices dropped even more. (Price estimates below $50 and above $499,000 were excluded as outliers, resulting in 24 values being deleted from the US panel and 11 values from the UK panel. The estimates were not normally distributed, so an analysis of statistical significance is beyond the scope of this short paper.)
We used the Name Dropper game to probe a separate random subset of players about the full range of their sentiments associated with the Rolex brand, with and without exposure to the treatment – see the game description above. Figure 1 shows the percentage of players that selected the different phrases with and without exposure to the treatment. The sample sizes were N = 188 without the treatment and N = 116 with the treatment, with data from the US and UK panels combined. Treatment exposure increased the selection rate for all phrases, indicating that respondents were more confident in their opinions. Positive and negative attitudes remained roughly balanced with treatment exposure, indicating that the treatment did not seem to do any net damage to general perception of the brand.
Figure 1. The selection rate for different responses to the game topic ‘Rolex’ in the Name Dropper game. Red indicates treatment exposure, and blue indicates lack of treatment exposure. Positive phrases are shown on the left, and negative phrases on the right. Responses chosen by less than 10% of respondents are not shown.
We used the Speed Stampede game to identify the predominant sentiment from a given list that another separate random subset of players associated with an image of a Rolex watch – see the game description above. (One can think of Speed Stampede as inquiring about a player’s top attitude, and Name Dropper as inquiring about all attitudes that apply.) Figure 2 shows the selection percentages for the different phrases with and without exposure to the treatment. The sample sizes were N = 150 before the treatment and N = 100 after the treatment, with data from the US and UK panels combined.
Figure 2. The selection rate for different responses to an image of a Rolex watch in the Speed Stampede game. Red indicates treatment exposure, and blue indicates lack of treatment exposure. Positive phrases are shown on the left, and negative phrases on the right. Responses chosen by less than 5% of respondents are not shown.
In terms of dominant attitudes, the treatment seems to have engendered a more favorable attitude about price, an enhanced sense of exclusivity around the brand, and reduced negative stereotyping of people who wear Rolex watches.
Exposure to a treatment featuring cheap Swiss watches had a measurable effect on attitude and price perception for the Rolex brand. Respondents exposed to the treatment had lower estimates for the price of a ‘genuine Rolex’. However, exposure to the treatment does not seem to have done any damage; instead of cheapening the brand, it seems to have improved attitudes toward Rolex. This might be because the treatment made Swiss watches in general and Rolex in particular seem more accessible, therefore causing respondents to revise their attitudes positively to the brand.
This was only a demonstrator study: the sample sizes were small, the experimental design was simple, and the analysis was straightforward. However, it is an example of the kind of field experiment that can be done with a panel of highly engaged mobile game players. Because the games are entertaining, the players remain engaged for a long time: the most-active players in our panels play hundreds of games over weeks or months. Thus we can expose large, nationally representative samples in a controlled, quantitative way to a variety of experimental treatments to see how their opinions are shaped by various marketing messages.
Stegemann, Nicole. “Unique brand extension challenges for luxury brands.” Journal of Business & Economics Research (JBER) 4, no. 10 (2011).