Customer Case Study

CompAsia

Rework Service Flows for a refurbished device operation racing toward 140 stores and an IPO

CompAsia Refurbished Electronics Service Flows Malaysia 2026

3 wks POC → Review

6 Workflows Built

30–35 Users in Rollout

5 days RMA SLA Enforced

70+ Stores Coordinated

Tracy Nguyen Implementation Lead

Kevin Long Kim Implementation

Harry Nguyen Business Consultant

compasia.com — B2C storefront

01 / 13

Agenda

What's Inside

01

Company

Company Overview + Industry Context

How CompAsia operates Malaysia's fastest-growing refurb device network — and why it outgrew its own coordination system.

02

Turning Point

Why They Acted Now

A 13× store expansion, an SLA nobody could enforce, and an IPO clock that demands operational proof.

03

Timeline

POC to Production

From first demo in January to a 6-workflow POC in April — two core supply chain processes built in three weeks.

📄 View PoC Plan →

04

Solution I

Order Fulfillment Flows

Production Main Flow + Outstanding Fulfillment Flow — replacing the shared Excel tracker with structured, stage-enforced coordination.

05

Solution II

Post-Sale RMA + Repair

Per-stage SLA timers, a BU decision block, repair sub-flows, and full IMEI-linked case history — replacing the RMA Tracker sheet.

06

BD Toolkit

Before/After + Sales Intel

Reference numbers, refurbished device industry key terms, and how to win deals in this space.

02 / 13

Company Overview

CompAsia — Malaysia's Fastest-Growing Refurb Retailer

CompAsia buys, grades, refurbishes, and resells second-hand iPhones, Samsung Galaxy devices, tablets, and wearables. Founded in 2012, it grew from ~5–6 stores in 2024 to over 70 stores by early 2026, targeting 140 stores by end-2026 ahead of a planned 2027 IPO.

The business runs a B2C channel (Shopify + 70+ physical retail locations) and a B2B channel (corporate device fleet purchases). Apple devices represent 81.6% of sales volume. Extended warranty products are sold under the CompAsia brand.

Before Rework, internal supply chain coordination ran entirely on shared Excel files, WhatsApp group chats, and Zoho. No structured handoff, no SLA enforcement, no way to see where any case was stuck without auditing a spreadsheet.

Tech Stack Before Rework

Shopify Excel / SharePoint WhatsApp Groups Zoho 3rd-Party Repair Tracker

70+

Retail Stores · Q1 2026

~200

Total Staff

4,074

Paid Orders · Q4 '25–Q1 '26

MYR 8.15M

Order Revenue · 3 Months

Device Mix — Paid Orders

Apple

81.6%

Samsung

7.0%

Others

11.4%

Industry Context

Every order is a service case

In refurb, "fulfillment" means finding and QC'ing a unique physical device — not pulling a SKU from a shelf. When the exact device isn't in stock, the process branches into a multi-day sourcing cycle. Generic retail tools break immediately.

The 30-day warranty window is operationally complex

Repair priority flips at 30 days: in-house first within, replacement first after. This decision matrix must be encoded in the workflow — 14 CS agents cannot be expected to remember it on every case.

Excel death spiral hits refurb faster

Multi-team Excel files — one column per department — fail quadratically as order volume and team size grow. A business going from 5 stores to 70+ in 18 months will hit this wall hard, fast, and all at once.

03 / 13

Turning Point

Why They Acted Now

Pressure 01

A 13× Store Count With a 1× Coordination System

CompAsia went from ~5–6 stores in 2024 to over 70 by early 2026 — roughly 13× in 18 months — while the internal coordination system remained unchanged: one shared Excel file, one WhatsApp group, one column per department. Tracy Nguyen (Implementation Lead) identified the failure point immediately: "They're tracking everything on Excel. Columns per department. No owner. No SLA. No alerts."

