Solutions/Ads optimization
Optimize ad spend on Shopee, Lazada, and TikTok Shop
Marketplace ad systems are tuned for GMV (hence the name GMV MAX), not your true profit. This can lead to unprofitable decisions and lost margin. DataGlass models True ROAS per campaign, flags the spend that is not paying back, and surfaces the budget moves that grow profit.
01/Problem
What sellers see
The platform dashboard shows a clean, high 40.0 ROAS and a green ACOS. However, the bank account tells a different story. E-commerce platforms make money from taking commission on revenue. Their incentive is to maximize order volume, not your profit. They do not know your COGS, your fees, or your margin. The more ads they show, the more volume there is, and the more money they make. You, on the other hand, are wasting money on ads when they are no longer profitable. The platform also shows raw ROAS — which includes canceled orders, duplicated attributions, and fees that should have been deducted. That inflates the number the seller sees, leading them to believe it is a high return.
02/Detection
What DataGlass detects
DataGlass reconstructs every ad-attributed order down to the SKU, deducts COGS, marketplace fees, vouchers, and fulfillment, and shows True ROAS per campaign. This is the very first layer of our decision making process: having clean data to make decisions on. Moreover, dataglass also compute the full unit profit breakdown of all skus, allowing us to work directly on profit not revenue. This is a very important distinction as your revenue can grow while profit shrinks.
- Campaigns with True ROAS below break-even after full cost
- Keywords pulling traffic but converting on low-margin SKUs only
- Ad groups where Shopee or Lazada keyword expansion has bled margin
- Daily budget pacing — over-spend before the profitable hours of the day
03/Action
Recommended actions
We take the above cleaned signal data and build a mathematical model to predict tomorrow's latent demand. By this construction, we optimize tomorrow's budget cap to maximize profit. We also account for downside risks, since the e-commerce marketplace landscape is uncertain and shifting. Risk is baked directly into the optimization to protect against losses on a bad day, giving stable long-run profit-maximizing performance.
- 01
Shift budget to profitable campaigns
Re-allocate spend from campaigns with negative contribution to those with the highest expected lift.
- 02
Adjust bids to break-even ACOS
Set bids that respect the SKU's break-even ACOS, not the platform's suggested target.
- 03
Restructure aggressive expansions
Tighten match types and negative-keyword lists where the platform's expansion is the source of the waste.
04/Platforms
Available on these platforms
05/Glossary
Concepts in this solution
Backed by research
The DataGlass research that grounds the recommendations on this page.
Working paper · May 2026
From Gut Feel to Posterior Inference: A Research Article on the DataGlass Decision-Intelligence System for E-Commerce Ad Budget Allocation
A rigorous public communication of the DataGlass system for daily ad-budget allocation on platform-controlled marketplaces — the analytical reasons rolling-mean heuristics fail, the Bayesian + bandits-with-knapsacks methodology, and the empirical 18–24% portfolio-profit lift.
Read paper
Working paper · May 2026
Prediction and Risk Optimization Under Uncertainty: A Cross-Domain Meta-Review of Methods in Finance, Operations, Causal Inference, and E-Commerce Decision Intelligence
A structured meta-review (213 primary works, 254 references) arguing that mature decision systems across finance, operations, insurance, energy, healthcare, causal inference, and e-commerce share four primitives — calibrated probabilistic models, coherent risk-aware objectives, explicit operational constraint sets, and principled exploration. Eleven worked cases ground the framework, with the DataGlass marketplace ad-budget system as the connecting tissue.
Read paper
Research report · May 2026
Decision Intelligence for E-commerce: How Retailers Optimise Pricing, Forecasting, Inventory, Promotions & Personalization
Pricing, forecasting, inventory, promotions, and personalization — a deep technical survey of the techniques large retailers use, the variants that matter, and how to deploy them.
Read paper
04/FAQ
Frequently asked
Platform dashboards show attributed revenue. DataGlass shows profit after fees, COGS, vouchers, and fulfillment — the actual number that determines whether the campaign was worth running.
DataGlass surfaces recommended bids and flags lines that are losing money. You decide when to deploy. Auto-bid loops are on the roadmap and will be opt-in.
Profit-aware bid adjustments target campaigns and keywords that are losing money, not the ones driving organic ranking. DataGlass shows the trade-off explicitly so the decision is informed.