RedBridge Global Living
Role: Head of IT · Sole author, analyst & document designer · Board-level decision paper for the Group Chairman · Bilingual EN/ZH · 15 sections · 12 decisions · 3 documents
Business Impact
A board-level investment case of this scope — primary data analysis, options appraisal, risk register, phased delivery plan, and full bilingual production — is the kind of engagement a mid-tier consultancy scopes at $25,000–$60,000 AUD. Written solo, in-house, with no external spend.
Analysis of the June and July creative reports surfaced a cost-per-lead spread from ¥417 on the best-performing post to ¥2,923 on the worst, and identified ¥16,666 of July spend across eight posts that produced no leads at all.
The register holds 2,829 leads (55.7%) with no salesperson, no outcome and no reason recorded. Against the Group's own stated cost-per-lead, that represents roughly A$960,000–1,440,000 of acquisition spend with no recorded result. The proposal quantifies the exposure, names the mechanism that closes it, and scopes the fix to ten weeks on existing internal capacity.
Scale at a Glance
Project Summary
| Field | Details |
|---|---|
| Project Name | RGL Ecosystem — Strategic Proposal, Adviser Companion Paper & Technical Backlog |
| 项目名称 | 红桥全球旅居生态平台 — 战略提案、顾问推演配套文件与技术待办清单 |
| Period | Aug 2026 |
| 周期 | 2026 年 8 月 |
| Company | RedBridge Group · prepared for the Group Chairman |
| 公司 | 红桥集团 · 呈报集团董事长 |
| Role | Head of IT — sole author, data analyst, and document designer |
| 职能 | IT 负责人 — 独立撰稿人、数据分析者与文件设计者 |
| Deliverables | Strategic proposal (bilingual HTML, 15 sections + 2 appendices), adviser questions companion paper, technical implementation backlog |
| 交付物 | 战略提案（中英双语 HTML，15 节 + 2 个附录）、顾问提问推演配套文件、技术实施待办清单 |
| Primary Source | Four-in-one inbound lead register — 5,076 rows, 38 columns, 21 Xiaohongshu accounts, 609 distinct posts |
| 主要数据源 | 「四合一」进线台账 — 5,076 行、38 列、21 个小红书账号、609 个不同帖子 |
| Secondary Sources | Product Book V1.0, Group organisational reform papers, two-committee meeting minutes, June & July creative performance reports |
| 辅助来源 | 《产品册 V1.0》、集团组织改革方案、两委会议纪要、六月与七月创意投放数据报告 |
| Benchmark | HM Treasury Five Case Model · PRINCE2 · APM · Governance Institute of Australia board paper guidance |
| 评估基准 | 英国财政部五案模型 · PRINCE2 · APM · 澳大利亚治理协会董事会文件指引 |
| Stack | Self-built bilingual HTML document system — CSS Grid pairing, three-mode language toggle, 12px baseline grid, print stylesheet |
| 技术栈 | 自建双语 HTML 文件系统 — CSS Grid 对照排版、三种语言模式切换、12px 基线网格、打印样式表 |
| Cost Impact | Est. $25,000–$60,000 AUD consultancy cost avoided |
| 成本影响 | 估计节省咨询费用 $25,000–$60,000 澳元 |
Key Highlights
- Sole author of a board-level decision paper — 12,950 English words with full Chinese parity, 15 sections, 2 appendices, and 12 discrete decisions each with a named owner and a stated dependency
- Primary analysis of the Group's own lead register — 5,076 rows across 38 columns, cross-tabulated by front-end status, sales owner, account, campaign, city and outcome
- Independently identified that the register under-records closes by a factor of four — 8 rows marked closed-won against 35 signed clients advised by the business, meaning every conversion ratio the Group quotes about itself was materially wrong
- Reframed the headline finding after challenging the initial thesis — the first draft argued that unassigned leads were the loss; further analysis showed only 264 carried a refusal reason and 2,829 carried nothing at all, so the finding became the missing recorded disposition rather than the missing assignment
- Located the funnel leak precisely — assignment was working as designed (386 of 389 rows marked assigned named a representative); the loss sat two steps earlier, where 1,540 of 1,553 leads marked only as arrived carried no outcome of any kind
- Established that the website form carries 1.3% of lead flow — 64 enquiries against 5,076 leads, with 85% arriving through RedBridge-branded Xiaohongshu accounts, making the acquisition channel the structural constraint rather than the software
- Quantified media waste from the creative reports — ¥16,666 of July spend across eight posts returning zero leads, against a best-performing post at ¥417 per lead
- Researched and applied a governance benchmark — built a 30-criterion rubric across nine categories from the HM Treasury Five Case Model, PRINCE2, APM, PMI and Australian governance guidance, then graded the document against it and published the gaps rather than concealing them
- Restructured the paper around the Group's August organisational reform — the two-president structure, KOS operating units, territorial CRM split, and the platform migration from Zoho to DingTalk, integrating each into the architecture and decision ownership
- Designed the boundary between the platform and DingTalk — identity, communications and approvals in DingTalk; leads, clients and money in the portal — with a researched integration path via H5 micro-app SSO, contacts API and work-notification API
- Separated the twelve adviser questions into a companion paper — removing roughly 2,200 words of non-decision material from the decision document while preserving the analysis for the audience that wants it
- Produced a technical backlog alongside the proposal — schema changes, import pipeline design, retention policy, e-signature architecture and security items, keeping implementation detail out of the board paper while making the build immediately actionable
- Built the bilingual document system from scratch — CSS Grid pairing with baseline-aligned EN/ZH columns, a three-mode language toggle persisted in localStorage, a 12px baseline grid governing every vertical gap, and a print stylesheet
The full bilingual proposal covers architecture, sequencing, commercial and compliance design, and the 12 decisions required to begin.