Pressure 02

The SLA Nobody Could Enforce

CompAsia commits 14 calendar days to customers. Internally: 5 working days from device receipt to return. But with no per-stage SLA system, the team relied on individuals remembering to follow up. Alvin Ho (RMA process owner): "Everyone's being very defensive... we cannot quantify that really it's here. This is going to clear up everything for us and this is a priority."

Pressure 03

The IPO Clock Is Running

CompAsia is targeting 140 stores by end-2026 and an IPO in 2027. Processing MYR 8M+ quarterly revenue through shared spreadsheets and WhatsApp is a structural risk any due-diligence team will flag. The Rework POC was, in Alvin Ho's framing, proof that the operation can be "feasible, effective and practical" at scale — not just functional today.

Decision Journey

2012

CompAsia founded — device trading and refurbishment

2024

~5–6 stores; Excel + WhatsApp coordination manageable

Q4 2025 – Q1 2026

~70 stores; 4,074 paid orders (MYR 8.15M) in one quarter

15 Jan 2026

First Rework demo — CFO Yap Kian Leong; CS + Supply Chain workflow identified as the fit

~8 Apr 2026 🚀

POC payment — 3 weeks, 6 workflows, 6 key users

May 2026

Pricing and implementation terms; production rollout planning

End-2026 → 2027

Target: 140 stores + production go-live; IPO 2027

04 / 13

Implementation

POC to Production

Payment ~8 Apr 2026 · POC review target ~25 Apr 2026 · 3 weeks end-to-end

3-Week POC Timeline — CompAsia × Rework — 3-Week POC plan — 7 working sessions, Mar 30 → May 3 📄 Open PoC Plan (Google Sheets) →

Jan 15, 2026

CRM Demo — Harry + Dolly + Yap Kian Leong (CFO)

B2C retail CRM ruled out. CS & Supply Chain workflow identified as the fit.

9 Apr, 2026

Process Discovery — Jeremy + Alvin Ho

OF flow and RMA flow documented from scratch. SLA sub-windows defined. Solution architecture split: Tracy → OF, Kevin → RMA.

10 Apr, 2026 · Go Live

Full Team POC Demo

Both workflows presented to Jeremy, Alvin Ho, Susan Tan, Lim Teck Shang. Edge cases surfaced: BU approval, IMEI history, round-robin.

~15 Apr, 2026

Kevin + Alvin Ho — RMA Deep Review

Field configuration. IMEI drop-down vs. free-text resolved. Button renaming logic confirmed. Alvin Ho independently created test jobs before this session.

21 Apr, 2026

CS Discussion — Susan Tan + Team

Round-robin assignment. Shift-schedule routing. CS view-only permission layer. CS acceptance checklist distributed.

May 2026

Pricing + Feature Progress

Subscription terms and implementation pricing discussed. Production rollout scoping. Next-wave department briefings planned (Ecom, Offline Retail, Device Service).

What shaped the timeline: Both process owners (Alvin Ho and Jeremy) arrived with documented current-state processes — the RMA Tracker sheet and Outstanding Order file — giving the Rework team a concrete baseline, not a blank canvas. Alvin Ho created test jobs independently before the Apr 15 review, without being prompted.

05 / 13

Needs & Goals

Before Rework vs. After Rework

Before Rework

📋 Outstanding Order Tracking

Shared Excel — one column per department. No case owner. No SLA per row. No alert when a case stalls. Management chases team leads verbally.

🔧 Post-Sale RMA Tracking

RMA Tracker Excel sheet — manual row updates. No per-stage deadline. No overdue flag. 5-day SLA tracked by comparing column header dates.

📡 CS ↔ Supply Chain Handoff

Inbound in Zoho → CS copies info → WhatsApp to Supply Chain group → SC updates Excel. No confirmation, no tracking, no owner.

🔍 Bottleneck Detection

Not possible. Delays identified only when customers complained. Responsibility disputes common — no timestamped record of who moved what and when.

📱 Device History

No cross-case linking — if a device returned for a second RMA, team searched Excel manually and unreliably. No system connected cases to the same device.

