For Bengaluru resident Srishti Shekhar, the stay-at-home situation and her last year of school made her try something she had never done before: online consultation to solve her acne issues. “I had been to two dermatologists before coming across Remedico’s service on Instagram. The sign-up process was very easy and all I had to do was send a few photos and I had a treatment plan designed for me within a day,” says Shekhar.
Like Shekhar, thousands of Indians turned to the internet when going to a clinic seemed risky. By September, the number of internet subscribers in India had risen to 776.45 million, up from 718.74 million in December 2019—474.11 million of them are urban subscribers and 302.35 million, rural subscribers, according to a January report from the Telecom Regulatory Authority of India (Trai).
This promises a market that goes well beyond urban areas. The Indian cosmetics market was valued at $13.19 billion (around ₹96,000 crore) in FY2020. And beauty brands have begun the shift to e-stores; everyone wants a share of the burgeoning skincare and cosmetics market as the country moves indoors to battle a second deadly wave of the coronavirus. The click-to-shop shift is crucial for the skincare and cosmetics sector, where social media trends often determine what reaches the stores. Pre-bookings and online launches also allow brands to launch new products faster in a tough market. But how does an industry that depends on swatching colours, experiencing fragrances and analysing skin types in person, make its online experience immersive and accurate? The solution, believe firms, lies in incorporating Artificial Intelligence (AI), Augmented Reality (AR) and Machine Learning (ML) technology to elevate the simple click-to-shop mode. With caveats, of course.
Globally, firms such as M.A.C and Clinique had been using AI and AR to help customise products to skin type much before the pandemic. M.A.C uses YouCam’s AR technology for a virtual try-on feature for lip and eye colours. Clinique’s technology, in turn, was designed to help you pick the perfect foundation shade and formulation.
This move to AR and AI-based technology, now gaining speed, is linked to the increasing demand for customised products, says Avni Sricharan, co-founder of Hyderabad-based custom lip colour brand LipHue. Sricharan and Dyuti Waghray launched LipHue in 2019 with a store in Hyderabad where customers have the option of creating their signature shade. Of course, they couldn’t try lip colours on the lips, so the colour would often look different from the way it looked on the wrist.
In 2020, LipHue launched its AR-enabled e-commerce platform for custom lipsticks: Their camera helps you try on shades as you mix colours and build your unique shade. “Our AR technology lets you try the lip colour on the lips from the comfort of your home,” says Sricharan.
It’s a big change. Hyderabadis used to prefer walking into a store to swatch and get their product in 20 minutes. “(Since the pandemic) our online sign-ups increased and people who would visit the website would spend six-seven minutes trying on lip colours. Because of covid-19, we have a lot more online orders now compared to on-ground sales.”
The Bengaluru-based dermatology consultation platform Remedico Health integrated AI technology in 2020. “We recently built an AI-based selfie checkup tool for people with acne which allows customers to understand the severity of their acne before a doctor gets involved. It allows a patient to understand what grade their acne is, as well as approximately how long it would take for their acne to be successfully treated by one of our dermatologists,” says founder Ranjit Bhatia. The company says it saw a rise in the number of patients with adult acne due to excess stress in 2020.
He believes image-based ML tools are now able to detect and provide basic provisional diagnostic information to both patients and doctors. This, he says, can help improve workflow. “It can reduce the cost of treatment as well as allow a dermatologist to simultaneously track, monitor and treat many more patients than they might have earlier.”
Another Hyderabad-based company that has been exploring the nascent tech-driven skincare market in India since 2018 is SkinKraft. Like LipHue, which entered the field to connect with the premium segment (custom-made solutions), SkinKraft makes customers answer a set of questions on skin, lifestyle and other health aspects. It then maps the data points with the help of an algorithm to create a cleanser, moisturiser or serum designed specifically for your needs. “These questions are designed by a dermatologist, ones that you would be otherwise asked during a personal consultation. And just like you would have a follow-up appointment for the doctor to understand if the products are working for your skin, SkinKraft too has a follow-up questionnaire which will even change your product based on your feedback if it isn’t working for you,” says CEO and co-founder Chaitanya Nallan.
But you need to be careful, warns celebrity skin expert Harshna Bijlani, medical head of The AgeLess Clinic, Mumbai. Most skincare concerns, she adds, still need a combination of in-clinic treatments and complementary skincare products. Data privacy issues could arise too.
There’s another limitation. The algorithms are based on the principles of allopathy, so when it comes to traditional systems like Ayurveda, you may need to refer to the classical texts, says Nallan, whose company Vedix offers customised Ayurvedic solutions.
For Bengaluru'Avneer Sidhu, the follow-up option was the clincher. “The first moisturiser I received didn’t work well for my skin type. The team reassessed my skin and I had a few more simple questions to answer, and the next product worked well. This made me sign my husband up for a custom regimen too,” says Sidhu.
Data: the more the better
The quality of the results, whether they match or better a personal consultation with a dermatologist, depends on the quality and quantity of data. With systems like SkinKraft, where the customer, who isn’t a skincare expert, responds to questions, there’s always a chance of an error (aren’t most of us confused about our skin types?).
This is why clinical studies serve as the base rules and the quantity of customer data plays a role. The more the data, the better the prediction. The quality of data is key; and the data bank grows only if sales increase.
Firms continue to work with dermatologists to better their models. “Human error is one of the reasons why an ML algorithm takes longer to learn. When we started, one of the parameters we measured was a sensory match. After two years of data, our ML has now started beating the in-house dermat’s cleanser prediction by a slight margin. This also differs from product to product,” says Nallan.
The shift seems to be here to stay.
Dhara Vora Sabhnani is a Mumbai-based journalist.