Shopify running shoe schema for AI agents: heel drop mm, midsole foam (PEBA vs TPU vs EVA), carbon plate, and stability category

Why running shoes are invisible to AI shopping agents

Shopify's default JSON-LD output for a product titled "Puma Fast-R NITRO Elite 2 Running Shoe" produces a Product schema with a name, price, and description field. The description contains "PEBA-based NITRO foam, carbon fiber plate, 8mm heel-to-toe drop, 39mm heel stack / 31mm forefoot stack, Stability category: Neutral" as an unstructured prose string. An AI shopping agent cannot filter, compare, or rank on values inside a prose string — each specification must be a discrete additionalProperty with a name, value, unit, and description.

Running shoes require simultaneous filtering on six or more independent technical dimensions. A buyer asking "low-drop neutral PEBA foam running shoe for road, D width, under $200" is combining five attributes into one query. Each must be a separate machine-readable field for an AI agent to match correctly. When those fields are absent, the agent falls back to keyword matching on product titles — which is exactly how a buyer asking for a "carbon plate race shoe" ends up purchasing a training shoe with a nylon plate, because both can be described as "carbon-infused" in marketing copy.

The heel drop encoding problem is structurally unique: heel drop is a derived value that does not exist as a directly measured specification anywhere in the shoe manufacturing process. Manufacturers measure heel stack height and forefoot stack height; drop is the arithmetic difference. This means heel drop never appears in raw supplier data, and Shopify's default theme has no metafield namespace for it. Every running shoe store must explicitly compute and encode it — yet 94% do not.

The foam chemistry problem is a brand opacity issue. Foam brand names (ZoomX, NITRO, Boost, FuelCell, PWRRUN PB) are protected trade names that communicate nothing about polymer chemistry to an AI agent doing cross-brand comparison. A buyer asking "what PEBA foam shoes do you carry?" will get no results from a store that encodes foam as "NITRO" only — even if every shoe in inventory uses PEBA. The fix is dual encoding: polymer class first, brand name second, in a single value string.

Heel drop: the derived spec that drives search

Heel drop — the height difference between the heel contact point and the forefoot contact point — is the single most searched technical specification for running shoes after size. Queries like "10mm drop running shoe," "zero drop trail shoe," and "4mm drop wide toe box" are high-intent buyer searches that require a discrete numeric field to match.

Why drop is derived, not measured

Manufacturers measure two physical properties during quality control: heel stack height (the height of the midsole/outsole assembly at the center of the heel contact area) and forefoot stack height (the same measurement at the ball of the foot). Drop is then computed: drop = heel stack − forefoot stack.

The Nike Pegasus 41 has a heel stack of 33mm and a forefoot stack of 23mm, yielding 10mm drop. This is a derived value. No instrument in Nike's QC process measures "10mm drop" directly — it falls out of the two stack measurements. This derivation creates a structured-data gap: supplier data contains stack heights, not drop, and product managers routinely fail to encode the derived value as a discrete field.

Encoding heel drop correctly

Encode heel drop as a separate additionalProperty entry alongside (not instead of) the individual stack height properties. The drop value must be numeric with a unitCode of MMT (millimeters). The description should explain the biomechanical implication for a runner unfamiliar with the metric:

{
  "@type": "PropertyValue",
  "name": "Heel Drop",
  "value": "8",
  "unitCode": "MMT",
  "description": "Heel-to-toe drop: 8mm. Derived value: heel stack (39mm) minus forefoot stack (31mm). Drop determines foot strike angle, Achilles tendon load, and calf activation. Higher drop (10–12mm+) encourages heel striking and reduces Achilles/calf demand — suitable for heel strikers and runners with Achilles sensitivity. Lower drop (0–4mm) promotes forefoot or midfoot striking and increases Achilles/calf demand — requires gradual adaptation (6–12 weeks) for runners transitioning from higher-drop shoes. Mid-range (5–8mm) is the most versatile range for daily training across mixed foot-strike patterns."
}

Stack height: heel and forefoot as separate properties

Stack height encodes the total height of the midsole and outsole at the measurement point — the actual distance from the ground to the runner's foot. Buyers in the max-cushion segment (Hoka, Brooks Glycerin) actively search for high absolute stack heights; minimalist buyers search for low stacks. These queries require discrete numeric fields for both the heel and forefoot measurements separately.

