2019 — 2024

Building Infrastructure That Changed
How a $50B Company Sizes Products

Turning 1.1 million body scans into a platform where anyone at Nike could query population data and get answers that shaped what they made. From research scientists to marketing directors.

Nike Fit for Everybody — inclusive sizing initiative
The Challenge

Sizing built for mannequins, not people

Nike's sizing system had been built around a small set of idealized mannequin measurements, the same approach the industry had used for decades. The result: products that fit a narrow slice of the population well, and everyone else poorly. High return rates. Inventory inefficiencies. A growing gap between Nike's stated commitment to serving every athlete and what the products actually delivered.

The data existed. Nike had 1.1 million body scans sitting in fragmented archives, external datasets, and portable scanner campaigns. But it was inaccessible, inconsistently formatted, and disconnected from the design and product teams who needed it. The goal: turn body data from a research asset into infrastructure that every team at Nike could use to make better products.

1.1M
body scans consolidated into a single platform
3
teams built from scratch: Engineering, ML/CV, Design
5yr
multi-year ML strategy with executive sponsorship
body data became permanent infrastructure, not a project
The Approach

From fragmented archives to living infrastructure

Building the Data Foundation

  • Consolidated 1.1 million body scans from three sources: portable scanner campaigns across global markets, external third-party datasets, and Nike's historical scan archives
  • Built the data pipeline to harmonize inconsistent scan formats, resolutions, and metadata schemas across decades of collection methods
  • Developed anthropometric diversity analysis: not just larger sample sizes, but ensuring the data represented the full range of body geometries Nike's products needed to fit
  • Created the first body type clustering system for Nike, a principled way to group the scan population so grading and design decisions could be made against real data
Data Sources
Portable Scanner Campaigns
Global markets, diverse populations
External Datasets
Third-party body scan research data
Historical Archives
Decades of Nike scan collections
1.1M Scans → Single Platform
BodyHub Platform — Query Layer
Stakeholder Types
NSRL Scientists
Apparel Designers
Marketing Directors
↓ ↓ ↓
Shared Infrastructure
Body Data Query Engine
Morphotype Architecture
3D-CAD Fit Tools
Outcomes
Grade Rules
Size Systems
Fit Validation

The BodyHub Platform

  • Built BodyHub: a query platform that let anyone at Nike, from sport scientists to marketing leads, get population-level answers from the scan database
  • Designed the platform to serve very different user types without requiring each to become a data scientist. The interface adapted to the question, not the other way around.
  • Integrated 3D-CAD fit tools directly into designer workflows, replacing subjective fit sessions with data-driven geometry validation
  • Secured executive sponsorship and multiple rounds of funding by demonstrating measurable ROI at each platform milestone

Teams Built from Scratch

  • Hired and led three teams from zero: Engineering (platform and data infrastructure), ML/CV (model development and validation), and Design (tooling and stakeholder interfaces)
  • Built a cross-functional operating model where research, engineering, and design worked in sprint cycles, not sequential handoffs
  • Established an inclusive, transparent leadership culture that maintained team motivation and retention across a five-year initiative
  • Partnered with product leaders across Nike's digital creation organization to align roadmaps with broader strategy

Inclusive Design Outcomes

  • Plus-size recalibration: Rebuilt grade rules from the data up for extended sizes, moving away from the industry-standard practice of proportionally scaling from a base size that didn't represent plus-size body geometry
  • Gender-fluid kids sizing: Developed size systems for kids' products that accounted for the full range of body proportions, not binary gender assumptions
  • MLB sizing consolidation: Reduced thousands of historical patterns to a 32-size matrix grounded in actual athlete body data. Less manufacturing complexity, better fit across the league.
  • High-difficulty silhouettes: Applied body type clustering to bra and tight fit validation, categories where poor fit had historically driven the highest return rates

Key insight: The goal was never to build a better sizing chart. The goal was to make body data a permanent capability, infrastructure that any team could build on rather than a project that any team could defund. That distinction is what determined every architectural and organizational decision along the way.

Key Outputs

Four initiatives that changed what Nike made

Inclusive Sizing

Plus-Size Recalibration

Rebuilt extended size grade rules from body scan data, not scaled-up base sizes. First time Nike's plus-size products were graded against the actual geometry of the bodies that wear them.

Kids Apparel

Gender-Fluid Kids Sizing

Developed size systems for kids' products grounded in actual child body proportions, removing binary gender assumptions from categories where they didn't reflect how kids' bodies are shaped.

MLB Partnership

Uniform Sizing Consolidation

Reduced thousands of historical MLB uniform patterns to a 32-size matrix derived from athlete body scan data. Less manufacturing complexity, better fit across the league.

Design Tools

3D-CAD Fit Validation

Built CAD fit tools that gave designers direct access to population body data during the creation process, replacing subjective fit sessions with geometry validated against the scan database.

Fit for Everybody — internal demonstration video available on request.
The full platform walkthrough is not publicly shared per NDA. Happy to discuss in detail during a discovery call.
The Outcome

Body data became infrastructure at Nike — not a project, a capability.

After five years, body data was no longer a research initiative at Nike. It was a permanent part of how the company designed, sized, and validated products. Design teams adopted new creation methods. Return rates on high-difficulty silhouettes improved. The BodyHub platform became a shared resource across multiple business units. The work established a new standard for what inclusive sizing can mean at global scale.

Why It Matters

Infrastructure thinking at product scale

The Nike work is the fullest expression of a principle that runs through everything since SOLS: physical products are data problems. The question was never "how do we make better orthotics" or "how do we make better leggings." It was always: how do we build the data infrastructure that makes better products inevitable?

At Nike, that meant building something that would outlast any individual initiative, a platform that didn't require ongoing advocacy to justify its existence because the teams who used it couldn't imagine working without it. That's the definition of infrastructure that worked.

Building something in this space?

I help teams build body data infrastructure that becomes part of how products are made — not a project that ends when the budget cycle turns.