What people say and what they do doesn’t always match, meaning that most market research findings should be taken with a grain of salt. So how can brands accurately predict which new products will perform best at retail? There’s a chatbot for that.
Claire, a Facebook Messenger bot designed to help businesses optimize products and ad campaigns through interactive surveys, is not a run-of-the-mill surveyor. Those only secure 17 percent completion rates. Claire, on the other hand, uses gamification to drive engagement, rewarding respondents with loyalty points and giveaways.
According to its founders, Claire is 10 time more engaging than the average survey and five times more accurate than traditional market research when it comes to predicting product performance.
“What makes us really unique is the machine-learning algorithms that we bring to the table,” explained Misha Laskin, co-founder of Claire, speaking Tuesday at Decoded Fashion NYC.
Laskin spent a lot of time analyzing data sets and measuring distant cosmic signals during his PhD in theoretical physics—and lo and behold it’s not that different from forecasting for retail.
“There are so many false positives, so much noise to filter through to get to the right signal, and so the job of a physicist is usually to write these algorithms that know where to search and what to listen to,” he said. “Just because a person says they want to buy something doesn’t mean they will. You have to know which users to listen to and how to distinguish good data quotes from bad ones.”
“We saw a huge problem in the industry that we were the right people to solve,” said Marta Jamrozik, a former consultant at Bain & Co. and Claire co-founder. “That problem is that more than 50 percent of new products fail. So $1.1 trillion are being left on the table every single year due to poor product decisions.”
Here’s how Claire is vying to change that. Brands email customers with a link to an interactive chat on Facebook Messenger. To encourage participation, brands might offer the chance to win a $100 gift card. Claire sends some instructions, asking questions like “Would you buy this dress?” and including photos that participants can zoom in on for a closer look. If the answer is yes, Claire asks why: was it the style, the color or the fit? The bot rewards consumers for playing by engaging them with points.
“The points system is basically related to your chances of winning that giveaway so we’ve seen that people find the points system so engaging that when we give them an opportunity to play another chat after they’ve finished the first one, 90 percent of people will opt in to the second one,” Jamrozik said.
Claire then analyzes their interactions to deliver accurate insights in order to optimize the products brands and retailers produce or buy and help them spend less time and money A/B testing ad campaigns.
“People oftentimes have a hard time understanding what they want until they see it in front of them. So if you show this kind of stuff to thousands of users, you’re very quickly able to cluster people into certain groups that resonate with this product, not this one,” Laskin said. “That’s really how personalization works. It’s not that this one person said yes or no across 50 things but it’s that the entire group that gravitates to that object seems to be resonating with it.”
The start-up’s client roster already includes Kohl’s and Rebecca Minkoff. In fact, when working with the latter brand to help identify which products would be bestsellers and which ones would perform worst, Claire’s predictions were 90 percent accurate—that’s three times better than the industry average.
“So not only are we delivering these very valuable business results but we are also providing an extremely engaging experience on the consumer side,” Jamrozik said.
“I think that chat is actually the very first step in a roadmap of what people are calling conversational commerce. You don’t have to be typing it out,” Laskin added. “I think that in general what’s happening with [Amazon’s] Alexa or the new product that Google released is that it’s very easy to interact with stuff via voice and so I think it’s that and augmented reality, where you can actually see products in front of you interact with your voice, that chat is really the first step in getting to that point.”