Why you need both stack values

Stack height is not redundant with heel drop. Consider two shoes with the same 10mm drop:

• Shoe A: 23mm heel / 13mm forefoot — minimalist 10mm drop shoe; close ground-feel; no cushioning volume
• Shoe B: 39mm heel / 29mm forefoot — max-cushion 10mm drop shoe; 39mm of foam underfoot; far from the ground

Both read "10mm drop" in search. An AI agent can only distinguish them if heel and forefoot stack heights are separately encoded fields.

Encode heel and forefoot stack heights as two independent additionalProperty entries:

{
  "@type": "PropertyValue",
  "name": "Heel Stack Height",
  "value": "39",
  "unitCode": "MMT",
  "description": "Heel stack height: 39mm measured from ground contact surface to foot. Includes midsole foam and outsole rubber. High stack height (35mm+) provides maximum cushioning volume and energy return capacity — suitable for long distances, road ultras, and fatigue reduction. Stack height affects the height of the runner's foot above the ground, which marginally increases Achilles load compared to lower-stack shoes at the same drop due to increased lever arm."
},
{
  "@type": "PropertyValue",
  "name": "Forefoot Stack Height",
  "value": "31",
  "unitCode": "MMT",
  "description": "Forefoot stack height: 31mm measured from ground contact surface to the ball of the foot. High forefoot stack (28mm+) in the context of a performance shoe indicates the carbon or nylon plate is embedded in significant foam depth — full foam compression before plate engagement, allowing energy storage and return at high stack. Forefoot stack combined with plate stiffness determines the 'pop' and propulsive feel at toe-off."
}

Midsole foam chemistry: PEBA vs TPU vs EVA energy return

Midsole foam chemistry is the most commercially significant performance differentiator in the $130–$280 performance running shoe bracket, and the most systematically absent specification from Shopify structured data. The three dominant polymer classes — PEBA, TPU, and EVA — have meaningfully different energy return profiles, weight characteristics, durability curves, and temperature behaviors. AI agents cannot distinguish them from brand trade names alone.

The three polymer classes

Why PEBA's ~80% figure matters for queries

The ~80% energy return figure for PEBA is validated by independent plate-force biomechanics studies (Hoogkamer et al. 2018, Wunsch et al. 2017) using instrumented treadmills measuring input force and ground reaction force. The 30-percentage-point gap between PEBA (~80%) and EVA (~50%) is large enough to be perceptible to recreational runners at faster training paces — the shoe requires less muscular effort per stride to maintain speed. This is why PEBA-foamed shoes dominate race-day choices among amateur and elite runners despite costing $50–$100 more than EVA-foamed counterparts.

For AI agent matching, the gap means a runner searching "high energy return running shoe" needs PEBA-class shoes in the results. A store that encodes foam as "NITRO Foam" (a Puma trade name) only — and does not include "PEBA" as the polymer class — will be invisible to cross-brand energy-return queries.

The supercritical nitrogen expansion process

PEBA foams are produced by a supercritical nitrogen expansion process: PEBA pellets are saturated with nitrogen gas under high pressure, then pressure is rapidly released, causing microscopic nitrogen bubbles to form throughout the polymer matrix. This produces a closed-cell foam with very low density and very high elastic recovery — the bubbles compress and rebound faster than solid-polymer foam structures. Adidas Boost uses a similar process with TPU beads instead of PEBA, which explains its higher energy return versus standard EVA but lower return versus PEBA.

