Brazil’s O Boticário cosmetics brand is launching two new fragrances specifically for millennials that have been developed through artificial intelligence.
Created in partnership with IBM, they are the result of a data-driven study by Symrise, a major producer of flavors and fragrances, that collected 1.7 million fragrance formulas – including scents sold to Coty and Estée Lauder.
This was combined with fragrance sales information, customers’ location and their age, human usage patterns and responses, to enable IBM to develop the scent AI tool called Philyra.
Philyra uses machine learning to create fragrance combinations that will match specific demographics. For Brazilian millennials, for instance, it suggests notes of fruits, flowers, wood, spices, and even caramel, cucumber, and condensed milk.
It’s not all algorithm however. The human element still exists on top with both perfumes ultimately tweaked by a master perfumer at Symrise to emphasize a particular note and improve how it lasted on the skin.
O Boticário is also known for being a brand that champions diversity and inclusion in their ads, so both AI-generated perfumes will be sold as genderless. “Fragrances are fragrances, and men and women should use whatever they prefer. We want to make our fragrance development process less bias,” said O Boticário’s marketing director, Alexandre Souza, to Exame’s publication.
Now, Symrise plans to distribute this technology not only to master perfumers but also to its Perfumery School to help train students. Beyond developing fragrances, IBM believes the technology can aid in other uses, like flavors, cosmetics adhesives, lubricants, and construction materials, as reported by Engadget.
The two O Boticário fragrances will hit the market on Monday, May 27.
How are you thinking about product innovation? We’re all about finding you the perfect partners to do so. The Current Global is a consultancy transforming how fashion, beauty and consumer retail brands intersect with technology. We deliver innovative integrations and experiences, powered by a network of top technologies and startups. Get in touch to learn more.
Google today announces the launch of an experiment in collaboration with innovation consultancy Current Global, to build-out a data analytics and machine learning tool powered by Google Cloud technology that will enable fashion brands to make more responsible sourcing decisions.
The initiative, revealed at the Copenhagen Fashion Summit, one of the fashion industry’s key sustainability events of the year, aims to focus on the raw materials stage (referred to as ‘Tier 4’ of the supply chain), providing brands with greater visibility as to the environmental impact of different textiles. The hope is to translate data into meaningful insights so the industry can take action.
Sustainability in fashion is a global environmental emergency. According to the United Nations Economic Commission for Europe (UNECE), the fashion industry accounts for 20% of wastewater and 10% of carbon emissions worldwide. The 2019 Pulse of the Fashion Industry report also shows the fashion industry is not implementing sustainable solutions fast enough to counterbalance the harmful environmental and social impacts of its rapid growth.
Current Global, an innovation consultancy that empowers fashion brands to reach their sustainability goals through the use of relevant technologies, analyzed where the industry’s largest environmental challenges are, and worked with Google to determine how it could help be part of the solution through the use of cloud-based tools for data collection and analysis.
What was identified was the need to focus on Tier 4, where brands have little to no visibility. This is an industry wide problem, where supply chains are highly fragmented, unregulated and with little transparency, yet where the majority of negative impact occurs.
Many organizations and brands have been trailblazers in an effort to collect and surface data that can lead to better sourcing decisions, but gaps in the data continue to persist due to its complexity and global nature. The aim of this experiment, is to bring together information in a way that will complement existing tools, consolidating and building on the data to shine a light into the furthest parts of the fashion supply chain.
“Lack of data in the fashion industry is one of the most pressing and complex issues we’re faced with. If you can’t see it, you can’t measure it, and if you can’t measure it, you can’t change it. In other words, without insights the industry is not empowered to make strategic and beneficial decisions for the sake of reducing their environmental impact,” Rachel Arthur, co-founder and chief innovation officer of Current Global, says. “We teamed up with Google to identify the strategic places within the supply chain that would benefit from its access to global data and its machine learning power to launch an experiment to create a decision making tool for the industry in order to enable a more sustainable fashion future. We know that if we could understand the nuance of the raw materials we source – information right now that is essentially impossible to accurately calculate – we could make an enormous dent into the overall composition of the clothes that are produced.”
