AI (Artificial Intelligence) seems to be the next big thing in many industries today. On Gartner’s 2020 Hype Cycle of Emerging Technologies, for example, we find no less than seven explicitly AI-related trends in the first steep curve of inflated expectations—such as composite AI, generative AI, responsible AI, embedded AI, and explainable AI. For a term that dates back to 1956 and celebrates its 65th birthday this year, this seems remarkable, especially since the productive application of the currently hyped AI variations is expected to take another two to ten years.
In this arena of promising AI technologies, the Dutch AI-based startup Lalaland is an interesting case. They have found a way to make AI work in a way that is both tangible and speaks to the imagination. Using AI technology, they are one of the front-runners that may change the online fashion industry and, arguably, make it more inclusive, sustainable, and profitable, thereby speaking to all three P’s of the Triple Bottom Line. I spoke to two of Lalaland’s founders, Michael Musandu and Harold Smeeman to find out what they do and how they do it.
Lalaland is a Dutch, Amsterdam-based tech startup that develops hyperrealistic AI-driven virtual models for e-commerce platforms. Founded in 2019, its founding team consists of three founders (Ugnius Rimša is the third) and they are with 16 people now. So far, they have attracted €625k of funding and won the Philips Innovation Award 2020, giving them the title of most innovative student start-up of the Netherlands.
Unlike one might assume, the company has nothing to do with the Academy Award winning 2016 musical romantic comedy drama La La Land. When asked for the origins of the name, Musandu explains “We wanted a name that triggers your imagination, since what we create is not real, or at least not in the traditional sense. We also wanted a name that forces oneself to form an opinion; you either like it or you don’t.”
As stated on their website, Lalaland “empowers e-commerce brands to capture the incredible diversity of humankind by generating artificial full-body fashion models.” This means that they create fully AI-generated fashion models that online retailers can use to replace their traditional living fashion models. The most important benefit is that, when shopping online, consumers can select a model that fits for example their own size, age and skin color and see how garments look on that model—rather than on the usually young and slim models that retailers currently use.
Competitive Edge and Impact
A relevant question in the midst of the AI hype is how Lalaland is different from others. With AI-generated images being their core technology, the difference from straightforward model photography and online fashion retail stores is obvious. But what makes them different from their main competitors?
As Musandu explains, when they started two years ago, it was their ability to do full body generation: creating an image of a human being from head to toe. Today though, Smeeman adds, it is their ability to show the user exactly what they want to see. Based on continuous development, testing, and learning over the past two years, they have developed an advanced understanding of what users want and how this translates into hyperpersonalized synthetic images.
In a world where online shopping has taken a further flight through Covid-19, the online fashion market is booming. According to Statista, fashion is the largest B2C eCommerce market segment. Estimated at US$525.1 billion in 2019, it is expected to grow further at over 10% per year and reach a total market size of US$1003.5 billion by the end of 2025.
A market of this size means that any impact that the use of synthetic imaging will have on eCommerce fashion companies, will be significant. According to Musandu and Smeeman, there are significant benefits to gain. Not only financially, be along all three aspects of the Triple Bottom Line (TBL): People, Planet and Profit.
People Impact of AI: Inclusiveness and Diversity
The most visible and obvious benefit of AI-based synthetic imaging is the diversity of models shown and the ability to personalize images to make them resemble the user’s themselves. Instead of just offering one picture of a single model, users can configure the models to make them look like themselves. While the extent of configuration will differ per eCommerce platform, something as “simple” as changing ethnicity has a direct impact on inclusiveness and diversity.
By adopting this AI-based technology, fashion brands and retail stores can be more inclusive in the way they sell and advertise their products. There is of course the practical benefit: if a customer sees how garments look on a model that looks like them, they can make a better choice. But there may be a long-term more important psychological effect too. When users see models that look like them, they are likely to feel more heard, seen, and respected, thereby gaining confidence. While it is too early to confirm whether such effect will be achieved using this technology, there is no reason to believe it won’t.
Oh, and in real life, Lalaland stimulates diversity and inclusiveness too. Their small team consists of eight nationalities and people of a wide variety of ages, colours, genders, and sizes.
Planet Impact of AI: Waste Reduction and Returns
Waste by returns is a major issue in fashion, and especially the online fashion industry. Every year, 2.3 billion kgs of waste is generated through returns. The reason is obvious: to be safe rather than sorry, people order their garments in various sizes and return the ones that don’t fit. The problem? The large majority of the returned garments aren’t resold anymore, but end up in landfill.
By better tailoring the “virtual fitting” online, Lalaland’s technology helps to improve the first-time right hit-rate and reduce the necessity to order multiple sizes. While the long-term effect still has to show, and return rates differ substantially between types of garments, Lalaland’s experience with its first customers has shown that return rates can drop from around 40% to around 30% in women’s wear—a 25% decrease.
The drop in returns does not only mean that less garments are wasted. It also implies less shipments—to the customer, and back to the retailer, thereby also saving on carbon emissions from transport.
Finally, while the evidence is not there yet, we can also expect an increase in actual usage rate of the garments. As research from the University of Manchester shows, about 12% of the clothes in women’s wardrobes are not used. Other reports show percentages as high as over 50% of unused clothes. Because of various psychological biases, when people have selected more carefully what they buy, they are also more likely to actually use what they have bought. Along those lines, a customer that has carefully selected the best fitting synthetic model when ordering a garment, may be more likely to actually wear the garment afterwards.
Profit Impact of AI: Cost and Conversion
Using synthetic imaging also has various financial benefits. In an industry where margins are thin, any percentage reduction in product returns has an immediate and substantial positive impact on profitability. The same applies to shipping costs, which are substantial in online retail where shipping is often offered for free or reduced rates. Typically transportation costs account for 7-10% of total sales revenues. Any reduction in sending products that customers won’t buy will therefore also have a direct impact on profitability.
Yet another source of cost saving is photography. Not using actual models, and not making actual photos, could cut the cost of photoshoots and post-production by up to 70%, according to Lalaland’s founders. Naturally, this depends on the quality of models and photos used, on the pricing level of the garments, and on the speed at which collections change. In any case, though, the cost savings can be substantial.
At the other side of the equation, using personalised synthetic models can drive sales as well. As Lalaland’s experience thus far shows, click-through rates may rise up to 140% and conversion rates show an increase of 15%. Part of this may be a result of the newness of the technology, which triggers customer’s curiosity now, but that may fade away once this technology has become mainstream. However, it is not hard to imagine that customer engagement and conversation will remain higher because of the more personalised experience offered.
Furthermore, by explicitly targeting customers that do not look like the typical models currently used at most online shops—which is the large majority of people—the market that is served is much broader. This means that the number of customers that feel they are actually served can dramatically increase by targeting them with personalised images.
Still in its early years, Lalaland’s true take-off still has to happen. But the signs are promising. Its first customers in the Netherlands are there: lingerie brand Sapph and Stieglitz women’s wear and, as of March, Wehkamp—one of the largest online retail stores in the Netherlands.
Furthermore, the application of Lalaland’s technology is not limited to garments only. It can be used for any industry in which human models are currently used. A good example is eyewear. With over 60,000 different faces in their library already, Lalaland is now working with a large retailer to use its technology for their sunglasses range. Accordingly, many other application areas can be thought of, making the application of their technology seemingly unlimited.