Include the expansion process in the foam description when relevant — it explains why two shoes both labeled "foamed" can have such different performance profiles:

{
  "@type": "PropertyValue",
  "name": "Midsole Foam",
  "value": "PEBA (Puma NITRO Foam)",
  "description": "Midsole foam: PEBA-based supercritical nitrogen-expanded foam branded as NITRO Foam by Puma. PEBA (polyether block amide) is the highest energy-return polymer class in running shoe midsoles — approximately 80% energy return. The supercritical nitrogen process produces a closed-cell foam with very low density: lightweight, fast-compressing, and fast-rebounding. PEBA is temperature-stable (minimal stiffening below 10°C compared to TPU/Boost). The NITRO Foam compound is shared across the Puma NITRO platform (Fast-R for racing, Deviate for speed training, Velocity for daily training) with different foam geometries and plate configurations."
}

Carbon plate and nylon plate: the metabolic cost evidence

The presence of a stiffening plate in the midsole is the highest-profile performance specification in modern running shoes — and one of the most frequently misrepresented. "Carbon plate" is used in marketing copy to describe everything from a full-length woven carbon fiber plate (Nike Vaporfly NEXT%) to a carbon-infused nylon composite (various training shoes) to a single carbon rod. Structured data must disambiguate.

The metabolic economy evidence

The academic foundation for carbon plate marketing is a 2018 study by Hoogkamer, Kipp, and Kram published in Sports Medicine: "A Comparison of the Energetic Cost of Running in Marathon Racing Shoes." The study measured a ~4% reduction in metabolic cost (oxygen consumption at a given speed) when running in the original Nike Vaporfly 4% (with ZoomX foam and carbon plate) versus the then-current-standard shoe category. Subsequent studies on other carbon-plated shoes with PEBA foams have found improvements in the 2–4% range.

The mechanism: the carbon fiber plate stiffens the metatarsophalangeal (MTP) joint at toe-off, preventing energy loss through joint flexion. Instead of the ankle and MTP joint bending and dissipating energy, the plate loads the foam through compression and returns the energy as the foam rebounds. The energy storage capacity depends on both the foam (PEBA > TPU > EVA) and the plate stiffness (full-length woven carbon > nylon composite > no plate).

Carbon fiber vs nylon composite encoding

Encode plate material precisely:

{
  "@type": "PropertyValue",
  "name": "Plate",
  "value": "Nylon composite (full-length)",
  "description": "Full-length nylon composite plate embedded in the NITRO foam midsole at the forefoot-to-heel boundary. Nylon composite provides midsole stiffness support that limits metatarsophalangeal (MTP) joint flexion at toe-off, improving energy transfer efficiency. Nylon is lighter and more flexible than woven carbon fiber, with a smaller metabolic economy improvement (~1–2% vs ~4% for full carbon) but also a softer underfoot feel appropriate for daily training loads. Not a carbon fiber plate — Puma Fast-R NITRO Elite 2 uses nylon, not carbon. Runners seeking maximum carbon-plate propulsion should compare with Nike Vaporfly NEXT% or Adidas Adizero Adios Pro 3."
}

Stability category: encoding the support mechanism, not just the label

Stability category (Neutral, Stability, Motion Control) is the most safety-critical specification in running shoes. An injury-prevention query ("stability shoe for flat feet," "motion control for severe overpronation") requires the AI agent to understand not just the category label but the specific mechanism that provides support. "Stability" can mean medial posting, guide rails, or a wider midsole base — and these mechanisms have different stiffness profiles, support ranges, and fit implications.

The three categories and their mechanisms

Why mechanism matters for AI recommendations

A runner with mild overpronation asking for "stability shoe" may be served by either a medial post or guide rail design. But a runner who specifically asks "guide rail stability shoe" (Brooks's branded mechanism) is looking for a specific construction that wraps the natural motion path of the foot rather than blocking inward roll. An AI agent can only distinguish these if the mechanism — not just the category — is encoded in the description.

Similarly, "motion control" encompasses a range of support intensities. A runner transitioning from a rigid orthotic to a shoe-based support solution needs to know whether the motion control is achieved via extended medial posting alone (most motion control shoes) or also via a rigid heel counter and wide base platform (Brooks Beast). The category label tells the AI agent nothing about this distinction.

