WATCH How Predictive AI is Transforming Merchandising

When you crave creative self-expression, chances are you turn first to your closet.

In every city at any given point, thousands or even millions of apparel lovers dig hopefully through wadded piles of fabric in closets, hoping to find the perfect reflection of their identity. Each person special, each style distinctly theirs.

And should their personal stockpile prove uninspiring, they turn next to the internet or a local store — the infinite extensions of their wardrobe.

For the retailer behind these storefronts, the challenge is clear. Helping shoppers express their unique selves is a core promise of almost any brand. But meeting diverse customer expectations is a logistical and demand planning nightmare.

It begs the question: How can retailers nurture individuality at scale?

When Fashion and Math Collide

People want to walk into a store, see something they like, and walk out with it in their size. Simple.

Or, not so simple, actually.

It’s easy to overlook how incredibly challenging the logistical ballet is underpinning inventory availability. Months before anyone has even thought to go shopping, merchandising teams have devoted countless hours anticipating demand, crafting the perfect item, and placing initial orders.

Allocators then work tirelessly to place inventory in store in the hopes of delivering that golden shopping experience. For the uninitiated, the complexity in making this happen is mindblowing.

As a “simple” example, take that bastion of individuality: the t-shirt.

Suppose an allocator wants to ensure every customer who walks into their 30 storefronts can get the shirt they want. They’re tasked with supporting near-daily availability across the size curve (say, XS through XL) and across colors (black, white, and blue).

Thanks to the jaw-dropping impact of combinatoric math, that allocator is now tasked with accurately anticipating demand 450 times. For one shirt.

Rules of thumb can get them part of the way, but in the pursuit of customer experience perfection, good enough is insufficient. Taking the “safe” route and applying generous allocations across sizes to each store is pretty much guaranteed to be a giant waste. Quite literally — wasted product going unsold and piling in landfills, wasted capital spent on unsellable items.

Merchandising teams are, arguably, the most overlooked and underappreciated workers in our modern economy. Precisely because they are so good at what they do, shoppers have now come to expect instant gratification as soon as the new shirt impulse strikes. (We should also send a special shoutout to the customer-facing retail workers who have to deal with the frustration stemming from out-of-stock sizes or styles).

To sum: Getting products to where there is demand is a very hard problem to solve. And when it’s not done well, it’s incredibly wasteful. Surely there’s a better way.

Solving Big Data Problems with the Right (AI) Tools

Even if the general populace had a better understanding of merchandising roles, they’d still be blown away at how well these teams are doing with (frequently) simple tools: a well-worn Excel sheet, best-guess forecasts, and the power of intuition that comes from subject matter expertise.

“The right tool for the job” is a timeless adage for a reason; we’re often not using the right tool, impeding our success without good reason. In the case of nurturing individuality, we can simplify the “job” down to “predict and fulfill demand,” which is ultimately a big data problem for two reasons.

The data is big, for one, because of the combinatoric example from above.

And the data challenge is also big because forecasting is an intricate science. It can benefit from a number of inputs, ranging from the tried and true (historical sales) to the more complex (purchase intent signals, trend implications, local weather). Even with all that data to hand, it’s a time-consuming and technically intensive process to decipher what combination of data applied to which model will deliver the strongest prediction.

Enter: AI.

No, not the AI you’re probably thinking of. While Generative AI has captured the hearts of budding artists and copywriters alike, Predictive AI hums quietly in the background, ready to improve the lives of anyone who aims to delight through anticipation.

Predictive AI focuses on anticipating future events based on data, unlike Generative AI, which creates new content.

So to answer our question from above: yes, we can nurture individuality at scale. Look no further than Desigual, the premier example of how possible it is to deliver starkly unique style to customers without sacrificing availability.

Desigual uses Syrup’s predictive AI to support intelligent inventory rebalancing — by chasing demand where it’s hottest, the merchandising team improved revenue by 12% in one quarter, all while eliminating 3+ hours of analysis from product managers’ workweek.

And if they can do it, so can any brand.