ecommerce // decisioning infrastructure

Product ranking, recommendations, and retail media,
resolved in one decision.

The ranking layer between your catalog and the products a shopper sees. Gortex orders every recommendation, search, and browse surface and places sponsored products in one call, so your team ships storefront instead of running a ranking stack and a separate ad auction.

request access → see the decisioning apis
# 01 / the problem

Ranking a store is a standing ML cost, and retail media makes it two systems.

Product ranking and recommendations never settle. Demand shifts, inventory turns over, and every surface, from the home grid to search to the cart, needs its own tuning. Then retail media arrives as a second auction, built apart from organic ranking and reconciled against it by hand, release after release.

01 recommendations as ongoing ml spend
Recommender models, feature pipelines, and an online serving stack that has to stay under budget through peak traffic and seasonal spikes.
02 latency on every product grid
Ranking runs on the request path. Every added signal is latency you pay on every browse page, every search, and every recommendation strip.
03 retail media as a separate auction
Sponsored products get built as their own system, then reconciled with organic ranking by hand, so a paid placement and a relevant one keep colliding for the same slot.
04 no record of why a product ranked
When a merchandiser questions placement or a seller disputes a sponsored slot, there is often no trace of why a product placed where it did.
# 02 / how gortex fits

One endpoint ranks the products and places sponsored slots in the same response.

You send a shopper, a surface, and the candidate products you already retrieved. Gortex returns the organic order and the sponsored placements together, inside one sub-200ms call, with a decision id you can audit later. The same call powers a recommendation engine api across your store, and it draws on the same primitives as our marketplace ranking api and sponsored listings api.

01
your retrieval
you gather the candidates: products, variants, or recommendations for the surface.
02
gortex decides
rank the products and resolve sponsored slots against that order, in one call.
03
your storefront
render the ranked grid with sponsored products already placed and priced.
01 you keep retrieval and your catalog
Gortex ranks the products you bring. No data migration, no re-platforming of your store.
02 ranking and monetization in one call
Organic order and sponsored products resolve against each other, so retail media only takes a slot when the placement still fits the shopper. See the sponsored listings api.
03 every decision is traceable
Each response carries a decision_id and a trace_id. The reason a product ranked is recorded, not reconstructed after a dispute.
04 seven typed sdks
Generated from one OpenAPI 3.1 spec, so contract drift across languages is mechanically impossible.
# 03 / the call

One request. A ranked product grid and its sponsored slots.

POST /v1/decide
200 OK · 69ms
// Rank a product grid and fill its sponsored slots.
$ curl api.gortex.ai/v1/decide \
  -H "Authorization: Bearer $GORTEX_KEY" \
  -d '{
    "recipient": { "id": "shopper_8120" },
    "context":   { "surface": "product_grid" },
    "items":     [ /* 200 candidate products */ ]
  }'

{
  "ranked":      [ /* ordered product IDs */ ],
  "sponsored":   [{ "slot": 2, "price": 0.51 }],
  "decision_id": "d_01HXZ3M",
  "trace_id":    "t_01HXZ3N"
}
# 04 / questions

Ecommerce ranking, answered plainly.

What does Gortex do for an ecommerce store?

It ranks the surfaces where shoppers find products: recommendations, search, and browse. For each request you send the candidate products you already retrieved, and Gortex returns the order plus any sponsored product placements in the same response, so organic ranking and retail media resolve together.

Do I have to replace my product catalog or search stack?

No. You keep retrieval and your catalog where they are, and send the candidates you already have. Gortex ranks what you pass in and returns the order plus sponsored slots. There is no re-platforming and no data migration.

How does retail media fit with organic product ranking?

Sponsored products resolve against the same ranked order in one sub-200ms call, so a paid placement only takes a slot when it still fits the shopper. You stop building and reconciling a separate ad auction against organic ranking release after release.

# 05 / start

See an ecommerce ranking decision, live.

The private beta is by invitation. Tell us the surfaces you rank and whether you run retail media, and we will get you a key.

request access → see the decisioning apis view the simulation
# related // decisioning apis