After Rework

📋 Outstanding Fulfillment Flow

Each outstanding order is a job card with an assigned Ecom owner, a daily SLA timer resetting on each sourcing attempt, and automatic routing back to Production Main Flow when stock is secured.

🔧 Post-Sale RMA Workflow

Job card per case, SLA timer per stage, automatic routing to repair/replace/refund sub-flows, timestamped history of every action taken across the 5-day internal SLA window.

📡 Job Card-Based Cross-Team Flow

CS creates a Rework job card. All teams update status within that card. No WhatsApp relay. No separate Excel update required. CS has read access to all workflow stages without calling SC.

🔍 Stage Statistics

Ticket count per stage, on-time vs. overdue rate, average dwell time per step — across all open and closed cases. Bottlenecks surface without a meeting.

📱 IMEI Dataset

IMEI as the unique device key. All job cards linked to the device record. CS searches by IMEI — full warranty history visible before responding to the customer.

The "Before" in the wild

What "tracking on Excel" actually looked like: a single Outstanding Order workbook where every team wrote into its own columns, plus a hand-built dashboard re-counted by formula. Functional at 5 stores — a liability at 70+.

CompAsia's current Outstanding Order Excel — Shopee OF sheet with one column per team — The live workbook: status, order number, marketplace ID, dates, and free-text "Feedback — Input by Production" / "Input by eCom" columns where IMEIs and notes are typed by hand. No owner, no SLA, no validation. *Click to enlarge.*

Excel dashboard counting pending orders by aging bucket and status — Aging counted manually into 0 / 1–5 / 6–10 / >10-day buckets per status. A snapshot that goes stale the moment someone edits a row — and tells you a case is late, never *why* or *who owns it*.

06 / 13

Service Flows · Part 1

Order Fulfillment Flows

Production Main Flow + Outstanding Fulfillment Flow — replacing the shared Excel Outstanding tracker

Production Main Flow · Stage Sequence

Stock Check

Picking by Production

Fulfillment / QC

CS Confirm with Customer

Fulfillment

Ship to Customer

Done

Outstanding Fulfillment Flow · Sub-flow (triggered when stock = not available)

Check Stock

Assign Serial

↩ Back to Production Main Flow

→ Refund Process

Production Main Flow

Every Shopify order enters this flow. The production team performs a stock check (confirming the specific device — iPhone 100% or SE policy), picks and QCs the device, and routes through CS confirmation if colour/capacity differs from the order. SLA timer per stage. Round-robin auto-assignment to the production team.

Why it matters: 4,074+ orders per quarter cannot rely on individuals knowing which device maps to which order and manually updating Excel rows. Every handoff is now timestamped. Every SLA breach surfaces automatically.

Outstanding Fulfillment Flow

Triggered automatically when a stock check returns "not available." The Ecom team logs each daily sourcing attempt — the SLA timer resets on each attempt (24-hour cycle). When stock is secured, the job card routes back to Production Main Flow. If no stock can be found, it routes to the Refund Process.

Before: The "Outstanding" column was a black hole — a device could sit there for days with no owner, no deadline, no alert. "Everyone writes into their own column. Nobody can see who owns it." — Tracy Nguyen

Before → After · Order Fulfillment

CompAsia's current Outstanding Fulfillment flow — hand-drawn process map — The Outstanding Fulfillment process as it existed — a static Visio map covering stock tiers, QC, repair, and refund branches. No live state, no SLA, no owner. *Click to enlarge.*

Production Main Flow on Rework Service Flows — full flowchart — The same logic, rebuilt as a live **Production Main Flow** in Rework — every stage executable, with durations and the Outstanding sub-flow branch wired in.

Production Main Flow zoomed in — Stock Check, Type of order, Stock Check Results, Outstanding Flow branch — Zoom-in: the stock-check decision splits into *Exact match*, *Different colour/upgrade*, and *Not available* — the last routing automatically into the Outstanding Flow sub-flow.