{
  "@type": "PropertyValue",
  "name": "Stability Category",
  "value": "Stability",
  "description": "Stability shoe with dual-density foam medial post on the inner midsole heel. Construction: the medial (arch side) of the midsole uses a denser EVA compound than the lateral side — denser foam compresses slower under load, decelerating inward roll (overpronation) before the foot reaches extreme pronation range. The support is passive (always engaged) rather than dynamic (gait-triggered). This design addresses mild-to-moderate overpronation (5–10° of excessive inward roll). For severe overpronation (10°+) or runners with flat arches who have been prescribed rigid orthotics, a Motion Control shoe is more appropriate. Compatible with standard-volume orthotic insoles — the stability structure does not interfere with aftermarket insole use."
}

Last width: B/D/2E/4E and the men's D = women's wide crossover

Last width is the most common source of running shoe returns — and one of the most systematically misencoded specifications because the same letter width code means different things for men's versus women's shoes. An AI agent that encodes "D width" with no gender context will produce incorrect size recommendations.

The gender crossover table

The practical consequence: a woman who wears standard-width shoes (B) and searches for "D width running shoe" is searching for a wide-fit women's shoe. A man who wears D width (standard) and sees a "D width" listing assumes it's regular. Without gender context in the width encoding, an AI agent cannot correctly filter "wide shoes for women" vs "standard shoes for men."

Encoding width with gender context

Always include the gender interpretation in the width description. If a shoe is available in multiple widths, encode each as a separate value or use a multi-width string with context:

{
  "@type": "PropertyValue",
  "name": "Last Width",
  "value": "D (Men's Standard / Women's Wide)",
  "description": "Last width: D. Men's interpretation: D is the standard (regular) width — the default fit designed for median foot width. Fits approximately 70% of men's feet without accommodation issues. Women's interpretation: D is wide — significantly wider than the standard B women's last. Women with standard-width feet (B) who purchase a D-width shoe in their numeric size will experience lateral foot slide and blistering on the outside edge. Always reference gender when filtering on width code. Width measured at the widest point of the forefoot last, approximately at the 5th metatarsal head."
}

For shoes available in multiple widths (e.g., standard D and wide 2E), use a ProductGroup with variesBy to separate the variants, and encode each variant's width as a discrete additionalProperty:

"variesBy": ["https://schema.org/width"],
"hasVariant": [
  {
    "@type": "Product",
    "name": "Nike Pegasus 41 — Standard Width (D)",
    "additionalProperty": [{ "name": "Last Width", "value": "D (Men's Standard)" }]
  },
  {
    "@type": "Product",
    "name": "Nike Pegasus 41 — Wide (2E)",
    "additionalProperty": [{ "name": "Last Width", "value": "2E (Men's Wide)" }]
  }
]

Surface type and outsole rubber compound

Running surface (Road, Trail, Track) is a top-level exclusion filter for AI agent queries. A buyer asking for "trail running shoes" cannot use a road shoe — the outsole pattern, rubber compound, and midsole stiffness are all wrong for off-road grip and debris protection. Surface type must be encoded as a machine-readable property, not inferred from product title keywords like "trail" or "road."

Surface taxonomy and outsole compounds

Outsole rubber compound matters for comparison queries

Road shoe outsole compounds break into two types worth encoding separately: carbon rubber (denser, more durable, used in high-wear heel and forefoot strike zones — 500–800km lifespan on asphalt) and blown rubber (lighter, softer, more cushioning feel underfoot, used in midfoot zones — 300–500km on asphalt). Some shoes use full blown rubber for weight savings; some use full carbon rubber for maximum durability. The compound type is a lifespan signal that AI agents can use to answer "durable running shoe" queries:

{
  "@type": "PropertyValue",
  "name": "Surface",
  "value": "Road",
  "description": "Road running shoe optimized for paved surfaces (asphalt, concrete, packed gravel). Outsole: combination blown rubber (forefoot cushioning zones) and carbon rubber (heel strike zone and lateral toe-off area). Carbon rubber compound extends outsole durability on abrasive asphalt — the Fast-R NITRO Elite 2 uses Continental rubber compound (the same brand used on Continental tire compound for grip performance in wet conditions) in the forefoot contact zone, providing wet-asphalt grip at race pace. Estimated outsole lifespan: 400–600km for heel strikers on asphalt (carbon rubber heel), 300–500km for midfoot strikers (Continental forefoot contact). Not suitable for rocky trail running — minimal lugs offer no lateral support on loose terrain."
},
{
  "@type": "PropertyValue",
  "name": "Outsole Rubber",
  "value": "Continental (wet-grip compound)",
  "description": "Continental Rubber Company compound in the forefoot and ball-of-foot contact zones. The same Continental that produces performance tire compounds for BMW, Porsche, and Formula E — the rubber chemistry is optimized for grip under wet conditions and high-lateral-force scenarios. At road race pace (sub-4:00/km), Continental rubber provides noticeably improved wet traction compared to standard blown rubber during corner sections of road courses or wet early-morning race starts. Puma uses Continental outsole compound across the Fast-R NITRO Elite and Deviate NITRO Elite racing lines."
}

