Written by Jess C
With Apple using face scans to unlock your iPhone X and Google incorporating FaceNet functionality to sort your Google Photos and tag people in them, it is evident that facial recognition has become a modern-day phenomenon on par with blockchain and artificial intelligence (Gemalto, 2018). While engineers and programmers are looking for new ways to improve the accuracy and effectiveness of this technology, marketers are more concerned with the ways it can be used to improve customer communication and personalisation. On the one hand, being able to identify consumers behind the device and even ‘read’ their emotions may be highly valuable for promoting goods and services that fit their mood and preferences (Yang et al., 2018). On the other hand, this raises multiple ethical challenges within the scope of the recent personal data protection regulations and the rights for privacy. Unfortunately, the field of facial recognition application in marketing is a relatively understudied area of marketing knowledge. This suggested the need to explore the emerging opportunities and challenges in this sphere that can be used by modern marketers.
Facial Recognition Opportunities
There are several ways face detection can collect consumer data for further processing. First, systems like Apple Face ID can make a 2D or 3D static ‘scan’ of facial features that are unique to a particular person (Newman, 2018). This data is widely applied for identifying users, which is convenient for maintaining security and preventing the access of unauthorised individuals to private accounts (see Figure 1). Second, this information can be processed in real time to identify consumer emotions (Hwangbo et al., 2017). This form of facial analysis may not require individual identification to be effective from the marketing standpoint. Marketers may choose to demonstrate specific advertisements to people experiencing joy, anger, or surprise. This can also be combined with geolocation analysis to promote local organisations offering specific goods and services. Finally, facial analysis may be beneficial for developing deep learning and big data capabilities (DeMers, 2017). The tracking of consumer reactions to certain advertisements or advertising methods, or customer communication channels can provide valuable feedback about the effectiveness of specific marketing instruments. This may be especially valuable considering the subliminal nature of the registered reactions.
Figure 1: Business Use of Face and Iris Recognition Technologies in the US and Europe in 2018 (%)
Source: Statista (2018, p.1)
Modern algorithms are also capable of ‘features extraction’ (Wang et al., 2017). This capability recognises consumers independently from their clothes, illumination intensity, or face expression. This is convenient for recognising visits to multiple stores of the same chain during different time periods to personalise marketing offerings. Similarly, such companies as Virgin Mobile are experimenting with eye movements tracking (DeMers, 2017). This instrument may be highly useful for identifying if the marketing message was actually received by the recipient. In many cases, the demonstration of an advertisement may not be sufficient as consumers may turn away, look in another direction, or simply hide the tab showing the information. Combined with emotion recognition, this instrument can greatly assist marketers in improving the quality of ads in the online environment (Yang et al., 2018). Highly personalised communication may be viewed more positively than generalised messages that presently raise substantial criticisms. As the demonstration and delivery of advertisements may not be fully avoided in the online environment, making them more focused and valuable may decrease their perceived intrusiveness.
Facial Recognition Adopters
Walmart can be recognised as one of the early adopters of the facial recognition technology (Anderson, 2017). The company has been experimenting in this field since 2012 and seeks to use this instrument to improve customer satisfaction and servicing standards. However, these intentions were criticised by both industry experts and regular consumers who viewed some of the suggested measures as obtrusive. Another example of facial recognition applications is the Whole Foods company utilising the Microsoft Kinect platform (Hwangbo et al., 2017). The ‘Smarter Cart’ solution recognises the gender, age, weight, and height of consumers and allows them to use voice commands. These capabilities are integrated into identification systems allowing them to create and share shopping lists online. Microsoft is planning to integrate emotion recognition in future versions of the software to identify behavioural patterns and offer better store layouts and customer servicing procedures.