To bring it to life, we’ll be collaborating closely with Stella McCartney on the first pilot project. This brand has been a pioneer in leading the fashion industry towards sustainability, helping to launch the UN Fashion Industry Charter for climate change and recently introducing Stella McCartney Cares Green, one of the arms of the Stella McCartney Foundation, to further promote sustainability and environmental protection.
As Kate Brandt, sustainability officer at Google, explains: “Stella McCartney has been a forerunner in the fashion industry embracing and leading the charge for sustainable fashion. At Google, we also strive to build sustainability into everything that we do whether that’s operating efficiency data centers to having our own Responsible Supply Chain Program. In 2016 we celebrated 10 years of carbon neutrality and we are the largest corporate renewable energy purchaser in the world. Outside of Google, we aspire to build tools to help people understand the planet, improve environmental impact, and take sustainable action. This pilot with Stella is a great step in the fashion industry’s bid to become more sustainable.”
The tool will use data analytics and machine learning on Google Cloud, focused on sources that allow companies to better measure the impact of their raw materials, relevant to key environmental factors such as air pollution, greenhouse gas emissions, land use and water scarcity.
To start, it will look at cotton and viscose, each chosen due to the scale of their production, data availability and impact considerations. More specifically, cotton accounts for 25% of all fibers used by the fashion industry, with a notable impact on water and pesticide use. Viscose production is smaller but growing in demand, and has links to the destruction of forests—some endangered—which are critical in mitigating carbon emissions.
The goal is not only to be able to determine the impact of producing these raw materials, but also compare the impacts of these in different regions where they are produced. This pilot will enable us to test the effectiveness of the tool on these different raw materials, building out the possibilities for expansion into a wider variety of key textiles in the market down the line.
Ian Pattison, customer engineering manager for Google Cloud UK, says: “Google’s mission is to organize the world’s information and make it universally accessible and useful. The challenge facing the fashion industry is one of information – taking fragmented and somewhat incomplete information and quickly translating it into meaningful insights to take action. In this case, understanding how fabrics are grown or made, what impact different sourcing decisions has on the environment, and ensuring that data is visible across the whole supply chain. Google’s 20-year leadership of data technologies, cloud computing and machine learning capabilities, coupled with our commitment to sustainability and our unrivalled global mapping, means that we are uniquely placed to work with the brands to address the challenge of reducing the environmental footprint of fashion.”
This is the first phase of the experiment. Google and Current Global are now actively working with further fashion brands, experts, NGOs and industry bodies with the ambition of creating an open industry-wide tool, and plan to continue driving collaboration with other key players—large and small.
The hope is that the experiment will give fashion brands greater visibility of impact within their supply chain and actionable insights to make better raw material sourcing decisions with sustainability in mind.
Adds Maria McClay, industry head of fashion and luxury at Google: “We have been hearing increasingly from clients, our industry partners and consumers the growing urgency around the fashion sector to make a dent in their negative environmental impact given the magnitude of the problem. If nothing changes, what is at stake is our future and that of our children’s. Google empowers its teams to find moonshots – really difficult, complex problems to solve where our technology can help make a 10x contribution, not just a marginal improvement. We believe that this could be our moonshot for the industry.”
How are you thinking about your sustainable innovation strategy? Want to learn more about how we worked with Google? The Current Global is a consultancy transforming how fashion and consumer retail brands intersect with technology. We deliver innovative integrations and experiences, powered by a network of top technologies and startups. Get in touch to hear more.
Why is someone who blew the lid on the Facebook data scandal talking about fashion? Canada-born Wylie was studying trend forecasting at the University of the Arts London while working at Cambridge Analytica, and has spent much of his career exploring links in culture.