07 / 13

Service Flows · Part 2

Post-Sale RMA + Repair Sub-flows

Per-stage SLA timers, BU decision block, repair routing, and full timestamped case history — replacing the RMA Tracker sheet

Post-Sale RMA Flow · Stage Sequence · Internal SLA: 5 working days

Arrange Device Collection

Device Arrival Confirmed

Unbox & Verify IMEI

QC / Verify Issue

Decision Branch

Ship to Customer

Done

Decision Branch

After QC, the job card routes automatically based on issue type and purchase date:

No Issue → Ship to Customer

In-House → In-House Repair sub-flow

Outsource → Outsource Repair sub-flow

Replace → BU Decision → CS Confirm → Ship

Refund → Refund Process sub-flow

30-day rule: Priority matrix flips at 30 days post-purchase — encoded in the routing logic, not left to agent memory.

BU Decision Approval Block

When a case requires a replace or refund decision, the job card auto-assigns to the designated BU representative (Ecom, Offline Retail, or Device Service).

The BU rep must fill in their decision (Replace / Refund / Repair Outsource) with a mandatory comment before the card can advance. Decision logged with timestamp.

Before: BU decisions via WhatsApp or email — untracked, no escalation path if no reply arrived for days.

Repair Sub-flows + Refund

In-House Repair

Device received by technician → repair in progress → complete or fail. Links back to parent RMA job card.

Outsource Repair

Device sent to 3rd-party → tracking reference logged → device returned → QC → back to RMA main flow.

Refund Process

Shared sub-flow triggered from either Outstanding Flow or Post-Sale RMA. Finance approval → CS notification. Same process regardless of entry point.

"The biggest thing for us right now is to highlight where cases are getting stuck — where is it that the highest delay or dwelling time of devices?"

Alvin Ho · Supply Chain Senior Manager · Apr 2026

Before → After · Post-Sale RMA

CompAsia's current Post-Sale RMA process flow for extended warranty (replacement priority) — The full RMA process as a static diagram: customer return → QC → 30-day check → repair/replace/refund branches, with the BU decision and outsource-repair loops. Accurate, but un-executable and impossible to track live. *Click to enlarge.*

Post Sales RMA Kanban board in Rework with stages and overdue SLA flags — Live **Post Sales RMA** board — every case a job card moving across stages (Device Collection → In transit → Unbox/IMEI → QC → Confirm Customer), each with an SLA deadline and automatic overdue flag.

RMA job card detail view showing order number, purchase date, 30-day check, model, IMEI — A single job card: order number, purchase date, the **"within 30 days?"** field that drives routing, model, capacity, colour and **IMEI** — all captured at collection and timestamped per block.

08 / 13

Service Flows · Part 3

Visibility, SLA Enforcement + CS Integration

Stage Statistics, IMEI dataset, and round-robin queue management across 14 CS agents

📊

Stage Statistics

Configured across all active workflows. Shows ticket count per stage, on-time vs. overdue rate, and average actual duration vs. configured SLA. Management sees bottlenecks at a glance — no meeting required.

"Everyone's being very defensive... we cannot quantify that really it's here. This is going to clear up everything for us."
— Alvin Ho

🔗

IMEI Dataset + Device History

IMEI as the unique device key. Every RMA job card links to the device IMEI record. CS searches by IMEI to see all prior job cards — dates, issue descriptions, repair outcomes, and resolution types — before responding to the customer.

"One device could have multiple job cards... you want to review the customer's profile in regard to this device's history."
— Lim Teck Shang

⚖️

CS Round-Robin + Queue Management

When agents set "active" status, new cases distribute evenly across the active team. CS team lead sees queue depth per agent in Kanban. CS has read-access across all workflows — agents can check any order or RMA status without messaging Supply Chain.