Another sphere utilising the new marketing opportunities is the fashion industry (Sampaio et al., 2017). The augmented reality capabilities offered by facial recognition software allow companies to let consumers to obtain certain experiences without purchasing physical products and services. The Sephora Visual Artist solution relies on Augmented Reality mechanisms to let clients try different makeup options before visiting company stores (Sonsev, 2018). Besides the benefits of promoting specific products, this instrument also reduces staff workloads as consumers become generally aware of their needs and the solutions that can address them. Finally, luxury brands including Rebecca Minkoff start using this technology to address the tastes of millennial audiences (Arthur, 2017). These consumers frequently avoid personal communication with store assistants and seek to fully control the purchasing process. Facial recognition allows them to speed up the alternatives evaluation and check out without saying a word to anyone or having to use cash.
Alibaba demonstrated a different method of utilising the facial recognition system as a payment instrument rather than an analysis tool (Gilchrist, 2017). Consumers can smile while looking into a specialised screen to confirm their purchase. This measure further contributes to the ‘fully controlled’ purchase experience combined with additional security preventing the use of stolen credit cards. Additionally, this further reduces the number of salespersons workloads and allows offline stores to service more clients. Finally, the implementation of facial recognition into modern devices developed by Apple, Google, and Samsung indicates a new trend in personal accounts management (Arthur, 2017). The development of e-commerce has raised the public awareness of cyber-crime and identification theft. In the case of the studied instrument, hacking into another person’s account may be substantially more difficult due to the additional security provided by face recognition.
One of the primary problems identified by Henderson et al. (2018) was the difficulty of analysing the facial data of people having different ethnic origin. Customers with darker skin were more difficult to analyse in terms of both identification and emotions. While this drawback is most probably temporary and can be addressed through technological advancement, it can presently create problems in terms of intercultural communication and claims of discrimination. Another technical drawback is the difficulty of repeatedly identifying consumers wearing different clothes, makeup, or hairstyles (Bedeli et al., 2018). Face recognition accuracy largely relies on big data and deep learning while the majority of brands have not accumulated substantial data about their consumers’ facial patterns yet. This sort of information may also be difficult to obtain without prior consent due to the current regulatory controversies regarding the right for privacy. However, the collection of anonymised data about consumer preferences may be recognised as a less intrusive instrument serving the same purpose.
One of the key problems hindering the wide-scale utilisation of facial recognition technologies in marketing is the problem of data ownership (Palfy, 2018). Consumer-related information regarding the preferences, emotional states, or behavioural patterns of individual clients has to be stored on corporate resources in order to be available for processing and analysis. This may violate the principles of confidentiality and informed consent and may create substantial security problems in the case of data theft. At the same time, Norval and Prasopoulou (2017) appraised this problem from the standpoint of the framework of contextual integrity. According to their findings, the studied phenomenon does not violate the existing status quo as personal photographs are readily available to the general public, which is socially accepted due to the spread of social media. While some practices devised by marketers may be ‘borderline’ by their nature, they can be made acceptable if the opt-out option is provided to all customers.
The findings of this essay suggest that the facial recognition technology has substantial potential for marketers as a method of analysing consumer reactions, learning more about client preferences, and offering a more personalised purchasing experience (Gemalto, 2018). While specific regulations in this sphere are still being developed, the collection of such data may be viewed as socially acceptable, especially in the case of informed consent. The most obvious benefit provided by this instrument is the capability to study subliminal reactions to advertising messages that reliably indicate the effectiveness of marketing tools (DeMers, 2017). It can be recommended that modern brands should start implementing facial recognition in a non-intrusive manner via the following consequent steps. First, they can offer this functionality as a security measure for authentication on their online platforms. Second, they should implement augmented reality options if they are compatible with their products and services. Third, consumer consent for data collection should be requested to personalise advertising messages and offerings. It should also be added the current provisions in this sphere do not prohibit the collection of anonymised data, which may be used as the fourth measure to improve the overall effectiveness of marketing instruments (Hwangbo et al., 2017).
Anderson, G. (2017) “Walmart's Facial Recognition Tech Would Overstep Boundaries”, [online] Available at: https://www.forbes.com/sites/retailwire/2017/07/27/walmarts-facial-recognition-tech-would-overstep-boundaries/#5f47ee9145f8 [Accessed on 19 December 2018].
