SaaS Distribution Channel: Partner Deals to $100M ARR

What It Covers

Zhong Xu, cofounder of Deliverect, explains how he scaled a restaurant SaaS platform to 80,000 locations across 50 countries and nearly $100M ARR by prioritizing distribution through POS partnerships, running a manual MVP before writing code, and now racing to build an AI intelligence layer.

Key Questions Answered

• •Wizard of Oz MVP: Validate demand before writing a single line of code by fulfilling orders manually. Deliverect's founders personally inputted orders into restaurant systems for the first 50 customers, charging just $50/month, until volume justified automation. This approach confirms real willingness to pay without wasting months on unvalidated engineering.
• •Partner-led distribution over direct sales: Rather than signing restaurants one by one, Deliverect partnered with POS vendors who already served hundreds of restaurants each. Signing 10 POS companies generated more monthly customer additions than direct outreach could. Eliminate channel conflict by always attributing customers to the partner, even when leads arrive directly.
• •Early adopter exchange formula: Offer steep discounts or extended free periods, but always extract two things in return: a reference case with video testimonial, and introductions to four or five industry peers. This converts each early customer into a distribution node, compounding social proof and referrals without additional sales spend.
• •Expand geographically before incumbents form: Deliverect opened 10 offices in a single quarter during COVID, entering 50 countries simultaneously. The logic: displacing an established local incumbent is harder than being first. Accepting a number-two position in any market is survivable; being number three is not. After seven years, Deliverect holds number one or two in every market entered.
• •AI moat requires owning the intelligence layer, not just the pipes: Connectivity infrastructure alone commoditizes over time. Deliverect's strategy focuses on becoming the decision-making layer — auto-adjusting menus, pricing, and promotions using AI agents — while installing guardrails that require human approval before autonomous changes go live, protecting brand reputation from hallucinated or unethical outputs.