The Impact of Magic Mirrors on Self Perception

Holition UX Series №4: Research Investigating Underlying Psychological Processes of Using Augmented Reality.

This research investigates how looking at yourself with augmented reality makeup on can affect self-esteem. We’ve looked into the entire experience of using a magic mirror with AR makeup, including the influence of factors such as personalisation feature and user expertise within the experience.

The research was presented at the leading international venue for consumer behaviour — Association for Consumer Research ’17 — in San Diego and published under the following title: Ana Javornik and Marta Pizzetti (2017) “Mirror Mirror on the Wall, Who Is Real of Them All? — The Role of Augmented Self, Expertise and Personalisation in the Experience with Augmented Reality Mirror” in NA — Advances in Consumer Research Volume 45, pages: 423–427

Index

1. Introduction
2. The self, make-up and augmented reality
3. Methodology. Study 1: Less isn’t necessarily more. Study 2: Experts and personalisation
4. UX Design implications

1. Introduction

Augmented reality (AR) simulations of make-up products provide an opportunity for consumers to see what they would look like with products on their faces without physically trying them. Through AR, numerous products can be virtually applied on a user with the push of a button. By removing the effort of physical application, the process of trying on make-up becomes more immediate and exploratory.

As well as making the process of trying make-up on easier, AR mirrors also reflect a different kind of self. However, little is known about how consumers perceive themselves when observing their augmented reflection in magic mirrors. Also less known is how do other features such as personalised looks affect consumers’ self-perception and their purchase behaviour?

2. The Self, Makeup and Augmented Reality

Augmented reality product simulation happens in real-time, fusing and syncing the real and the augmented self. Viewing an augmented image in the mirror can potentially lead to a modified perception of the self — at least for the time of the interaction. The crucial question is — how much does our opinion of ourselves change when looking at our reflection through an AR magic mirror?

Makeup augmentation can vary in its intensity. It can be only partial, such as lipstick covering merely one part of the face, or it can be holistic, overlaying the whole face with full virtual make-up. Gestalt theory has shown that the perception of the whole is different from the perception of the sum of singular parts. Such a seemingly small detail — viewing a virtual lipstick as part of a full makeup virtual look — can therefore provide a substantial difference in consumer perception.

Rather than seeing an augmented product in isolation, a holistic combination of products (a.k.a. a makeup look) allows consumers to observe the stimuli in a more realistic context, making it more believable. We therefore hypothesized that makeup augmentation is more effective when holistic (e.g. a full makeup look) rather than as a single augmentation product (e.g. only a lipstick visualisation).

Hypothesis 1:

Holistic augmentation (full AR makeup look) has significantly stronger effects on self perception than partial augmentation (AR lipstick).

However, makeup preferences are personal and subjective. Consumers do not like different make-up looks equally — rather, they often wish to personalise their makeup to achieve the desired effect that is aligned with their individual taste. The same goes for magic mirrors; not every holistic augmentation is perceived as the desired augmentation of one’s face and subsequently of the self. The ever-growing desire for a tailored experience and the importance of personalisation has been reflected across many industry trends. However, as shown by numerous research studies, a crucial component in the process of personalisation is personal expertise. We thus expected that the individuals’ expertise in makeup will have a significant effect on the positive perception of holistic augmentation of a self.

Hypothesis 2:

Personalised holistic augmentation (full makeup look) leads to a more positive perception of the self-augmentation. These effects are observed with consumers that possess an expertise in make-up.

And finally, how does augmentation of the self link to the actual product? Abundant marketing research has shown that consumers can develop a sense of psychological ownership of a product without actually owning it. Such psychological ownership significantly affects purchase behaviour. If positive perception of the self-augmentation is established, one would expect this to be associated not only with the virtual simulation of the products, but also with the actual products. In other words, seeing virtually overlaid products that a consumer likes would create the sense that the individual already ‘owns’ the products — psychological ownership. As choosing a full makeup look requires a certain level of expertise, we would expect for this process to appear only among those that possess expertise in makeup.

Hypothesis 3:

Personalised holistic augmentation (full makeup look) positively affects self-perception leading to psychological ownership of the products, which further leads to purchase intentions — for the consumers that possess expertise in make-up.

3. Methodology

Two experimental studies were conducted in a lab setting with female participants at two large British universities. In both studies, participants were asked to interact with a mirror interface built by Holition. When observing themselves in the magic mirror, participants were able to explore augmented reality makeup on their faces and pick the one they liked the most. Participants were asked to complete two questionnaires, one before the interaction and one after. The self-reported measures were: self-esteem, psychological ownership, purchase intentions (of tried-on products) and makeup expertise. All the responses were measured on a scale ranging from 1 (the lowest value) to 7 (the highest value).

Study 1: Less Isn't Necessarily More

51 participants were randomly assigned to one of the two conditions. In both conditions, their task was to choose a lipstick that they liked the most. The main experimental group had a pre-set full virtual makeup overlaid on their face, while the control group only saw virtual lipstick. In both conditions, users were asked to browse through a wide range of virtual lipsticks and choose one that they liked the most.

Self-esteem

Results showed that users experienced significantly higher self-esteem when seeing their face fully augmented in comparison to when only one part of the face (only lips) is overlaid with a virtual product (Figure 1).

