Owning the Buying Workflow

Diving deeper into

Ameet Shah, partner at Golden Ventures, on the economics of vertical SaaS marketplaces

Interview
it’s really about protecting the fortress.
Analyzed 4 sources

The real move is not replacing the POS first, it is making the marketplace hard to dislodge by owning more of the buying workflow around it. For Faire, chat, invoices, payments, and inventory tools make wholesale ordering happen inside one system, so a retailer gets more value every time purchasing data, payment history, and supplier communication stay connected. That makes a rival need to beat not one feature, but the whole operating loop.

  • In B2B marketplaces, the sticky product is the workflow around the transaction. Research on vertical SaaS marketplaces shows the winning pattern is to sit inside sourcing, procurement, logistics, payments, and terms, because that is where buyers and sellers already do the work that leads to the order.
  • Faire has the scale to turn adjacent tools into defense. It reached an estimated $616M revenue in 2023, served 800K+ retailers, operates in 100+ countries, and is expanding into retailer discovery, net 60 lending, inventory management, and operations, which gives it many chances to pull more merchant activity into its own surface area.
  • There are two playbooks for this problem. Faire is leaning toward bundling more native tools, while companies like Shopify and ServiceTitan also use ecosystems and app stores to cover workflow gaps. Native products defend the core by keeping data and actions in one place. Ecosystems defend the core by making the platform the default hub other vendors plug into.

The next step is a fight over who becomes the system that sees every wholesale dollar before it reaches the register. If marketplaces like Faire keep capturing communication, ordering, invoicing, payments, and financing in one flow, they can move upstream from discovery into the merchant back office, and eventually pressure POS vendors from a position of daily habit rather than feature parity.