Much like fashion trends, politics is cyclical, and encompasses the idea of presenting an aesthetic, or narrative, he explains. “Trends are just as important in politics as they are in fashion; just that rather than an aesthetic trend, it might be an ideological, behavioural or cultural trend,” he says. “You need to keep track of all kinds of trends in politics because you need to know if you come out and say something, what the adoption of that will be six months down the road. And is that going to help you win an election.”
Given the nature of his role at a data business, unsurprisingly he also has a big view on the impact of artificial intelligence and machine learning on the fashion industry too.
Fashion’s intuitive nature is not only hard to measure in trends, but also incredibly complex for machines to learn, he explains. He believes fashion is contextual because its trends and aesthetics are hard to quantify. “How do you define bold? If I go to a black tie dinner but I’m decked out in camo, I’m actually wearing quite muted colours, but it’s quite bold,” he exemplifies, saying boldness is contextual, depending on where you are.
“How do you tell a computer about that? Amazingly, with fashion, is that not only is it visually and aesthetically really enriching, computationally and mathematically it’s really hard. It’s a hell of a lot harder than politics.”
Computer vision could be the solution, he says, because an outfit is essentially visually-encoded information. In that sense, it is human beings who would need to look at pictures of people wearing clothes and choose the relevant adjectives that describe what they look like. They then need to work closely with computers to teach them about fashion.
“Everybody needs teaching, even computers. You learnt fashion in the first place, so the computer needs to learn fashion from people as there are no machines that know fashion yet,” says Wylie.
We saw this movement towards focusing on the human earlier this year at SXSW. While every conversation was underpinned by the concept of artificial intelligence, the topic kept highlighting the sense of instilling humanity in all interactions – from robots learning from humans, to humans being freed from minimal tasks to focus on what matters.
Another strong theme from SXSW – and one permeating consumer trends full stop today – is around the lack of trust in society. The Edelman Trust Barometer has reported a straight-line decline for 25 years, and Wylie likens the rise of ironic fashion such as Vetements to this too. “If you have a lot of designers who are starting to make stuff that is ironic – or stupid like the €200 DHL t-shirt – and people are buying it, it’s because you have a total collapse of trust in institutions, including fashion institutions,” he says, adding that this is where fashion and culture in general have a lot more power than they give themselves credit for.
The Vogue Italia interview otherwise covered Wylie’s involvement in the Cambridge Analytica and Facebook data scandal more broadly, and exactly why he decided to go public with the information.
For more on the future of data regulation and privacy, listen to our episode of TheCurrent Innovators podcast with Amnesty International’s Sherif Elsayed-Ali.
There’s a very simple filter that comes with working at Google, and it’s about putting the user first, says Tomas Roope, creative lead at Google Zoo, the tech giant’s think tank focused on pushing the limits of creativity through technology.
Talking to Rachel Arthur in a live recording of TheCurrent Innovators podcast from the FashMash Pioneers event in London, he said: “The way we think is always user-first. Are we really solving something for somebody here? …At Google we’re about solving problems at scale.”
While Roope admits some are more PR or headline-driven than others, his process, whether the result incorporates buzzworthy terms like augmented reality, artificial intelligence or beyond, always comes back to whether the solution is something that answers a consumer need. “What shifts the bottom line is making things more relevant, and making them simpler. [It’s about answering] what do people really want?” he asks.
Anchoring much of that work these days however, comes data. “[At Google], we have seven to eight products that have over one billion users monthly, and so we have a really great understanding of what people are doing… and what they’re thinking,” he explains.
That insight is what informs the work his team does as a result, while machine learning (ML) then takes it to the next level, Roope notes. He refers to ML as an area that’s not yet being explored to its full potential.
“We’re in the middle of two massive revolutions – one of which is still the smartphone coming from 10 years ago, and now the rise of machine learning.” He refers to this as not only a powerful and extraordinarily interesting tool that allows you to fix problems in a way you couldn’t have done before, but as the most exciting underpinning to the future we’re currently building.