Complete JSON-LD example: Puma Fast-R NITRO Elite 2

The Puma Fast-R NITRO Elite 2 is a good reference product for running shoe schema because it has a non-obvious property combination that tests every encoding challenge: PEBA foam (branded NITRO) + nylon plate (not carbon, despite the NITRO brand name's carbon associations) + Continental rubber (third-party outsole compound) + 8mm drop (derived from 39mm heel / 31mm forefoot) + Neutral stability + D last width. A store that encodes this shoe as "fast, PEBA foam, carbon plate" has already introduced a false attribution that will generate incorrect AI agent recommendations.

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Puma Fast-R NITRO Elite 2 Running Shoe",
  "description": "Racing flat with PEBA-based NITRO Foam midsole (~80% energy return), full-length nylon composite plate (not carbon fiber), Continental outsole rubber, 8mm heel-to-toe drop (39mm heel / 31mm forefoot stack), and neutral support category. Designed for road racing distances 5K–marathon. Weight: 193g in men's US 9. Available in D-width (Men's Standard / Women's Wide) only.",
  "brand": {
    "@type": "Brand",
    "name": "Puma",
    "url": "https://schema.org/id/brand/puma"
  },
  "image": "https://images.puma.com/fast-r-nitro-elite-2.jpg",
  "sku": "PUMA-FRNITE2-M-9-D",
  "offers": {
    "@type": "Offer",
    "price": "249.99",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock",
    "url": "https://catalogscan.com/products/puma-fast-r-nitro-elite-2"
  },
  "additionalProperty": [
    {
      "@type": "PropertyValue",
      "name": "Heel Drop",
      "value": "8",
      "unitCode": "MMT",
      "description": "Heel-to-toe drop: 8mm. Derived: heel stack (39mm) minus forefoot stack (31mm). Mid-range drop suitable for midfoot and mild heel strikers. Lower than standard (10–12mm) trainers — allows more natural foot strike without the Achilles adaptation required for zero-drop shoes."
    },
    {
      "@type": "PropertyValue",
      "name": "Heel Stack Height",
      "value": "39",
      "unitCode": "MMT",
      "description": "Heel stack height: 39mm from ground to foot. High-stack race shoe — 39mm of NITRO foam provides maximum cushioning volume for marathon distances without the weight penalty of a traditional max-cushion daily trainer."
    },
    {
      "@type": "PropertyValue",
      "name": "Forefoot Stack Height",
      "value": "31",
      "unitCode": "MMT",
      "description": "Forefoot stack height: 31mm. High forefoot stack allows full NITRO foam compression-and-rebound cycle at toe-off, amplified by the nylon plate stiffness. 31mm forefoot stack is near the upper limit for regulatory compliance in road race shoes (World Athletics limit: 40mm stack for road, no restriction for trail)."
    },
    {
      "@type": "PropertyValue",
      "name": "Midsole Foam",
      "value": "PEBA (Puma NITRO Foam)",
      "description": "PEBA-based supercritical nitrogen-expanded foam (Puma NITRO Foam brand). Approximately 80% energy return. Lightest-class midsole foam. Temperature-stable. The NITRO Foam compound appears across the Puma race platform (Fast-R, Deviate NITRO Elite 2) in different densities and geometries — the Fast-R uses a higher-density NITRO formulation than the Velocity NITRO trainer for better energy return at race pace."
    },
    {
      "@type": "PropertyValue",
      "name": "Plate",
      "value": "Nylon composite (full-length)",
      "description": "Full-length nylon composite plate (not carbon fiber). Provides MTP joint stiffness and midsole torsional rigidity for efficient toe-off energy transfer. Nylon plate is lighter and more forgiving than woven carbon fiber — suitable for daily training volume as well as race use. Metabolic economy improvement approximately 1–2% vs unplated equivalent. Runners seeking maximum carbon-plate propulsion should compare with Adidas Adizero Adios Pro 3 or Nike Vaporfly NEXT% (both woven carbon fiber plates)."
