When beauty brand Glossier wanted to create its first face wash in 2015, the retail-meets-tech company turned to its blog ‘Into the Gloss’ to crowdsource ideas. Founder and CEO Emily Weiss asked the online community: “What would your dream cleanser look like? Smell like? Feel like? Do for you? Not do for you? Who would play this cleanser in a movie?”

Weiss’ post lit up with 380 comments. Readers wanted a product that would leave their skin clean, without making it dry. It should have no alcohol and be environmentally friendly. The bottle should have a pump dispenser and it should be embodied by actresses Emma Stone or Julianne Moore - both redheads with milky complexions.

That’s how the ‘Milky Jelly cleanser’ was born. The pump bottle was advertised by Emma Stone lookalikes. The crowd-sourced product remains a bestseller and has since become a defining feature of a brand that now describes itself as a “people powered beauty ecosystem.”

Since its launch in 2014, the beauty-brand-meets-tech-company has gained a reputation for its innovative use of customer data. Glossier treats conversations on its blog and social media pages as strategic opportunities to listen to their audience - creating products or altering the user journey through their website in response.

According to Weiss, consumers are more than willing to volunteer their opinions or advice to Glossier. On Instagram - where it has 1.7m followers - Glossier receives five DMs each minute, with the overwhelming majority containing ideas. It also maintains a Slack channel, where several hundred of its “top customers” are invited to share their feedback.

Because these conversations are so important to Glossier, they maintain their brand-to-customer relationship by selling to their audience directly, not through middle men like Amazon or department stores.

But as the company’s popularity grows, Glossier is keen to scale its ability to listen. In the past, the brand has employed data analysts to sift through comments on articles, social media, user generated content and product pages. But Glossier wants to automate this process, using machine learning.

Speaking on the Recode Decode Podcast in January, Weiss said she had been concerned with the question: “How are we going to be able to kind of listen at scale to user feedback, or to people talking about what sucks about beauty, and what they want in beauty?”

The 33-year-old has decided the answer is to create Glossier’s own beauty focused social media platform - due to launch later this year.    

While Weiss wants to create a positive online space for women, where they feel empowered to talk about beauty products, she also wants a site that enables Glossier to listen to user feedback on a massive scale.  

“it’s just again about listening so we don’t have to manually go through every single comment and be like how many times did someone say “conditioning,” right?” Weiss told Recode Decode. “We can say, alright, people are really excited about conditioning. Are they happy with the products that they’re using or is there an opportunity to do something better?”