Arthur, R. (2017) “The Automated Future of the Fashion Store: Where Self-Checkouts and Human Touch Collide”, [online] Available at: https://www.forbes.com/sites/rachelarthur/2017/02/01/the-automated-future-of-the-fashion-store-where-self-checkouts-and-human-touch-collide/#5fdd28985d6c [Accessed on 19 December 2018].
Bedeli, M., Geradts, Z. and van Eijk, E. (2018) “Clothing identification via deep learning: forensic applications”, Forensic Sciences Research, 3 (3), pp. 219-229.
DeMers (2017) “Facial Recognition Could Drive The Next Online Marketing Revolution: Here's How”, [online] Available at: https://www.forbes.com/sites/jaysondemers/2017/11/27/facial-recognition-could-drive-the-next-online-marketing-revolution-heres-how/#5ef10b9f529a [Accessed on 19 December 2018].
Gemalto (2018) “The top 7 trends for facial recognition in 2018”, [online] Available at: https://www.gemalto.com/govt/biometrics/facial-recognition [Accessed on 19 December 2018].
Gilchrist, K. (2017) “Alibaba launches ‘smile to pay’ facial recognition system at KFC in China”, [online] Available at: https://www.cnbc.com/2017/09/04/alibaba-launches-smile-to-pay-facial-recognition-system-at-kfc-china.html [Accessed on 19 December 2018].
Henderson, G., Rank-Christman, T., White, T., Grantham, K., Ostrom, A. and Lynch, J. (2018) “Intercultural competence and customer facial recognition”, Journal of Services Marketing, 32 (5), pp. 570-580.
Hwangbo, H., Kim, Y. and Cha, K. (2017) “Use of the smart store for persuasive marketing and immersive customer experiences: A case study of Korean apparel enterprise” Mobile Information Systems, 1 (1), pp. 1-17.
Newman, D. (2018) “Facial Recognition Software: The Future Is Here”, [online] Available at: https://www.forbes.com/sites/danielnewman/2018/09/18/facial-recognition-software-the-future-is-here/#76b0b466299d [Accessed on 19 December 2018].
Norval, A. and Prasopoulou, E. (2017) “Public faces? A critical exploration of the diffusion of face recognition technologies in online social networks”, New Media & Society, 19 (4), pp. 637-654.
Palfy, S. (2018) “Five Reasons Why Passwords Aren't Going Away Any Time Soon”, [online] Available at: https://www.forbes.com/sites/forbestechcouncil/2018/09/04/five-reasons-why-passwords-arent-going-away-anytime-soon/#22a0f8007b7d [Accessed on 19 December 2018].
Sampaio, J., Zonatti, W., Mendizabal-Alvarez, F., Rossi, G. and Baruque-Ramos, J. (2017) “New Technologies Applied to the Fashion Visual Merchandising”, Modern Economy, 8 (3), pp. 412-429.
Sonsev, V. (2018) “How Sephora Makes Beauty A Two-Way Conversation”, [online] Available at: https://www.forbes.com/sites/veronikasonsev/2018/04/12/how-sephora-makes-beauty-a-two-way-conversation/#1ee214157f51 [Accessed on 19 December 2018].
Statista (2018) “Business use of face and iris recognition technologies in North America and Europe as of 2018”, [online] Available at: https://www.statista.com/statistics/938679/north-america-europe-business-use-face-iris-recognition-technologies/ [Accessed on 19 December 2018].
Wang, H., Hu, J. and Deng, W. (2017) “Face Feature Extraction: A Complete Review”, IEEE Access, 6 (1), pp. 1-39.
Yang, D., Alsadoon, A., Prasad, P., Singh, A. and Elchouemi, A. (2018) “An Emotion Recognition Model Based on Facial Recognition in Virtual Learning Environment”, Procedia Computer Science, 125 (1), pp. 2-10.