"Do we need to assign manually to each one of them... or is there a round-robin?"
— Susan Tan, CS Team Lead · Apr 21 2026

Production Main Flow table view — job, tags, current block, Shopify date, status, type of order, serial, created by — The same workflow as a spreadsheet-style grid: job, tags (KPDNKK, Pending Refund), current block, Shopify date, status, order type and assigned creator — sortable, filterable, exportable. The Excel mental model preserved, but now backed by live SLA-enforced state. *Click to enlarge.*

Kanban columns Outstanding Flow and Picking with job cards, tags, deadlines and overdue counts — Each pipeline column shows its live cards — device, assignee, tags (KPDNKK, Pending Refund), deadline and an overdue count. Where cases pile up and how long they have dwelt is visible without opening a single row.

Assignment Rules dropdown: Auto Assign, Open Job for Claim, Users Decide, Keep Previous Assignee — Assignment is configurable per block — **Auto Assign** (round-robin), **Open Job for Claim**, **Users Decide**, or **Keep Previous Assignee**. CS picks the model that fits each stage rather than being locked into one.

09 / 13

End-User Adoption

Acceptance Checklist + POC Testing

Acceptance Checklist

A structured Rework sign-off document, divided by process owner — Jeremy (Outstanding Fulfillment) and Alvin Ho (Post-Sale RMA). Each tab includes step-by-step test cases from login through sub-flow triggering and SLA verification. A Q&A section at the bottom collects open questions and adjustment requests.

✓ Login and navigate the service menu

✓ Create a job card and complete each stage

✓ Trigger sub-flows (repair, refund, outstanding)

✓ Verify SLA calculations and overdue flagging

✓ Completion percentage tracked via built-in formula

OF Checklist Owner

Jeremy

SC Senior Manager, Outstanding Fulfillment

RMA Checklist Owner

Alvin Ho

SC Senior Manager, Post-Sale RMA

POC Testing Approach

After the Apr 10 full-team demo, each process owner received owner-level access to test all roles within their workflow using one login — simulating the full team's journey.

Alvin Ho independently created test jobs before the Apr ~15 review session — arriving with specific questions about field configuration, button renaming, and IMEI drop-down logic.

Next Steps Confirmed

Next Mid-week check-in with Jeremy (OF) and Alvin Ho (RMA) before formal Friday review

Next Briefings for Ecom, Device Service (Cheryl), and Offline Retail teams on Rework scope

Next Individual user accounts for each team member post-POC sign-off (target: 30–35 users)

Next Shopify API integration feasibility — auto-create Rework job from new Shopify order (post-POC scope)

Proof of adoption

Rework — CompAsia System Acceptance Checklist, Outstanding Fulfilment tab, 87.1% progress — Jeremy's Outstanding Fulfilment tab — structured test items (login → navigate → flow diagram → sub-flows → SLA), each marked Pass/Fail with a live progress formula. Separate tabs for Alvin (RMA) and Susan & James (CS).

Customer feedback and question list tracking feature requests, questions and issues during the POC — Every question, feature request and issue raised by Jeremy, Alvin and Susan — triaged by type and status (Solved / Developing / Not support). Live items like Kanban filtering, business-hours SLA and Quick Sight API are scheduled into the next phase.

10 / 13

Summary

Before & After Summary

Category	Before Rework	After Rework
Outstanding order coordination	Shared Excel — one column per department; no owner; no SLA; no alert when cases stall	Outstanding Fulfillment Flow — job card per order; daily SLA timer; auto-routes to refund or back to production when resolved
Post-sale RMA management	RMA Tracker Excel sheet — manual row updates; no per-stage deadline; no overdue flag; 5-day SLA unenforced	Post-Sale RMA workflow — per-stage SLA timers; auto-routing to repair/replace/refund sub-flows; full timestamped case history
CS ↔ Supply Chain handoff	WhatsApp group → manual Excel update → no confirmation or tracking	Job card created by CS; all teams update in Rework; no chat relay; no separate Excel update required
Bottleneck detection	Not possible — Excel audit required; delay attribution unresolvable; "everyone's being very defensive"	Stage Statistics — dwell time per stage, on-time vs. overdue rate, visible without a meeting
SLA enforcement	5-day internal target tracked by comparing column dates; no system alerts	Per-stage SLA configured; overdue cards surface in Kanban; CS and SC managers see late cases automatically
Device history	No cross-case IMEI linking — same device returns with no prior history visible to the agent	IMEI dataset links all job cards to the device; CS searches by IMEI; full warranty history in one view
Repair routing + BU decisions	Verbal decisions via WhatsApp/email — untracked; no escalation; delays when BU didn't respond	BU Decision approval block — auto-assigned to BU rep; decision required to advance card; full audit trail
Tools	Excel + SharePoint, Zoho, WhatsApp groups, separate repair tracker — siloed and unconnected	Rework Service Flows — 6 workflows, unified across Customer Excellence and Supply Chain

