Retail / CPG

Digital Twins for Ecommerce: Zalando Scales 3D Visualization

Objective

Zalando is making online fashion shopping more realistic by using ALLSIDES’ NVIDIA‑powered 3D platform to turn footwear into high‑fidelity digital product twins that customers can explore from every angle.

Customer

Zalando

Partner

ALLSIDES

Topic

Content Creation / Rendering

Key Takeaways

From Scan to Storefront in 48 Hours

  • With ALLSIDES 3D scanners and NVIDIA Cosmos™, Gen‑3C, and DiffusionRenderer, Zalando can bring products online in 3D within 48 hours, accelerating content production while boosting add‑to‑cart rates by 3%–4% and reducing returns in footwear.

Maximizing ROI Through GPU Acceleration

  • Powered by NVIDIA RTX™ 5000 and RTX PRO™ 6000 Blackwell GPUs, Zalando scaled high‑fidelity 3D product production to 10,000 footwear SKUs in 2025 and is targeting 45,000 SKUs in 3D for 2026, supported by seven‑minute, fully automated scans.

Premium Quality at 1/10th the Price

  • Using ALLSIDES’ NVIDIA‑accelerated 3D pipeline, Zalando reduced the cost of each 3D scan by up to 92.1% and improved geometry and texture accuracy by 10x, with scanning running 99.4% faster than traditional methods, so high‑fidelity 3D content can be produced economically at scale.

How Zalando scales 3D digital twins to transform online fashion?

Founded in Berlin in 2008, Zalando has grown into a leading European technology platform for fashion and lifestyle, connecting around 62 million active customers with more than 7,000 brands across 29 markets and processing over 278 million orders a year. With that level of scale, even small improvements in how products are presented online can translate into meaningful changes in customer experience and operational efficiency.

Why Zalando Needs High-Fidelity 3D Visualization

As a leading European fashion and lifestyle platform, Zalando wants its digital experience to match the confidence customers feel when they can pick up a product in a store. Standard product photos on a white background give customers a clear overview, but cannot fully capture shape, depth, and materials, especially in footwear and accessories where silhouette and design drive purchase decisions. As customer expectations for richer digital experiences evolve, this created both a challenge and an opportunity to move beyond static imagery.

Behind the scenes, Zalando relied on repeated studio shoots for different formats and markets, which were time‑consuming, expensive, and difficult to adapt quickly for new campaigns. Early 3D experiments existed, but they sat outside the main workflow, took about three times longer than the new approach, required specialist skills and manual modeling, and didn’t scale across hundreds of thousands of SKUs. Visual quality was inconsistent, and teams still needed separate shoots and asset pipelines for product pages, marketing, and other channels.

How is Zalando achieving scalable 3D digital asset creation for ecommerce?

Against this backdrop, Zalando set out to give customers a clearer, more realistic view of products before purchase, helping them make more confident decisions and reduce returns—starting with the high‑impact footwear category. The team wanted to move beyond static imagery into richer visualization—interactive 3D, video, and immersive experiences—that feel closer to an in‑store view.

To make that vision viable, Zalando needed a scalable 3D asset pipeline rather than isolated pilots. The company also wanted to automate image and video creation from 3D assets and build a reusable digital‑twin foundation to support future generative AI and immersive commerce experiences.

ALLSIDES

Side‑by‑side view of a physical shoe from Zalando’s assortment and its high‑fidelity 3D digital twin captured with ALLSIDES.

From a Single Scan to 3D Views and Videos, AR Try‑On, and Marketing Content

To address these challenges, Zalando integrated ALLSIDES’ AI‑powered photometric 3D scanning platform into its content production pipeline, with NVIDIA GPUs providing the compute backbone. ALLSIDES’ scanners capture products in a controlled multi‑view setup and generate high‑fidelity digital twins that accurately reflect geometry, materials, and color, with each footwear scan taking around seven minutes.

These digital twins become reusable assets that power multiple content formats at scale through Zalando’s ShopAR platform, which operationalizes 3D content across the business. From a single scan, Zalando can generate:

  • Interactive 3D product views on product detail pages
  • Classic packshots on white backgrounds
  • Lifestyle and campaign imagery
  • Marketing videos and product spins
  • Augmented reality experiences, including virtual try‑on

This replaces many separate shoots and workflows with a single capture that serves multiple teams and channels.

