The GPO was good at the hard part — negotiating pricing — and constrained by the software around it. Members could not easily see their contracted prices, ordering happened outside any system the GPO could observe, and the savings claim that justified membership was assembled manually and imperfectly at renewal.

We modelled members as a hierarchy from the outset, which proved important. A DSO member is a parent organization with many locations ordering independently while volume commitments aggregate upward, and building this as a flat member list would have required an expensive rebuild once the first large group joined.

The catalogue was the largest engineering commitment. Products from multiple distributors and manufacturers, each with different feed formats and product identifiers, had to carry each member's negotiated price and stay current — because a catalogue showing last quarter's pricing is actively worse than none at all. Keeping those feeds accurate is ongoing work rather than a one-time integration.

For ordering we favoured punch-out into distributor systems over asking practices to adopt a new purchasing workflow. Adoption drives the volume the GPO's agreements depend on, and members overwhelmingly prefer to order the way they already order.

The savings engine was built to withstand scrutiny: calculated against a defensible baseline and broken down by category and location, so a member asking what the membership saved them received a number they could interrogate. On the GPO side, participation and contract-compliance data became direct leverage in the next round of vendor negotiations.