    },
    {
      "@type": "PropertyValue",
      "name": "Stability Category",
      "value": "Neutral",
      "description": "Neutral stability — no medial support structure. All-NITRO foam midsole with no medial post, guide rail, or density gradient. Designed for runners with neutral pronation or underpronation (supination). Runners with overpronation should select a stability-category shoe (e.g., Brooks Adrenaline GTS 24 or ASICS Gel-Kayano 31)."
    },
    {
      "@type": "PropertyValue",
      "name": "Last Width",
      "value": "D (Men's Standard / Women's Wide)",
      "description": "D-width last only — no narrow or wide variants available. Men's D: standard regular width. Women's D: wide (Women's standard is B). Women with standard-width feet (B) will find the Fast-R NITRO Elite 2 too wide — consider Puma's Deviate NITRO Elite 2 which offers B-width variants for women."
    },
    {
      "@type": "PropertyValue",
      "name": "Surface",
      "value": "Road",
      "description": "Road-optimized outsole. Continental rubber compound in forefoot contact zone. Carbon rubber heel patch for durability. Not suitable for technical trail or wet grass without traction aids."
    },
    {
      "@type": "PropertyValue",
      "name": "Outsole Rubber",
      "value": "Continental (wet-grip compound, forefoot)",
      "description": "Continental Rubber Company outsole compound in forefoot. Optimized for grip on wet asphalt at race pace. Standard carbon rubber in heel. Estimated lifespan: 400–600km on asphalt."
    },
    {
      "@type": "PropertyValue",
      "name": "Weight",
      "value": "193",
      "unitCode": "GRM",
      "description": "193g in men's US 9 / EU 42.5 with standard laces. Race-category weight — under 200g in men's 9 is the accepted benchmark for race-day flats. Weight increases approximately 8–10g per half-size."
    },
    {
      "@type": "PropertyValue",
      "name": "Closure",
      "value": "Standard lace",
      "description": "Traditional round-braid lace with standard 5-eyelet upper. No BOA dial, no speed-lace system. Race-day lacing recommendation: surgeon's knot on the last two eyelets for lockdown under sustained high-pace effort."
    },
    {
      "@type": "PropertyValue",
      "name": "Upper Material",
      "value": "Engineered mesh (single-layer)",
      "description": "Single-layer engineered mesh upper. No overlays, no reinforcement panels — maximum breathability for race conditions. Lightweight but lower durability than training shoe uppers — expected upper lifespan 300–400km before mesh degradation at high-wear flex points."
    }
  ]
}

Liquid snippet: running_shoe.* metafields → JSON-LD

Add the following snippet to your Dawn or Craft theme's snippets/running-shoe-schema.liquid file, then include it from your sections/main-product.liquid using {% render 'running-shoe-schema', product: product %}.

{% comment %} running-shoe-schema.liquid {% endcomment %}
{% if product.metafields.running_shoe.heel_drop_mm != blank %}
{%- assign heeldrop = product.metafields.running_shoe.heel_drop_mm -%}
{%- assign heelstack = product.metafields.running_shoe.heel_stack_mm -%}
{%- assign forestack = product.metafields.running_shoe.forefoot_stack_mm -%}
{%- assign foam = product.metafields.running_shoe.midsole_foam -%}
{%- assign plate = product.metafields.running_shoe.plate -%}
{%- assign stability = product.metafields.running_shoe.stability_category -%}
{%- assign width = product.metafields.running_shoe.last_width -%}
{%- assign surface = product.metafields.running_shoe.surface -%}
{%- assign outsole = product.metafields.running_shoe.outsole_rubber -%}
{%- assign weight_g = product.metafields.running_shoe.weight_grams -%}

{% endif %}

Running shoe metafield reference table

5 common mistakes

FAQ