View Strategic Proposal →Phase Breakdown
- Analysed the four-in-one inbound register: 5,076 rows, 38 columns, 29 Dec 2025 to 12 Aug 2026, spanning 21 Xiaohongshu accounts and 609 posts
- Cross-tabulated front-end status against sales ownership, revealing the five-step funnel decay: 1,553 → 895 → 590 → 489 → 389
- Established data reliability per field — lead volume, paid share, assignment rate and duplicate rate reliable; meeting and close counts not
- Computed duplicate rate on Xiaohongshu ID — 875 of 5,076 rows (17.2%), of which roughly two-thirds sat unflagged
- Verified the wet-signature recommendation against the data: of 8 recorded closes, 6 had a documented in-person meeting and 7 were onshore
- Wrote 15 sections covering evidence, design principles, target architecture, product-to-system mapping, payments, data and CRM, content, compliance, delivery, resourcing, investment and risk
- Mapped the Product Book's twelve products onto five transaction models rather than twelve builds
- Designed the payment architecture across AUD and CNY with rails assigned by transaction value, separating transactional checkout from milestone-billed engagements
- Specified the legal matter boundary at record level with row-level database policies, a classification pass, and a retention policy keyed to matter closure
- Scoped adoption as a deliverable with a named owner and cutover date, addressing the documented failure where the existing portal was offered for testing and never opened
- Researched what constitutes an industry-grade board-level investment proposal across HM Treasury Five Case Model, PRINCE2, APM, PMI, and Australian governance guidance
- Compiled a 30-criterion rubric across nine categories with excellent/adequate/inadequate descriptors and category weightings
- Graded the document against the rubric and published the failing criteria — no quantified investment appraisal, no do-nothing option — rather than omitting them
- Produced a remediation plan tiering the sixteen shortfalls by what unblocks them
- Incorporated the Group's new governance structure: board, two presidents, HR governance committee, two committees, KOS operating units, and the T0–T7 level system
- Reassigned decision ownership from a single principal to the correct function under the new structure
- Reconciled the platform proposal against the Group's move to DingTalk, defining the boundary and researching the integration path rather than arguing against the decision
- Integrated lead recovery SLAs, L0–L3 lead grades, A0–A4 authorisation levels, and per-unit funnel reporting, establishing the platform as the mechanism that makes the reform operable
- Updated the economic case using the Chairman's own stated figures rather than externally sourced estimates
- Built a CSS Grid pairing system wrapping each English block with its Chinese counterpart at runtime, rendering side by side above 1180px and stacked below
- Implemented a three-mode language toggle — English, Chinese, both — persisted across sessions, with each mode using the full available width
- Matched Chinese type size to its English counterpart at every level and baseline-aligned the columns, resolving misalignment caused by differing CJK and Latin font metrics
- Placed every vertical gap on a 12px baseline grid derived from the 1.5 line-height, so paragraph, callout and section spacing tile to the same rhythm as the text
- Audited the prose against a specific AI-writing checklist, reducing a repeated rhetorical construction from 16 instances to 1, and eliminating second-person address entirely
Technical Highlights
Runtime DOM wrapping pairs each English block with its Chinese sibling, tagged by element type so heading, paragraph and callout rows each receive the correct spacing and alignment treatment.
English, Chinese, or both. Each mode releases the width constraints appropriate to it; the both-mode retains two columns for prose while callouts, lists and tables span full width.
12pt Arial at 1.5 gives a 24px line box; every margin in the document is a multiple of 12px, so vertical rhythm holds across both languages and all three modes.
Chinese and Latin faces place their baselines differently within an identical line box. Heading and paragraph pairs use baseline alignment rather than box alignment so the text lines up rather than the containers.
Register analysed across status cross-tabulation, per-account performance, duplicate detection on composite keys, monthly cohort coverage, and outcome distribution by funnel stage.
30 criteria, nine weighted categories, scored with reasoning per criterion and a threshold rule identifying which failures warrant returning the paper regardless of total score.
Adviser analysis split into a standalone companion paper with its own cover, navigation and status; implementation detail split into a technical backlog. Each document addresses one audience.
Page-break control on tables and callouts, cover page break, sidebar suppression, and grid preservation for two-column output.