Ownership

In addition, users experienced significantly higher psychological ownership of the virtual lipstick when the full makeup look was overlaid, as opposed to when only the lipstick was overlaid.

Figure 1 ComComparison of change in self-esteem in the two groups after trying on lipstick:

— “Lipstick Only” where the virtual lipstick was tried on without any other virtual makeup

— “Full Makeup” where the virtual lipstick was tried on as part of a predesigned makeup look * Self-esteem change was calculated as a difference between self-esteem measured before and after lipstick try-on

Purchase

However, overlaying a pre-designed look on the participants’ faces caused issues. Not everyone necessarily liked the makeup look to the same degree, and the same makeup looked very different on different faces. Indeed, three participants in the full makeup condition did not like the full makeup at all. After excluding those from the analysis, we found a significant difference between the two groups also in terms of their purchase: the participants that were choosing their lipstick while wearing a full augmented reality makeup reported significantly higher purchase intentions for the lipstick (Figure 2).

In comparison to the partial augmentation, the holistic augmentation (full makeup look) thus had a significantly stronger effect on the positive perception of self, product ownership and product purchase. One key factor that impacted this was the users’ makeup expertise. We then investigated further how different ways of creating holistic makeup and makeup expertise affected this process.

Figure 2: Reported intentions of purchasing the lipstick (on a scale of 1–7) across the two groups: — “Lipstick Only” where the virtual lipstick was tried on without any other virtual makeup

- “Full Makeup” where the virtual lipstick was tried as part of a predesigned full makeup look

Study 2: Experts and Personalisation

In the following study we asked participants to find a complete makeup look (composed of lipstick, eyebrow enhancement, foundation, blush, eyeliner and mascara) that they liked the most.

95 participants were split into two conditions. In the non personalised condition, participants had 13 pre-designed looks that they could try on to find the one they liked the most. In the other condition, they were able to personalise their looks by selecting different virtual products. The participants also differed in terms of their level of makeup expertise. The experimental design was thus composed of 4 groups (Figure 3).

Figure 3: Four experimental groups — participants with low or high make-up expertise tried on personalised or pre-designed makeup

Expertise

As in Study 1, we measured self-esteem before and after the augmentation. The self-esteem did not differ significantly across the four groups prior to the exposure to AR stimuli, but after trying on the full AR makeup, the participants across the four groups reported significantly different self-esteem levels. The combinations of expertise and personalisation conditions thus led to different effects on self-perception.

Those with high makeup expertise and a personalised AR makeup experience reported the highest increase in self-esteem after the augmentation. Surprisingly, those with a low level of expertise reacted very differently — they even reported a slight decrease in self-esteem when personalising the full look (Figure 4).

Figure 4: Graph showing change in self-esteem across the four conditions

* Self-esteem change was calculated as a difference between self-esteem measured before and after lipstick try-on

Ownership

Importantly, the results confirmed the findings from the first study in terms of psychological ownership of the products. The positive perception of the self (i.e. the increase in self-esteem) that participants experienced after seeing themselves augmented was shown to lead to psychological ownership of the products, which in turn increased purchase intentions.

This shows that due to the increase of self-esteem, augmented reality make-up elicits feelings of being an actual owner of the virtually tried-on product. Such ownership facilitates the intent to purchase. As anticipated in Hypothesis 3, this process was observed only for the consumers with makeup expertise.

Figure 5: Model showing the effects of full virtual makeup on self-esteem, which further affects the psychological ownership of the virtually tried products and finally, impacts the purchase intentions of those products

4. UX Design Implications

Augmented reality magic mirrors offer more than the ability to try products. As well as visualizing products, they offer a user the platform to create and explore their own image. However, this process can in turn affect consumers’ self-esteem. This change in self-perception also depends on different application features that suit user profiles differently. The design of the AR mirror experiences should take these variations into account.

The Magic of the Full Look

Even if consumers are only looking for a single product, the experience is more impactful with a complete augmented reality visualization. Magic mirrors should offer default or generic looks that can serve as a basis or canvas for single product try-ons.

Celebrating the Self

Mirror functionalities should not be limited to simply ‘trying on’ the product, but also focus on the experimentation with self. Design features should thus support a positive perception of self-image.

Experts See it Differently

Mirror experiences should be dynamic and cater to users with varying levels of expertise. For users with low expertise, more support is crucial in the form of predesigned options and product combinations. Personalisation, on the other hand, is more important for those with high expertise and can foster higher positive perception of the self and in turn product purchase. Knowing your audience and adapting the experience to them is key.

Connecting the Dots

Given the significant link between self-esteem, psychological ownership of products and purchase behaviour, easy routes to purchase should be integrated in the try-on experience and be especially prominent at the points when the positive self-augmentation is particularly salient.

Authors:

Ana Javornik, PhD

Marta Pizzetti, PhD

Anisa Motala, MSc

Ana Maria Moutinho, PhD

Publication Design:

Yasmeen Ayyashi

Technical support:

Ana Maria Moutinho, PhD

Maria Jose Garcia Sopo, MSc

Russell Freeman, PhD

Academic support:

Prof Yvonne Rogers

Prof Anthony Steed

Funding:

UCL Knowledge Exchange and Enterprise Funding EPSRC IAA

Newcastle University Business School

Holition Ltd

2019

@Holition