Store management
Manage shop settings, channel availability, operating status, and the complete product catalog from one place.
Marketplace seller operations
Backend maintenance and a renewed seller experience for managing commerce, performance, and growth.

Role
Backend & maintenance
Scope
Emirates LS team
Status
Production
Overview
My work with Emirates LS supported Tokopedia's marketplace environment from a maintenance and backend perspective. This case study expands that foundation into a unified seller command center: daily operations, shop health, reporting, and AI-assisted decisions in one coherent experience.
Seller operations
The seller experience brings everyday operations together instead of making teams move between disconnected tools. It covers the work from creating a product to fulfilling an order and understanding the result.
Manage shop settings, channel availability, operating status, and the complete product catalog from one place.
Create listings, organize categories and variations, improve product quality, and handle high-volume catalog updates.
Track new orders, packing deadlines, shipment progress, cancellations, returns, and warehouse origin.
Keep stock synchronized across sales channels and identify low stock or overselling risk before it affects customers.
Bring chat, questions, complaints, and response performance into the seller's everyday operational queue.
Connect promotions, advertising, affiliates, and live-shopping activity to product and revenue performance.
AI listing assistant
This feature direction reduces repetitive seller work without publishing unverified AI output. It prepares the difficult fields, explains what may be missing, and gives the seller final approval.
The seller starts with a clear product photo instead of a long empty form.
AI identifies the likely item, category, color, material, visible attributes, and possible variations.
The system drafts the title, description, attributes, search terms, price range, and shipping information.
Image quality, missing information, prohibited claims, duplicate listings, and risky content are flagged.
Every generated field remains editable. The seller approves the final listing and stays in control.
Decision intelligence
Reporting should not stop at charts. The product compares the signals behind a result, explains the strongest contributing factors, and recommends what the seller can do next.
Separate traffic problems from conversion problems, then compare price position, stock, delivery speed, ratings, returns, listing quality, and ad efficiency to identify likely causes.
Find products and bundles that match growing search demand, related purchases, category gaps, seasonal movement, and the seller's existing strengths.
Compare product interest, available stock, shipping coverage, competition, and conversion by city or region to guide inventory and campaign decisions.
Show the seller how pricing compares with similar products and whether a discount, bundle, ad, or improved listing is the strongest next action.
Reports
Revenue, orders, fees, discounts, refunds, cancellations, and net payout.
Views, search visibility, conversion, stock cover, returns, and margin.
New and returning buyers, repeat behavior, cohorts, and leading locations.
Spend, attributed sales, ROAS, search terms, and underperforming campaigns.
Packing time, delivery service level, cancellation reasons, and response time.
Ratings, listing quality, policy risks, service quality, and growth readiness.
New versus returning customers, purchase frequency, leading cities, category interests, and retention patterns.
Every recommendation includes the signals behind it, so the seller can understand and challenge the reasoning.
AI drafts and recommends. Pricing, stock movement, promotion changes, and publishing still require seller approval.
Research basis
The operational scope is grounded in documented Tokopedia seller capabilities, including real-time sales and operations, inventory, buyer analytics, TopAds, reporting, and the unified Tokopedia and TikTok Shop Seller Center. The image-to-listing and recommendation features are clearly presented as product direction.
Next case study
More product work across AI, commerce, delivery, payments, internal systems, and Kurdish digital experiences.
All projects