11 / 13

BD Toolkit

Win More Deals Like This

Industry Key Terms

Outstanding Fulfillment

Orders where the exact requested device isn't immediately in stock — enter a multi-day sourcing cycle. The Outstanding Fulfillment Flow manages this as a structured, daily-SLA loop.

Post-Sale RMA

End-to-end warranty and return process. Internal SLA: 5 working days (device receipt → return shipment) against 14 calendar-day customer commitment. The gap absorbs customer response delays.

30-Day Window

Policy boundary separating two repair priority matrices. Within 30 days: in-house repair first, then replacement. After 30 days: in-house then outsource, then replacement. Must be encoded in the workflow — not left to agent memory.

IMEI

The 15-digit hardware ID embedded in every device. In refurb, IMEI is the only reliable way to link a physical device to its order record, warranty history, and repair history — across multiple cases and years.

BU (Business Unit)

The commercial/sales team (Ecom, Offline Retail, Device Service) that owns the revenue relationship. BU decides replace/refund/repair. Supply Chain executes. They do not make BU decisions unilaterally.

Acceptance Checklist

Rework's structured POC sign-off document. Step-by-step test cases (login → navigate → create → move → verify SLA → trigger sub-flow) plus Q&A section. Both training guide and go/no-go gate for production rollout.

BD Insights

Open with the SLA question, not the Excel question

Every refurb operator knows they're on spreadsheets — that's not news. Ask: "When a device comes in for repair, how do you know if a specific case is on track right now?" The gap between SLA commitment and visibility into it opens the conversation.

"Everyone's being defensive" = the buy signal

When a prospect says delays can't be proved because "everyone points fingers" — name it directly: "What if we could show you average dwell time at each step? So the late-case conversation starts with data, not accusations?"

Sell the Outstanding Flow as the anti-commitment-breach tool

Ask: "How do you make sure every outstanding order is actively worked, every day, until resolved?" If the answer involves checking an Excel sheet, there's a deal.

The IMEI dataset is the differentiation proof

Ask: "If a customer calls about a device repaired twice before, how does your CS agent see that history in under 30 seconds?" Generic ticketing tools treat each case as independent. The IMEI dataset doesn't.

IPO-track businesses move fast

CompAsia signed within weeks of a first demo — operational auditability is a board-level requirement, not a nice-to-have. Lead with: "How would you demonstrate SLA compliance to a due diligence team today?"

The expansion conversation is built into the first deal

CompAsia's POC covers SC + CS. Ecom, Offline Retail, and Device Service are already identified as next-wave. Start with the most painful process (usually RMA or Outstanding), prove the model, then expand laterally.

12 / 13

Conclusion

CompAsia didn't just get a case management tool. They got an operations layer built for a business racing toward 140 stores and an IPO.

From the moment a customer places an order to the moment a returned device ships back repaired — every stage is now tracked, every handoff is timestamped, and every SLA breach surfaces automatically. The team that spent hours chasing Excel rows and WhatsApp threads can spend that time resolving cases instead.

3 wks POC → Review

6 Workflows Configured

30–35 Users in Initial Rollout

5 days RMA SLA Per Stage

70+ Stores Coordinated

REWORK · CUSTOMER CASE STUDY · COMPASIA · 2026

13 / 13