Referral Business and Architecture in the Home Services Industry

Referral business, broadly defined as "old customers bringing new customers," is treated here as a form of distribution where users act as agents selling products and earn commissions.

What is Referral?

Distribution essentially means "agency sales of goods"; any scenario where a user sells a product and receives a commission qualifies as distribution.

Who can be the user?

a) Partner: has community resources and sells products professionally
b) Existing customer: purchased before and recommends to others
c) Companies or individual businesses specializing in product distribution
d) ...

E‑commerce Referral Model

The article outlines typical e‑commerce referral flows such as Taobao’s affiliate system, JD Distribution, and various social‑commerce platforms, illustrating a simple "recommend‑once, commission‑once" process based on order amount.

Characteristics of Referral in Service Industries

Unlike e‑commerce, service‑industry referrals often follow a "recommend‑once, commission‑multiple times" pattern due to the long service cycle and risk management considerations.

A comparison table highlights differences:

E‑commerce

Home‑service

Recommendation

Once

Once

Commission

Once

Multiple

Service Cycle

None

Exists

Commission Calculation

Based on order amount

Variable (node‑based)

Operation

Presentation layer

Commission strategy & presentation layer

Service products are diverse (worker leads, client leads, cleaning services, training courses, etc.), and each has flexible commission nodes and amounts.

Design of the Referral Architecture for "DaoJia" Platform

The architecture abstracts three core concepts: Resource (product/lead), Referrer (distributor), and Commission (reward).

System Overview

The backend consists of three subsystems:

Commission platform (distribution platform)

Partner management platform

Shared basic service platform

The frontend (small‑front) provides multiple channels such as mini‑programs, public accounts, and H5, achieving "diverse promotion, unified data".

Commission Platform

The platform is built around the notion of "commission nodes"—specific points in the service cycle where a commission is payable.

@CommissionComputeStrategy(incomeNodes = {IncomeNode.WORKER_BAOMU_AUTHENTICATION, IncomeNode.WORKER_YUESAO_AUTHENTICATION, IncomeNode.WORKER_YUERSAO_AUTHENTICATION})
@Component
public class WorkerClueAuthenticationComputeHandler extends WorkerClueCommissionComputeHandler

Key technical features include:

JSON‑based rule storage with abstract‑class hierarchy.

Validation hooks ("verification points") in abstract rule classes.

Versioning of rule items to handle concurrent calculations.

Commission calculation uses the Strategy pattern to dynamically add computation nodes whenever a new rule is configured.

Rule Configuration and Dynamic Calculation

Rules are configurable, allowing rapid onboarding of new resource types and commission nodes without code changes. The system listens for node events (e.g., worker authentication) and triggers commission calculations accordingly.

Standardized Commission Flow

The main workflow remains stable regardless of added nodes:

Pre‑calculation: estimate future commissions for the referrer.

Eligibility check: monitor service cycle events to determine when commissions become payable.

Commission order generation: compute actual commissions based on rules.

Payout: after business and financial audits, funds are transferred to the referrer's account.

Query: provide full visibility of commission status throughout the service cycle.

Practical Example

Referrer Xiao Huang recommends worker Zhang Aunt for a maternity‑nurse role via a mini‑program. The platform pre‑calculates expected commissions, monitors Zhang’s certification process, generates a commission order once certification is complete, and finally pays out after audit.

Conclusion

By abstracting service‑cycle commission nodes and making commission rules configurable and computation‑driven, the architecture can be applied to any service‑oriented referral scenario, emphasizing thorough business research, logical axis identification, and systematic decomposition during design.