Case Study — E-Commerce Analytics System

When Creative Strategy Meets Hard Data

"I bridge the gap between creative storytelling and hard-data profitability."

As a Digital & Brand Marketing Manager overseeing Rp 7.7B+ in total ad budget across TikTok Shop and Shopee, I built this full analytics system from scratch — in Google Sheets — to ensure every rupiah of budget was allocated based on real margin data, not just surface-level GMV. No third-party BI tool. No data team. Just structured thinking, advanced formulas, and a relentless focus on profitability.

Role Digital & Brand Marketing Manager
Industry Beauty / Personal Care (FMCG)
Channels Managed TikTok Shop + Shopee
Tools Google Sheets — ARRAYFORMULA, QUERY, VLOOKUP, IFS
Scope 3 trackers · 5 months · 43 active SKUs
Data Notice Figures anonymized & proportionally scaled · Brand identity withheld
Total Budget Managed
Rp 7.7B+
Across TikTok + Shopee · 5 months
Peak Monthly GMV
~Rp 3.5B
Combined channels · Jan 2026
Top Video ROAS
14.4×
TikTok GMV Max · top creator
Margin Risk SKUs Caught
38 / 43
Platform fee >15% settlement — auto-flagged
⚠ All brand identifiers & exact figures anonymized. Metrics are proportionally scaled from real operational data (Jan–May 2026).
00

The Problem

Context

A growing beauty brand was running simultaneously on TikTok Shop and Shopee, managing 43 active SKUs and an affiliate program through GMV Max — with a total 5-month budget exceeding Rp 7.7 billion.

Despite strong top-line GMV, there was no clear view of which products were actually profitable after platform admin fees, affiliate commissions, and seller discounts were deducted. Revenue was being mistaken for business health.

All data lived in isolated raw exports from three separate seller dashboards. There was no infrastructure to connect them, no threshold logic to trigger decisions, and no consistent method to evaluate affiliate video efficiency.

Critical Questions That Couldn't Be Answered

Which SKUs are margin-positive after all fees? Which affiliate videos deserve more ad budget? Is GMV growth translating to actual profit? How much is admin fee eating into settlement? Which creators drive the highest ROAS?

My Approach

I designed and built 3 interconnected Google Sheets systems — each solving one layer of the analytics problem — with automated formulas that refresh whenever fresh exports are pasted in. No BI tool license, no data engineer, no dashboarding software required.

Self-service — zero formula knowledge needed Auto-refresh on raw data paste Decision-ready output ARRAYFORMULA + QUERY + VLOOKUP + IFS
01

The 3-Layer System

01

Daily Revenue Tracker

Unified dashboard aggregating daily GMV from TikTok Shop and Shopee. Tracks MoM delta, channel revenue mix, ads spend efficiency, and cost-per-conversion — updated by pasting raw platform exports into designated input sheets.

ARRAYFORMULA MoM % delta Multi-platform merge Daily granularity
02

GMV Max Video Tracker

Evaluates TikTok affiliate videos at the individual video level. Pulls creator data from 3 separate raw exports, calculates ROAS per video, and auto-flags scale / optimize / kill with a month-filter dropdown for period comparison.

VLOOKUP cross-sheet QUERY function ROAS scoring Decision automation
03

SKU Margin Analysis

Calculates true per-SKU profitability after deducting: COGS, platform admin fee, affiliate commission, seller brand discount, and platform subsidy. Auto-classifies via multi-condition IFS into Winning / Potential / Slow Moving with margin health flags.

IFS multi-condition Settlement calc Fee waterfall Auto status badges
02

Revenue Performance Analysis

Monthly GMV — TikTok Shop vs Shopee (Jan–May, indexed Rp M)
Daily TikTok GMV Trend — Campaign Spikes Visible (Jan–May)
📉 Channel Decline Pattern
TikTok GMV dropped −57.6% Jan→May. Shopee held stronger at −49.4% — suggesting TikTok is more sensitive to ads-spend fluctuation.
✅ Shopee ROAS Advantage
Shopee ROAS ran consistently 2–3× higher than TikTok with lower CPC. Data supported a budget reallocation recommendation.
🔥 Campaign Spike Behavior
Daily TikTok GMV peaked at Rp 209M on double-date campaign days vs. organic baseline of ~Rp 30–45M — a 5–7× lift pattern visible in the tracker.
💡 Spend Allocation Finding
Initial spend was 80% TikTok / 20% Shopee — but Shopee produced proportionally stronger ROI. Rebalancing flagged as immediate recommendation.
📊 5-Month Totals
TikTok: Rp 6.28B GMV · Shopee: Rp 5.09B · Combined monthly avg: ~Rp 2.27B
03

Affiliate Video Performance

The GMV Max tracker evaluates performance at the individual video level. Each video gets a ROAS score derived from cross-referencing 3 separate raw data exports (campaign spend, creator data, video GMV) — and is automatically routed to a scale / optimize / kill decision. Creator IDs anonymized below.

