How No-Code Platforms Revolutionize Backend Page Production and Boost Efficiency

Background

With e‑commerce traffic growth plateauing, cost reduction and refined operations are critical, leading to a surge in B‑side backend demands. Standardized operational workbenches require consistent page development to ensure product experience, making efficient page production essential.

Problem & Strategy

The goal is to upgrade page production tools for the Taobao backend, focusing on two ideas:

Scene‑centric design: abstract the 86% of form‑list‑detail scenarios into standardized scene templates, reducing marginal costs and boosting overall efficiency.

Pipeline‑style production: treat the entire workflow as an ordered pipeline with standardized outputs for each role, optimizing the whole chain rather than shifting workload to a single stage.

Overall Solution

Collaboration model upgrade : Define standard scene specifications, let products design around scenes, pages build from scenes, and APIs derive from scene configurations, shortening translation steps and improving delivery speed.

Development model upgrade : Provide a no‑code UI generation capability that configures scenes (including interactions, conditional rendering, data binding) and automatically generates required API specifications for backend implementation.

Product standard & experience assurance : End‑to‑end no‑code development guarantees consistent interaction language and product tone, lowering learning costs and enhancing user experience.

Technical architecture

Scene Standardization

Platform architecture centers on scenes: products design around scenes, pages are built from scenes, and APIs follow scene‑defined data structures. Scenes are broken down into two granularities:

Scene materials (template level) such as filter forms, operation bars, tables, and pagers.

Scene controls (component level) like employee selectors or date displays.

Scene materials embed business logic and API requests, forming a "scene model" that standardizes required interfaces and data formats.

Standardization process includes establishing a scene‑specification group, creating a scene‑deposit mechanism for new or customized scenes, and building a scene‑ecosystem that unifies standards across multiple business domains.

Data Standardization

Data standardization consists of model definition, data model generation, and data entity production. Three models are defined:

Gateway model : universal wrapper for all backend APIs, indicating success, errors, and debug info.

Scene model : describes the full set of interfaces for a specific scene, including fixed parameters and fields derived from business models.

Business model : captures domain concepts, objects, and attributes, reusable across multiple pages.

Using these models, developers select business model fields, bind them to scene configurations, and automatically derive API definitions, similar to a template engine.

Backend implements the generated APIs and provides reusable Java classes (e.g., pagination, cascade queries) to convert DO/DTO to VO, reducing development effort. For legacy or third‑party services, mapping and orchestration capabilities transform non‑standard interfaces into the defined API shape.

No‑Code Page Production

The core of no‑code is to generate complete functional pages from standardized scene materials and data models without writing code, addressing low efficiency, quality, and front‑back coordination issues.

Pages consist of UI, interaction, and data; scene and data standardization supply a structured framework, leaving only minimal configuration work. Visual configuration handles scene layout, interaction linking, and data binding, lowering the barrier for non‑frontend developers and improving delivery quality.

Compared with generic low‑code platforms, this solution abstracts non‑UI logic and data integration based on standard scenes, eliminating hand‑coded effort and enabling full‑process no‑code development.

Standard protocol : Extends group low‑code and OneAPI 2.0 protocols with scene, business, and gateway models.

Development assets : Scene materials, controls, and models serve as reusable assets.

Development configuration : UI configuration of scenes, layout composition, interaction linking, and API binding.

Build & release : Generate page schema and API models, extract component dependencies, scaffold code, and publish to CDN or as micro‑modules.

Runtime rendering : Engine parses schema, renders scenes and controls, executes interaction flows, and lets scenes handle API calls.

Backend Development Efficiency Measurement

Existing metrics focus only on code complexity and ignore change impact and cross‑mode comparisons. A new efficiency model normalizes minimal scope complexity and measures total development and integration time:

Define efficiency index = (minimal scope delta complexity / standard page complexity) / (continuous dev time + continuous integration time).

Enhance Hollestad complexity with diff, dependency, and AST analysis to capture change‑induced complexity.

Construct continuous time windows from dev/low‑code/no‑code activity logs.

Create a normalized standard page by sampling scenes across modes.

Design a full measurement pipeline covering source code, low‑code, and no‑code workflows.

Conclusion

The no‑code platform has been built, supporting multi‑role collaboration and reducing translation steps, achieving a 96% coverage of scene business, 39 scene data standards, and 81 business data specifications. The Orca‑Efficiency framework measures full‑stack development and integration, showing a 5× efficiency gain over source code, 1× over low‑code, and a 68.6% overall productivity boost, while ensuring consistent product experience and quality.