ALLSIDES Photometric 3D Scanning Technologies Powered by NVIDIA

ALLSIDES runs its reconstruction pipeline on premises using NVIDIA RTX 6000 Ada and NVIDIA RTX PRO Blackwell GPUs as the computational backbone for large‑scale 3D reconstruction and neural processing. At the software layer, NVIDIA technologies sit at the core of the 3D engine:

  • NVIDIA® CUDA® and OptiX™‑based differentiable ray tracing extract high‑fidelity, physically accurate materials from multi‑view scans.
  • Neural texturing with nvdiffrec recovers detailed BRDF properties, enabling advanced operations such as material segmentation and inpainting on the resulting digital twins.
  • NVIDIA Cosmos, Gen‑3C, and DiffusionRenderer generate photorealistic environments, camera‑controlled videos, and material‑aware relighting from a single scan.

Together, this stack allows Zalando teams to create marketing‑ready visuals without scheduling additional photo shoots, backing a pipeline that can support tens of thousands of 3D assets per year.

Automate 3D Asset Pipeline: Scan Once, Use Everywhere

The new workflow is designed to be simple for Zalando teams while handling complexity behind the scenes. For a typical footwear SKU:

  • A product is placed in an ALLSIDES scanner, which captures a multi‑view photometric dataset. NVIDIA GPU‑accelerated reconstruction creates a high‑fidelity digital twin.
  • The 3D asset is automatically ingested into ShopAR, where it passes quality checks.
  • Derivative content—interactive 3D views, AR try‑on, contextual videos, and marketing imagery—is generated from the same asset and distributed to the right channels.

Multiple Zalando teams now work from this shared 3D asset foundation, including ShopAR, Zalando Marketing Services (ZMS), Content Strategy & Curation, Product Curation, 3D Design, and Content Supply. Instead of commissioning new shoots for each use case, they request a scan and derive the formats they need—product images, video, lifestyle content, or virtual try‑on—from the same digital twin.

This “scan once, use everywhere” model enables Zalando to move from capture to campaign‑ready 3D content in as little as 48 hours. It speeds planning and execution, keeps visual standards consistent across brands and markets, and lowers overall cost even with more enhanced product detail pages. Brand partners selling on Zalando also benefit from richer product presentation without producing their own 3D assets, and ShopAR is offered as a B2B service to extend these capabilities across markets.

ALLSIDES

“Together with ALLSIDES and powered by NVIDIA, we are building a new foundation for how fashion content can work in the future. Richer 3D experiences help customers understand products better and make more confident decisions, while also opening up new possibilities for our brand partners. When a single digital asset can be used across product pages, virtual try‑on, localised marketing campaigns and multiple platforms, it creates the most efficient, flexible and scalable content model. We believe this can reshape how fashion content is created and used across the industry.”

Matthias Haase
VP Content Solutions, Zalando

Richer 3D Experiences Drive Measurable Impact on Engagement, Returns, and Scale

The NVIDIA‑accelerated ALLSIDES pipeline is driving measurable improvements in how customers interact with Zalando’s product pages. Key results include:

  • 15%–20% increase in time spent on product pages featuring elevated 3D and video content
  • Around 30 seconds of average engagement time in AR experiences built on digital twins
  • 3%–4% uplift in add‑to‑cart rates for products with enhanced 3D, with significant return reductions in footwear

These changes show that more realistic views of products help customers make better‑informed purchase decisions.

Scaling 3D Production

Operationally, the 3D pipeline is built for high volume and speed. Between April and December 2025, Zalando produced 10,000 footwear SKUs in 3D and plans to deliver a further 35,000 in 2026, targeting 45,000 3D footwear assets overall. Automated scanning and reconstruction allow each scan to complete in about seven minutes, with 3D production fully automated. According to ALLSIDES data, the platform delivers:

  • Up to 92.1% lower price per 3D scan
  • 10x more accurate geometry and textures
  • 99.4% faster scanning times compared with traditional approaches

Together, these gains make high‑fidelity 3D content a viable, everyday part of Zalando’s operations.

Powering Next‑Generation Commerce

The 3D digital‑twin pipeline positions Zalando for future innovation. With a growing library of accurate digital twins, the company can explore AI‑generated product photography, automated video creation, and increasingly immersive AR and VR experiences. These assets can support physical AI applications today and, over time, feed into simulations for logistics and supply‑chain scenarios. Together, Zalando, ALLSIDES, and NVIDIA are creating a flexible content and data foundation that can support the next wave of digital commerce.

“I am proud that ALLSIDES is enabling Zalando to build one of the largest 3D production workflows in fashion ecommerce today. By combining our 3D scanning technology with NVIDIA’s GPU‑accelerated AI infrastructure, we are unlocking entirely new possibilities for digital commerce and generative content creation with a focus on driving results and growth on the major metrics in ecommerce.”

Franz Tschimben
Cofounder and CEO, ALLSIDES

Discover how NVIDIA-powered 3D digital twins are reshaping ecommerce and retail.

Related Customer Stories