The Item Primitive

Business software platforms become infrastructure when they own a primitive - a fundamental unit of data that all workflows reference and transact on.

  • Salesforce understood this with the customer record. Every interaction, every deal, every support ticket flows through the customer.
  • Stripe understood this with the payment record. Every transaction, every subscription, every payout flows through the payment.
  • Rippling understood this with the person record. Every payroll run, every benefit, every access permission flows through the employee.

But in commerce, point solutions have missed the opportunity to build true infrastructure, because they’ve built on the wrong primitives (or none at all).

Commerce Has Been Building on Quicksand

The modern commerce stack is built on transactions—orders, invoices, shipments. We treat these events as primary and items as secondary. This is backwards.

The result? Fractured item data across systems (Shopify says one thing, your ERP says another, retailers see a third version). Manual reconciliation every time you onboard a new retailer. Compliance chaos when GS1 codes don’t match UPCs don’t match retailer SKUs. AI that can’t help because it’s drinking item data “through a straw” via integrations.

Point solutions tried to solve pieces of this: EDI providers move item data around but don’t own it. PIMs organize it but don’t transact on it. ERPs claim to be authoritative but are terrible at commerce-specific attributes. Shopify owns DTC but abdicated wholesale.

Nobody owns the item primitive in wholesale commerce. So nobody can build the platform.

The Item Is the Primitive

Here’s what we believe: Every commerce operation references back to items.

  • A purchase order is a collection of items + quantities + prices + delivery terms.
  • Inventory is items + locations + availability.
  • EDI is items + trading partner mappings + compliance data.
  • Fulfillment is items + routing + packaging specs.
  • Pricing is items + channels + customer tiers.

The item is not just product data. It’s the semantic core of commerce: what it is (attributes, descriptions, images), where it comes from (suppliers, manufacturers, lead times), where it goes (channels, retailers, pricing), how it moves (packaging, shipping, compliance), and what happened to it (transactions, inventory events, returns).

Why Now?

Three things make this possible now that weren’t possible before.

First, the bundling moment has arrived. Point solutions are dying because AI needs context. You can’t build useful AI on partial item data from integrations. You need the complete graph: item → inventory → orders → fulfillment → pricing. The company that owns the item primitive + the data graph around it wins the AI layer.

Second, wholesale is finally digitizing. For decades, wholesale ran on phone calls, spreadsheets, and EDI from the 1980s. COVID accelerated digitization by 10 years. Brands that ignored wholesale now depend on it. Retailers that resisted automation now demand it. The infrastructure layer is up for grabs.

Third, Shopify abandoned wholesale. Shopify won DTC by owning the item record for direct-to-consumer. But they never built for wholesale’s complexity—compliance requirements, multiple pricing tiers, retailer-specific packaging, chargebacks, consignment terms. There’s a $10T wholesale market with no authoritative item primitive.

What We’re Building

Endless is the authoritative item record for wholesale commerce.

We’re not building another PIM that organizes data but doesn’t transact. We’re not building another EDI provider that moves data but doesn’t own it. We’re not building another order management system that tracks transactions but treats items as secondary.

We’re building the item primitive with all commerce workflows on top.

The wedge: Purchase order management + EDI. Brands need to transact with retailers. This forces them to give us authoritative item data. Every transaction validates and enriches the item record.

The expansion: Everything that touches items. Inventory management (items + locations). Fulfillment orchestration (items + routing). Pricing + promotions (items + channels). Compliance + certifications (items + requirements). Supplier relationships (items + sources).

The platform: The commerce data graph. Every brand’s complete item catalog. Every retailer relationship and their requirements. Every transaction across the wholesale channel. The semantic layer that makes AI actually work.

Why This Wins

Network effects compound. Every brand that onboards enriches item data. Every retailer that transacts validates it. Every compliance requirement captured makes the next one easier. The item record becomes more valuable over time.

AI becomes possible. We alone have the complete context—what the item is, where it’s been, where it’s going, what’s in stock, what’s on order, what retailer requirements apply. The AI can work magic because we own the primitive.

Lock-in is structural. Once we’re the authoritative source for your item data in wholesale, we’re not a tool you use. We’re infrastructure you depend on. Every new retailer, every new channel, every new workflow reinforces this.

The Competition Can’t Do This

NetSuite and ERPs were built for manufacturing and accounting—terrible at commerce-specific attributes, can’t move fast enough. Shopify won DTC but never built for wholesale complexity. SPS Commerce has a massive EDI network but just moves pipes—no authoritative item data, no platform vision. Anvyl was smart on purchase orders but treated items as secondary, got acquired before building the primitive. PIMs like Salsify and Akeneo organize item data beautifully but don’t transact—no workflow layer, no commerce graph.

Nobody else is positioned to own the item primitive because nobody else is capturing it at the transaction layer while building the platform on top.

What This Means

If we’re right, Endless becomes the commerce infrastructure layer. Every wholesale brand needs authoritative item data. Every commerce workflow is built on the item primitive. Every AI application in commerce needs the context we have.

We’re not building software to manage your purchase orders.

We’re building the item primitive that wholesale commerce runs on.