It’s completely reshaping what our world looks like, and what opportunities there are for brands in it as a result, he explains.
To get there, he says experimentation for all industries – including fashion and retail – is key. “For me, you’re not going to sit and discover the future by dwelling on it… it’s all about test and learn,” he explains.
As to where it will take us, he adds: “There’s a great quote by Bill Gates that says we tend to overestimate what’s going to happen in two years, but underestimate what will happen in the next 10. If you look back 10 years, we didn’t have smartphones, but in two years nothing’s happened. Only when we look over a good chunk of time do we see how much it’s changed.”
Catch up with all of our episodes of TheCurrent Innovators here. The series is a weekly conversation with visionaries, executives and entrepreneurs. It’s backed by TheCurrent, a consultancy transforming how consumer retail brands intersect with technology. We deliver innovative integrations and experiences, powered by a network of top technologies and startups. Get in touch to learn more.
US performance label Ministry of Supply has launched an intelligent heated jacket that uses machine learning to adjust the garment’s temperature.
The Mercury jacket creates a microclimate optimized to the wearer’s body by using a custom microcontroller heating system to heat up carbon-fiber heating pads sewn in the garment’s lining. The system takes in to consideration the weather and body temperature, motion data, and user preference to modulate power. For example, when walking to a train stop the jacket senses temperatures and an elevating heart rate, as well as user behaviour learnt through time, to regulate the system.
The machine learning element ensures that the more feedback the user gives its accompanying app, the better the system gets at learning their preferences. Meanwhile an added voice element allows wearers to naturally activate the jacket through a smart assistant like Amazon Alexa.
“Our mission is to invent clothing that blends form and functionality — and temperature regulation is one of the most important factors in comfort,” says the brand’s team. “We’re excited to present our vision of what wearable technology can become, not just a way to monitor our vitals – but also act on it allowing us to become more comfortable and capable because of it. The Intelligent Heated Jacket is just that literally putting a learning thermostat in your jacket.”
Since Ministry of Supply’s inception, it has approached clothing through a human-centric, design-led methodology that takes into consideration both aesthetic and function. The jacket has been developed to replace any other outerwear alternative.
This is Ministry of Supply’s third successful Kickstarter campaign. In 2012, it launched the Apollo shirt, which controls body temperature after raising over $400K. Following that, the Atlas socks, which are made out of coffee beans that filter out sweat, raised over $200K or its $30K goal.
Imagine this: You walk into your favorite store and the sales associate welcomes you by name. She or he lets you go about your business, but on-demand shares with you which of their latest products you would most likely be interested in.
Such recommendations, powered by artificial intelligence, are a very familiar experience online these days, but they’re also increasingly being worked towards in the brick and mortar retail world.
A multitude of different technologies lie at the heart of achieving this, but namely it’s a connection between CRM and machine learning, all with that layer of identification placed on top to deliver results for the specific customer in question.
Your mobile device usually plays a key role in making the ID part possible, but facial recognition is another such way.
Lolli & Pops, a candy store based in the US with roughly 50 doors, is one such retailer experimenting with this. A proof of concept called Mobica, which is powered by Intel, was on show at NRF’s Big Show in New York this week. Using computer vision, it’s a facial recognition loyalty scheme designed to drive VIP consumer engagement.
The opt-in experience (shoppers literally have to enrol their face to be a part of it), means anyone entering the store is recognized in real-time by an app the sales associates are using on their tablet devices. From there, they are able to tell the individual’s taste profile, know for instance if they’re allergic to peanuts, and be able to personally recommend great products to them via AI-enhanced analytics accordingly.
“It’s designed for their loyalty shopper, so about wanting to make them feel really special,” Stacey Shulman, Intel’s chief innovation officer for its Retail Solutions Division, told me. “Privacy isn’t an issue because they have such a strong relationship with their customers and are trusted by them already. It all starts with service and a connection to the customer.”
