Solutions/Optimize e-commerce profit
Optimize e-commerce profit across Shopee, Lazada, and TikTok Shop
Optimizing e-commerce profit means naming one objective, listing the decisions that move it, and pricing each decision against the constraint that binds it. Marketplaces optimize a different objective — gross merchandise value, which is what they take commission on — so a seller who follows the platform’s recommendations is optimizing someone else’s function. DataGlass reconstructs contribution per SKU from order lines across all three platforms and ranks every available move by expected profit.
01/Problem
What sellers see
Three seller centres, three definitions of revenue, and no number anywhere that answers "did this month make money?" Each platform optimizes and reports on gross merchandise value, because commission is charged on GMV — so the recommendations a seller receives are optimal for an objective that is not theirs. The result is a business run on five or six proxies at once: ROAS on the ads tab, conversion rate on the product tab, stock cover on the inventory tab, none of them commensurable, and none of them profit. Profit optimization is not a harder version of this. It is a different exercise: a single objective, a list of decisions that move it, and a stated constraint on each.
- Platform-reported ROAS looks healthy on every campaign while net profit falls
- A Target ROAS set at break-even, which targets exactly zero profit rather than protecting it
- Adjusting a Target ROAS that is not the binding constraint, so nothing changes and the lever looks weak
- The same SKU priced independently on three platforms, with three different contribution margins nobody has computed
- Bestsellers that are the lowest-margin products in the catalogue once commission, vouchers, and the free-shipping cost share are attributed
- Stock decisions made on units sold rather than on contribution at risk
- Selling mostly on Shopee? The per-SKU break-even playbook for one platform
02/Detection
What DataGlass detects
Optimization requires an objective you can compute. DataGlass reconstructs contribution per order line — selling price less cost of goods, category commission, transaction and payment fees, seller-funded vouchers, the free-shipping cost share, fulfillment, and a returns reserve — on Shopee, Lazada and TikTok Shop, then joins the three into one canonical catalogue so a product is one row rather than three. That is the objective function. Everything below is measured against it rather than against a platform proxy.
- Products whose contribution margin is negative at their current price once every platform fee is attributed per unit
- Campaigns whose marginal return has fallen below the product’s true contribution while the reported average still clears the target
- Target ROAS values set at or below the product’s break-even ROAS — the line where profit is zero for every volume
- Campaigns where the daily budget is the binding constraint, so the entered target is inert and adjusting it changes nothing
- The same canonical product carrying materially different contribution margins across platforms, unpriced
- Stock positions where the contribution at risk from a stockout exceeds the holding cost of covering it
03/Action
The five decisions that move e-commerce profit
A profit objective is only useful if something can be done about it. These are the decision variables — the things a marketplace seller can actually set — with the constraint that binds each. DataGlass ranks the available moves across all three platforms by expected contribution gained, and attaches the arithmetic to each one so the recommendation can be audited before it is deployed.
- 01
Ad target, bounded by the profit optimum rather than break-even
Break-even ROAS is the line where the whole contribution of a click goes to the auction, so it earns exactly zero at any volume. The profit-maximizing target is break-even marked up for the demand the ads did not create, for the fact that winning more clicks reprices the clicks you already had, and for the gap between the average click and the marginal one.
- 02
Ad budget, which is the only lever denominated in money
A Target ROAS fixes a price and says nothing about volume — click count cancels out of it. The budget is therefore the only instrument that bounds quantity, and the only one that can bound a loss. Knowing which of the two is currently binding is a single division: entered target divided by reported ROAS.
- 03
Price, against measured elasticity rather than a margin rule
A uniform target margin across a catalogue is not optimization, because demand does not respond uniformly. Contribution is maximised where marginal revenue meets marginal cost per product, which requires an estimate of how each product’s demand responds to price — including across platforms, where the same product can carry very different fee loads.
- 04
Assortment, judged on contribution rather than velocity
The products that sell most are frequently not the products that fund the business. Ranking the catalogue by contribution — not units, not GMV — usually reveals a tail of high-velocity SKUs that are cash-flow-negative once every fee is attributed, and a quiet middle that pays for everything.
- 05
Stock, priced as contribution at risk
A stockout costs the contribution of the sales it prevents plus whatever ranking decay follows; excess stock costs holding and obsolescence. Both are denominated in profit, which makes the reorder point an optimization rather than a rule of thumb — and campaign-driven demand spikes make the distribution, not the mean, the thing that matters.
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.
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
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
OPEN DATASET
Thai marketplace benchmarks
Aggregated first-party measurements across Thai Shopee shops: what an ad click costs by hour of day, when buyers actually order by category, and the attribution mechanics that make those numbers read wrongly. Observational, aggregate-only, minimum 8 shops per cell. CC BY 4.0.
Read paper
04/FAQ
Frequently asked
E-commerce profit optimization is choosing the values of the decisions you control — ad target, ad budget, price, assortment, and stock — so as to maximise contribution after every real cost, subject to the constraints that bind each decision. It differs from e-commerce analytics in that analytics reports a state and optimization selects an action, and it differs from following platform recommendations in that marketplaces optimize gross merchandise value, which is the base they charge commission on, rather than your profit. In practice the hard part is not the optimization but the objective: contribution per order line has to be reconstructed from cost of goods, category commission, transaction and payment fees, seller-funded vouchers, the free-shipping cost share, fulfillment, and a returns reserve before any decision can be scored against it.
The same five decisions apply, with one Shopee-specific starting point: contribution per SKU after category commission, transaction and payment fees, vouchers, the free-shipping cost share and returns — which gives break-even ROAS as one divided by that margin. Treat break-even as the zero line rather than a floor, because a target set there permits a bid equal to the whole contribution of a click. The full Shopee playbook, with the worked example and the operator checklist, is the Increase Shopee profit pillar linked from this page.
Because they are optimal for a different objective. Marketplaces earn commission on gross merchandise value, so their ad systems, their campaign suggestions and their bid recommendations are tuned to maximise GMV — which is why the flagship Shopee campaign type is called GMV Max. Maximising GMV and maximising profit coincide only when contribution margin is constant across everything you sell, which it never is. A recommendation that lifts GMV by pushing volume into low-margin SKUs, or by lowering an ad target below break-even, is doing exactly what it was designed to do while making you poorer.
Only after the same product is one row rather than three. Shopee, Lazada and TikTok Shop each carry their own fee schedules, voucher mechanics, attribution windows and cost-share rules, so the identical unit can hold materially different contribution on each. Optimizing per platform in isolation therefore produces a portfolio that is locally sensible and globally wrong — most obviously in budget allocation, where the correct rule is to equalise marginal return across platforms rather than to hit a target on each. DataGlass builds a canonical catalogue across the three, computes contribution on each binding, and ranks moves against one objective.
Revenue and profit move together only while contribution margin is positive and roughly constant, and neither condition survives contact with a real marketplace catalogue. In the worked example on our blog, a campaign whose Target ROAS is set two points below break-even posts the largest GMV of any configuration tested — four and a half times the profit-optimal one — and loses THB 6,173 a day doing it. Every visible metric improves as the campaign walks off the cliff. That is the general shape of the problem: revenue is observable and profit is not, so optimizing what you can see is a reliable way to lose money slowly.