For Your Eyes Only: Lingerie’s Supply Chain Secret

How data-savvy Adore Me is giving inventory management a new look?

Netflix. Amazon. Spotify. 18,000 titles. 32 million books. 100 million songs. All three brands dominate through unmatched choice. They don't just provide the most popular content – they have the most content – and that makes them the most dependable.?

There’s no real equivalent in fashion retail. That’s because hosting countless pieces of digital content may be technically demanding – but is logistically very simple. On Amazon, six in ten sales are fulfilled by third-party merchants , which removes much of the risk around holding stock. As for its distribution network, reducing Prime delivery from two days to one cost the company $800 million in a single quarter , more than double the entire market value of the fashion retailer ASOS.

A book is not available in multiple titles. A song is not available in multiple recordings. But in clothing retail, size, color, style, and fit variations can be put together in endless combinations. Add in omnichannel selling, home delivery, and returns, and inventory management quickly becomes an operational minefield.?

Plus-size, minus-profits

Today, the industry is rightly looking to embrace inclusivity, and 85% of U.S. clothing retailers now return online searches for plus-size options. But actions speak louder than search results – and only half of these will take customers to actual ‘plus-size’ items.?

Despite 70% of U.S. women regularly wearing a 14 or above, in 2021, just one in five clothing items sold was plus-size. Last year even saw major brands further scale back their size ranges . Retailers are waking up to a harsh reality: stocking plus-size is hard. It’s not only logistically difficult, but doing so poorly can harm business .??

Lingerie brand Adore Me is an outlier. Stocking an impressive 67 different sizes, customers can be confident almost any item they see on a hanger will fit them. With six outlets, an online store, subscription services, international shipping, and available in both parent company Victoria’s Secret and other major outlets, the brand is a rare success story for the fashion long-tail. With a thriving creator community built on body positivity, the medium-sized B-Corp boasts 1.5 million online followers.?

It’s not just social media; Adore Me’s tech-savviness is the engine behind their commercial success. In 2017, their 73-robot-strong smart warehouse was featured in MIT Technology Review. Today, the same facility processes up to 20,000 online orders every 24 hours. They’ve used Generative AI to free up nearly 40 hours a month for staff writing product descriptions. Executives have even pioneered the use of data to track supply chain emissions through their partnership with Carbonfact – reducing the footprint of their Spring 2024 collection by 12%.?

Love at First Byte?

Behind all of this has been a long-held belief in the value of data science. This love for modeling (software, not the runway) as a driver of sustainable growth has seen them make huge strides in addressing global fashion’s $500 billion problem : waste.?

It’s an inability to accurately forecast product demand – and prevent waste from overstock – that has led so many brands to turn their backs on plus-size. In an industry with tens of thousands of product size combinations, even the most meticulous sales planning will leave retailers with excess stock.?

Adore Me turned to Pendulum in 2023 to build on their existing work solving this very data challenge. With a decade spent developing solutions for national security and global health, swapping hardware for hosiery may seem like an interesting pivot for an AI company – but if anything, it shows just how universal the principles of data science can be. Whether it’s critical medicines in Central Africa, U.S. warfighters on the frontlines, or lingerie stock in New Jersey – effectively integrating multi-channel and multi-modal data is a powerful force multiplier wherever the problem space is.?

Pendulum deployed Predict\Products – an AI-powered demand forecasting tool – for the fashion retailer. The software builds on Adore Me’s efforts leveraging machine learning to detect patterns and trends across their product lines that would otherwise be totally invisible to conventional ERP solutions.?

Leveraging insights from across over 160,000 different SKUs, the solution provides their data science team with over 5,000 unique, highly accurate forecasts every single week. They can use these to make much more informed decisions about how they allocate this stock across their distribution channels and reduce potential overstock across this huge number of size and fit variations.?

Today, Pendulum’s software produces these weekly forecasts for Adore Me’s online store, as well as their VIP channel – where recurring customers can access special offers and discounts. In real terms, this means significantly more accurate insights into how much of each line item will be sold in any given week. These forecasts will soon also integrate directly into the retailer’s inventory planning systems, where sales and operation planning teams can also easily access them.??

Leveraging thousands of unique forecasts that span 67 individual sizes every week is impressive enough – but Adore Me has taken this machine-learning technology one step further. They have also integrated these forecasts with Plan\Products – Pendulum’s AI-powered optimization solution.?

Working with their already impressive data infrastructure, Plan\Products automatically allocates Adore Me’s stock-on-hand between their online store and VIP offering, as well as their Elite Try-On boxes. By holistically distributing across these channels, it optimizes to reduce unsold inventory based on each of their relative demand.?

These optimizations are not only slashing waste. Plan\Products is also helping them prevent avoidable stockouts – where available inventory would be allocated to the wrong channel. This has the potential to reduce missed sales from improper allocation by 16% – a further revenue opportunity for a retailer that is already proving the value of applied data science.?

Together, these two solutions are helping to power an incredibly complex omnichannel sales ecosystem a decade in development. Thousands of recurring customers have the ability to try on outfits at home – all either selected by themselves or chosen uniquely for them, all while Adore Me keep their commitment to inclusivity, and all while reducing product waste. A remarkable feat of in-house engineering.?

Less Waste, More Revenue.

Sustainability and profitability shouldn’t be competing priorities. Investing in ethically sourced materials, recyclable packaging, or greener transport is often perceived as coming at the expense of top-line growth – better inventory management isn’t.?

Whereas most software solutions will optimize with just one KPI in mind – here, operational AI is doing both. By preventing these current and future missed sales opportunities through misallocation – Pendulum’s artificial intelligence is both fulfilling its core purpose of decarbonizing the supply chain – and making a tangible commercial difference.?

Adore Me’s online store and subscription boxes are just the start. Predict\Products long-term forecasting models are also currently being deployed to more accurately inform their wider product procurement strategies. That means they can reduce overstock not just across their existing inventory – but also directly at the source. Both solutions are also set to be rolled out to prevent missed sales and waste for Adore Me products on other third-party marketplaces. This collectively stands to cover almost every channel across their distribution landscape.?

Improvement starts with a single number. What you send – and where you send it – changes everything when deployed at scale. Look no further for a better example of this. Adore Me prove that supply chains shouldn’t be a barrier to inclusivity and sustainability – and that applying machine learning across complex operations can simultaneously benefit people, products, planet – and bottom lines.?

Pendulum’s AI demonstrates that it doesn’t matter whether it's lifesaving drugs or lingerie, when you use the right tools, supply chain management is ultimately very simple: Optimize with what you have. Best predict what you need. Do both, and you’re on the road to creating more from less.?

Benjamin Fels is CEO and Co-Founder of Pendulum Systems – an AI company that specializes in supply chain optimization.