You can easily imagine the same VIP concept being applied at the likes of Sephora for beauty, or even in an apparel merchant.
Other facial recognition technology on show at NRF enabled special, personalized deals to surface on screens in real-time, demonstrated a restaurant that allows customers to pay by face, and also touted broader data collection opportunities around demographics and store-traffic patterns.
It was the customer service piece that felt particularly pertinent however. As Shulman explained: “Technology today needs to not be at the forefront. It needs to be the helper at the back. When done right, it enables people to get back to the customer and back to what’s important. That’s what we see here; it’s not about the facial recognition or the AI, it’s about the experience the customer then has. The differentiator between a brick and mortar store and Amazon today is customer service. We can’t compete on price and selection anymore, so we have to go back to service. If we don’t we will have a problem.”
The Lolli & Pops facial recognition initiative will roll out to stores in the coming weeks, according to Shulman.
Young fashion shoppers today are demanding personalization more than ever. According to an IBM study, 52% of female Generation Z would like to see tools that allow them to customize products for themselves.
This coincides with an ever-increasing expectation for speed in delivery of product. While several fast fashion retailers can get product to shelves in weeks, the majority of clothing items take anywhere from six to 12 months of development.
Technology is impacting throughout the supply chain to shift this forward, including in the creative process itself. Artificial intelligence (AI) for instance – incorporating computer vision, natural language understanding and deep learning – is being used to produce key insights on trends to both expedite the initial design process and better predict demand for hyperlocalized products.
IBM has teamed up with Tommy Hilfiger and The Fashion Institute of Technology (FIT) Infor Design and Tech Lab on a project called Reimagine Retail to demonstrate this. The aim is to show how AI capabilities can give retailers an edge in terms of speed, and equip the next generation of retail leaders with new skills using AI in design, according to Steve Laughlin, general manager of IBM Global Consumer Industries.
To do so, FIT students were given access to IBM Research’s AI capabilities including computer vision, natural language understanding, and deep learning techniques specifically trained with fashion data.
Those tools were applied to 15,000 of Tommy Hilfiger’s product images, some 600,000 publicly available runway images and nearly 100,000 patterns from fabric sites. They then brought about key silhouettes, colors, and novel prints and patterns that could be used as informed inspiration to the students’ designs.
Product recommendations for e-commerce sites are not new in concept, but the suggestions they present to shoppers are increasingly getting smarter thanks to the algorithms behind them.
And the result of delivering more relevant product ideas? Higher spend of course. When Jewelry.com partnered with omnichannel personalisation technology firm, Dynamic Yield, to integrate personalised product recommendations on its website, for instance, it saw revenue increases per visitor of 39% from the homepage, 13% from product pages, and 18% from cart pages.
The key, according to the team, was not just to focus on the usual ‘most popular’ or ‘similar to current item’ suggestions, but instead to turn to machine learning to automatically select the most effective strategy for each user.
That meant finding a personalisation strategy that would work for both visitors with a rich history of behavioural interactions, and those who are new to the site, thus for whom minimal information is known. Doing so is about capturing signals from shoppers about their buying intentions and preferences for specific products as they move through the sales funnel, the Dynamic Yield team explained, and then providing upsell and cross-sell opportunities throughout.
“Traditional retail is beginning to have what we like to call a ‘moneyball moment’ where the old way of simply making gut decisions on which experience to serve your customers is being challenged. As machine learning technology becomes more advanced, algorithms will outperform humans in recommending products that users are more likely to show an affinity for, and ultimately buy,” Mukund Ramachandran, CMO of Dynamic Yield, notes.
“With Dynamic Yield, we can use machine learning to make data-driven recommendations based on where visitors are in the sales funnel. The ability to assess the level of valuable information about each visitor and automatically serve the most effective strategy has empowered us to increase revenue across our site,” said Jon Azrielant, director of marketing at Jewelry.com.
