MAKEUP AI - Machine learning

The Company

A makeup and beauty retailer founder by the number 1 make up artist in the world. Working to break ground within the beauty industry, using innovative technology for customers and make up artists.

THE goal

To create a personal connection between Charlotte Tilbury (the world’s number 1 Makeup Artist), and each and every Charlotte Tilbury customer.

THE PROBLEM

As a woman or man who wears makeup, it can be hard to know what colours and shades suit me best.

The customer

  • Women and men who wear makeup

    • In LA

    • In London

Makeup AI is the first of it’s kind as a highly personalised recommendation and virtual try on tool, using; facial recognition, machine learning and augmented reality. Makeup AI intelligently predicts which makeup products will suit you the best based on your skintone and facial features.

Discovery

Sketching and ideation

Makeup AI was born out of a conversation with our founder who wanted to extract all her expert knowledge and share it with everyone and anyone who had an interest in makeup.

“Bottle me up and put my brain in a computer”

The initial discovery was geared around understanding two things;

  1. Understanding Charlotte’s breadth and depth of knowledge, and unveiling her secrets

  2. Understanding the underserved needs of our customers and identifying where the pain points are when using and choosing makeup

  3. Conducting competitor analysis and market research.

1 UNVEILING CHARLOTTE’S SECRETS

  • Ask the experts interview sessions with Charlotte and her professional artistry team

  • Ethnographic studies with Charlotte, to see her processes during a makeover. Understanding her way of working was key to us being able to emulate this process through a digital experience

  • Ethnographic studies with Makeup Artists in Charlotte Tilbury stores, who are trained in the Charlotte way and also deal with clients on a day to day basis.

Guerilla style user interviews

2 UNDERSTANDING THE CUSTOMER

  • Assumption gathering using the hundreds of women we have working at Charlotte Tilbury head office. Running group sessions to understand their behaviours and validating at scale with a survey to all staff

  • Guerilla style user interviews with non Charlotte Tilbury customers to gain breadth of understanding across different customer groups

  • Customer interviews with out target users; in the US and London. Deep dive sessions to validate our hypotheses, identifying the most impactful area for enhancement and change within the current customer journey when a user; chooses, uses or buys makeup

  • Sharing sketches and conducting co-design sessions with users to start to visualise potential solution

  • Large scale surveys with existing clients to validate at scale, the direction for MVP.


3 COMPETITOR AND MARKET RESEARCH

There are many AR experience in the makeup industry, predominantly in the form of mobile apps. These experiences allow for virtual try-on and experimentation. At Charlotte Tilbury we had already developed a virtual try on experience in store, “The Magic Mirror”. It was important for us to understand these experiences in terms of; tech, expected customer behaviour, interactions, intentions and primary goal of each. This way, we could take learnings from existing experiences and enhance our own whilst layering on new capabilities around analysis.

Competitor analysis across experiences with differing intent; try-on, skin analysis, tutorials and in-store upsell.

This multi pronged approach to discovery allowed us to reduce the level of risk upfront and start to identify the MVP and initial design with a certain level of confidence.

Synthesis and idenitfying MVP

Our discovery unveiled 3 top issues for our customers when it comes to make up;

  • Choosing the right makeup

  • Knowing how to apply the makeup

  • A feeling of intimidation

We chose to focus our attention on helping customers choose the right makeup and removing the feeling of intimidation. Starting at the very beginning of their journey with make up also made logical sense. If we can help them to choose the right makeup, we can build up their trust in us as a brand, and can then overcome the problem of applying makeup later down the line.

Desired journey mapping

Other consideration at this stage included;

  • Identifying the desired journey

  • The technical viability

  • The device and experience context

  • The speed to market

  • The potential business value



Design, test and build

Following the discovery and ideation phase, we defined our MVP as “a private consultation with the world’s number 1 Makeup Artist”. Essential to the success was the immediate customer understanding that this was no normal AR try-on tool and in fact was using Charlotte’s Magic Makeup Algorithm to recommend the best products for skintone and facial features. Building out different flows and prototypes to test with customers meant we could validate design decisions as we went. In conjunction to this, it was important to maintain the essence of a Charlotte Tilbury consultation, therefore tying the customer need in with the brand identity.

Key challenges were around building trust between the user and the machine, not only in terms of the AR try-on looking realistic but also building a belief that products recommended to the user follow Charlotte’s algorithm.

Approach and final designs are still under embargo until the product launches in July.