Top 10 Creators by GMV Contribution (Feb, indexed)
Video Decision Matrix — ROAS × vs Ads Spend (indexed)
Creator Product Type GMV (indexed) ROAS Ads Spend (indexed) System Decision
04

SKU-Level Margin Analysis

Revenue Waterfall — Where Gross GMV Goes (Portfolio Average)
⚙ Technical Build Notes

The fee waterfall is calculated using a multi-step settlement formula built in Google Sheets. Each deduction layer (seller discount → platform admin fee → affiliate commission → COGS) is computed sequentially, enabling the system to isolate exactly where margin is lost before any ad budget decision is made.

ETL process follows a Power Query-style paste-and-refresh workflow: raw settlement exports are pasted into input tabs, and all downstream calculations update automatically via ARRAYFORMULA chains — no manual recalculation needed.

The Decision Framework logic uses nested IFS with 5 concurrent conditions evaluated per SKU: sales volume tier, net margin threshold, admin fee rate, seller discount level, and HPP ratio — outputting a single actionable status badge.

🚨 Auto-Flag — Admin Fee >15%
38 of 43 SKUs had platform admin fee exceeding 15% of settlement — the system's critical threshold. These are automatically tagged as "Margin Risk" before any ad spend is factored in. This flag is designed to prevent budget being allocated to SKUs that appear profitable at GMV level but are margin-compressed at settlement level.
🚨 Auto-Flag — Seller Discount >40%
37 of 43 SKUs ran at seller discount exceeding 40% — the system's danger threshold for discount depth. These are flagged as "Diskon Tinggi (Danger)" — a signal that the brand may be buying GMV at the cost of actual margin, and that pricing strategy needs review before further scaling.
✅ Top Margin SKU
The 5-unit bundle variant achieved 62.7% net margin — highest in the portfolio — driven by a favorable HPP ratio and strong retail price anchor. Identified by the system as the #1 candidate for budget scaling.
⚡ Danger Zone Caught Early
2 SKUs actively being scaled were flagged as "Berbahaya (Margin Tipis)" — margin near zero after all fees. The system surfaced these before additional ad budget was pushed, preventing further wasted spend.
🚨 Flag Condition 1
Admin Fee Rate > 15% of settlement → auto-tag: Margin Risk. Applies regardless of GMV or order volume.
⚠ Flag Condition 2
Seller Discount > 40% → auto-tag: Diskon Tinggi (Danger). Triggers pricing review recommendation.
SKU Code Product Category Qty Sold Avg Sell Price Net Margin % Platform Fee Rate Seller Discount % Sales Status Margin Health
05

Automated Decision Framework

The system doesn't just report — it produces clear operational signals. Every SKU and every affiliate video is automatically routed to a recommendation based on multi-condition IFS logic evaluated in real time.

🚀 Scale — All Conditions Met
  • Orders ≥ 500 in period
  • Net margin ≥ 30%
  • Video ROAS ≥ 10×
  • Affiliate fee ≤ 10%
  • Platform fee rate ≤ 15%
→ Increase ads budget, protect stock levels, replicate winning creative format.
⏳ Optimize — Mixed Signals
  • Orders 100–499
  • Net margin 15–29%
  • Video ROAS 5–9.9×
  • High discount flag active
→ Test new creatives, review sell price, reduce discount depth before scaling.
🛑 Kill / Pause — Red Flags
  • Orders < 100 or zero
  • Net margin < 15% or negative
  • Video ROAS < 5×
  • HPP > 50% of sell price
→ Pause all ads, stop affiliate push. Review pricing or SKU viability.

06

What This System Enabled

Before the System

Budget decisions were made on raw GMV alone — no visibility into whether revenue translated to settlement after fees. Platform admin costs, affiliate commissions, and seller discounts were siloed or untracked.

Affiliate video evaluation required manually cross-referencing 3 separate exports. No standardized ROAS threshold existed for scale vs kill. Decision-making was slow and inconsistent.

After the System

Any team member can paste fresh platform exports and instantly receive: a margin-ranked SKU table, creator ROAS leaderboard, and daily GMV trend — with zero formula knowledge required.

The tracker caught 2 SKUs being actively scaled at near-zero margin after all fees — enabling course-correction before significant additional spend was wasted.

Reporting time ↓ ~80% Per-SKU margin visibility Self-service for any team member Proactive risk flagging Rp 7.7B+ budget managed with data confidence