On the homepage, for instance, the Dynamic Yield widget leveraged affinity-based recommendations, recommending products according to a weighted score of what returning users had added-to-cart, viewed, or purchased in the past. To induce engagement among new visitors, the widget presented products with the highest amount of page views and click-through-rate on the site.
“While ‘point solutions’ for deploying product recommendations have existed in the market for decades, these solutions are limited by data silos that restrict their algorithms to only making decisions based on a user’s interactions with product recommendation widgets. With Dynamic Yield’s unified data stack, information onboarded from all onsite interactions, third party data, CRM data and loyalty data can be ingested,” Ramachandran explains.
On the product pages, the team has been running A/B tests, comparing 45% of users who were recommended products ‘similar to the current item’, 45% who were recommended ‘top-selling’ products, and 10% who received?a control variation. As a result, Jewelry.com revealed that recommending ‘top-selling’ products provided a 10% uplift compared to the other variations.
Finally, an additional widget was introduced on the bottom of the cart page to showcase items frequently bought together with the current item.
“These results are very strong compared to industry benchmarks. We think this is the case because with Dynamic Yield product recommendations are only part of the puzzle. The entire site starts working better for you – the homepage engagement is higher which leads more people to discover the most relevant products as they browse,” Ramachandran adds.
Topman’s global digital director, Gareth Rees-John, took to the stage at Shoptalk Europe this week with a welcome reminder of the things it’s possible to do without huge budgets.
He noted how many retailers are still operating on legacy systems with “jumbled data” making it hard to move forward fast, and said his focus is on “making little changes that have robust business cases”.
The key, he said, is about doing things the retail board will understand – referring to Sir Phillip Green as an owner that is becoming increasingly tech savvy but still at his roots a traditional shopkeeper – and said it’s about nudging people along.
He highlighted three simple ways his team is personalising the e-commerce experience for shoppers in order to help drive conversions.
The first is dedicated to students. A simple switch at the top of the website, facilitated by SaaS company Qubit, enables users to toggle all products to student prices – a 10% discount. “Normally we see 38% of spend on the website is with students, when we do this then we see 50%, so it’s huge – just by taking the friction out,” Rees-John explained.
The second he said is about personalising the website based on geography. “We see trends in the data as to what people are buying and where. Sterotypically, for instance, we don’t sell as many coats in [the northern city of] Newcastle – it’s all lads in short sleeve shirts – compared to in the south-east.” So the website is set up to over-show on categories where they do sell.
The last pulls in artificial intelligence: Canadian company Granify helps optimise Topman’s conversion rates by serving different messages to shoppers when they are at flight risk. The notifications use machine learning to address issues that will help retain the individual in question, such as letting them know an item is low in stock, as one example. It’s seeing an uplift of 3-5% in doing so.
Long-term Rees-John is looking to streamline the creative process for personalised content. “One of the biggest barriers to personalisation is the creative output – dynamic ads have their limits and if you have multiple segments then you need multiples of artwork. Our view is by the end of the year to have six different modules on the homepage and every person will see them in different orders but only see three at one time.”
The result will equal 720 different permutations of the website. “It isn’t a big data exchange it’s just a different experience of the brand going forward,” he said.
If time is the greatest luxury for modern consumers, Nordstrom is steadily proving that convenience is one of the foremost things it can offer its shoppers.
The department store is expanding its Reserve Online & Try In Store service to over 40 stores nationwide, following the success of its pilot in six last year.
The premise, which is built around making it easier for customers to shop in the way that they want to, enables app users to select items they like, then book to have them set in a fitting room for them in the store of their choice, ready to try on in person. There is no commitment to purchase at any stage.
“Many of our customers like to feel and try on clothes and shoes before they purchase them and we’re excited to offer them a more convenient way to do so,” says Shea Jensen, senior vice president of customer experience at the company. Read the full story, including further insight from Jensen, via Forbes.