Cookie banner

This site uses cookies. Select "Block all non-essential cookies" to only allow cookies necessary to display content and enable core site features. Select "Accept all cookies" to also personalize your experience on the site with ads and partner content tailored to your interests, and to allow us to measure the effectiveness of our service.

To learn more, review our Cookie Policy, Privacy Notice and Terms of Use.

or
clock menu more-arrow no yes mobile

Filed under:

Your Next Beauty Recommendation Might Come From a Robot

New, 1 comment

Unlike SmarterChild, L'Oréal's chatbot wouldn't tell me I'm pretty.

Photo:  Innocenti

Racked is no longer publishing. Thank you to everyone who read our work over the years. The archives will remain available here; for new stories, head over to Vox.com, where our staff is covering consumer culture for The Goods by Vox. You can also see what we’re up to by signing up here.

Beauty brands, so often at the forefront of splashy and innovative ways to get us to spend money, want to reach us via robot. With Facebook Messenger’s addition of a chatbot component, brands have found an additional opportunity to whisper in your ear — and now, you’ll also be asked to respond. Beauty fans can ask a bot for the perfect foundation match or the best lipstick shade for a friend with green eyes. After it algorithmically calculates the best product for you, the bot presents the opportunity to purchase, like the futuristic version of a beauty counter employee. Who among us never asked AIM’s SmarterChild for a word of advice?

Over the past eight months, huge beauty brands like L'Oréal, Estée Lauder, and Sephora have developed Messenger bots that share beauty tips and recommend products based on your conversations with the A.I. If we’re already wasting time on there stalking an acquaintance from high school’s wedding photos, why not be vaguely more productive and restock our tinted moisturizer, too?

With 1.2 billion Messenger users, even a niche initiative like a bot still has the potential to reach hundreds of thousands of new customers. The huge user base means brand reps can’t (or won’t) narrow down the bots’ targeted demographic. There are bots that can help you with anything from a timed call to prayer to ordering a morning coffee to finding a Mother’s Day gift. According to Facebook, bots account for 2 billion messages each month. With over 100,000 bots on the platform, you can now find someone willing to talk to you about pretty much anything, even if that someone is artificial.

“People were already using Messenger every day to talk to their friends and family, so it’s only natural that businesses would come to where people are to drive their business objectives,” says Messenger product manager Allison Swope. Objectives vary bot to bot. Sephora Assistant, for example, wants you to book a makeover at a local store. Over the holidays, Estée Lauder offered to deliver a gift ordered over Messenger to your home within 60 minutes. Madison Reed’s bot “Madi” will analyze your existing hair color and offer suggestions (and products) for your next dye job.

It’s free for brands to host a bot on Facebook, says Swope. Developers can either submit one for Facebook’s approval, or partner with one of the in-house developers to create a bot on a brand’s behalf. Facebook’s “developer checklist” is straightforward enough: Bots should have a start button, a greeting, the ability to converse, and email notifications. Beauty bot conversations feel more like an especially personalized quiz than like the smutty back-and-forth I remember from SmarterChild. “We usually focus on bots that do one thing really well,” says Parham Aarabi, the founder and CEO of beauty tech company ModiFace. “Beyond that, absolutely everything in the experience and personality of the bot is customized.”

“A year ago, when we opened the platform, people were just thinking about bots as chatbots, like a robot that you can interact with through chat,” says Swope. At their core, bots are still robots that you can chat with. But they now also have real-world applications, from booking a makeover appointment to trying on virtual-reality lipsticks over a selfie of your face. Customer service is also a big application initiative — instead of a janky technical assistant in a pop-up box on a brand’s website, customers can complain to a more sophisticated bot inside the platform they already use, and get the relief of an instantaneous, if artificially engineered, response.

When I asked the Sephora Virtual Artist bot for coral lipstick options, I was offered 10 products priced from $17 to $54. Four of those 10 suggestions were Sephora Collection’s own lipstick. You can ask for more results or refine within your choices, like requesting a coral gloss or matte. The system isn’t perfect — when I wrote in “coral lipstick under $20,” I was given only two options, and both were over $30. If you’re looking for something very specific, it saves you a few clicks from browsing through pages of search results on the Sephora website — and it’s oddly satisfying to have the endless options narrowed down by a well-meaning robot. “Carmen” from L’Oréal helped me pick out a gift for my best friend, myself. The chat was like a choose your own adventure of “pampering” options, from skincare to bath products to breakfast in bed.

One snag: When I tried to go back to an earlier option, Carmen wasn’t having it. The conversation was a train and it was only moving forward. Unlike the bots of AIM past, these have little room for shenanigans: no one would tell me their a/s/l; I couldn’t get any to tell me I was pretty. “The goal is not to mimic human behavior, but to provide a useful service,” says Aarabi. “Often, this means making it clear that it is a bot and not pretending to be a human. So, for example, limiting answers to a few options, directing users on what they can ask, etc., are all ways of making the experience more efficient and successful.” I would call any chat filled with compliments both “efficient” and “successful,” but it turns out that bot doesn’t (yet) agree.

Brands are reaping the rewards of the ability to reach the entire Facebook-using world — a quarter of the world’s population. Most traffic to bots comes from the brand apps, Facebook pages, or websites and not from Messenger itself. (Facebook has an “Explore” page in the works to help people connect with bots more easily.) It follows that the customers engaged enough to like the Sephora Facebook page would also be intrigued by a Messenger bot. The Messenger version of Sephora’s reservation assistant has an 11 percent higher booking rate than options like calling the store, booking on Sephora’s actual website, or booking through a mobile app. “What that really shows is that if you go to where people are, you can drive meaningful business outcomes,” says Swope.

Beauty customers, who by now are accustomed to rapid-fire tech innovation, are already used to drawing inspiration from Instagram and YouTube. It’s a natural jump for beauty lovers to consume and discuss beauty content on another social media platform. Compare Sephora’s bot stats to those from the Absolut Vodka messenger bot, which, just like his human counterparts, will offer you a free drink if you spend a few minutes chatting him up. The bot led to a 4.7x lift in Absolut sales since its launch in October. Respectable, but not beauty-brand level.

Customers who have a successful chat can buy a product without ever leaving Facebook. “One of the capabilities allows you to pop up the checkout experience in the bot,” says Swope. “You never have to leave Messenger, but you can go through the checkout on the brand’s website.” Does that mean companies will be able to send you targeted emails about coral lipstick? “We really think carefully about those experiences,” is all Swope would say. I was expecting to get an email with the subject line “We Just Randomly Figured You Might Like Some Coral Lipsticks,” but I haven’t received any unsolicited follow-up from Sephora or any of the other branded bots.

As with most new technology, beauty bots feel removed from my needs. But that doesn’t mean customers won’t develop a new need for instant beauty product feedback now that we have the option. One day soon, we’ll all be buying our clothes on Amazon and our makeup on Facebook.