Ads
The standard ad-audit advice on Lazada is the same advice every Lazada Ads agency offers: find the campaigns with low platform ROAS and pause them. The advice is not wrong; it is incomplete in a structurally important way. Lazada’s ad inventory is split across two products with different mechanics, different conversion behaviour, and different break-even economics. Treating them as one ROAS distribution and pruning the visible underperformers is the right answer to the wrong question.
In our data across ~180 Thai SEA-6 Lazada accounts, the structurally loss-making ad spend on a typical account is concentrated not in the campaigns that look broken on the dashboard, but in Sponsored Discovery placements with strong-looking platform ROAS, healthy click-through, real attributed orders — and unit economics that turn negative the moment LazMall commission, the seller-funded portion of Voucher Wallet, the Free Shipping Logo seller-share, and the SKU's actual COGS are properly counted. Sponsored Search is more intent-driven and clears its true-ROAS bar more often. Pruning Sponsored Search before Sponsored Discovery is the audit-order error this note documents.
This note argues that reducing Lazada ad waste is a per-placement decision, not a per-campaign decision. The minimum useful resolution is keyword × SKU for Sponsored Search and audience × SKU for Sponsored Discovery, with each placement evaluated against its underlying SKU's break-even ROAS rather than the account-wide target. With chart-level visualisation of the structural difference between the two ad products, sensitivity analysis on the cost-stack inputs that move most, the two-week audit procedure that produces the per-placement decision list, and the operating signal that survives Pay Day, Mega Sale, and 11.11 campaign-window inflation.
Most Lazada ad waste is not in the campaigns that look broken. It is in the Sponsored Discovery placements the dashboard ranks as winners.
Thesis: in our data, accounts running the per-placement audit recover 20–30% of monthly Lazada ad spend in the first 30 days without losing meaningful attributed revenue. The saved budget reallocates to top-decile Sponsored Search keywords and to high-margin SKUs whose Sponsored placements were funded below their potential. The framework is mechanical; the obstacle is data plumbing (per-placement attribution from order-line data); the moat is running it weekly rather than at quarter close.
Sponsored Search vs. Sponsored Discovery — the structural difference
Lazada Sponsored Search captures intent. The buyer typed a query; the seller's bid wins an impression on the search-results page. Sponsored Discovery is broad-reach: the platform shows the ad in feeds, recommendation rails, and category pages to buyers who did not necessarily express intent for the SKU. Both report ROAS in the same way through the Sponsored Solutions seller portal. Both have very different break-even economics, and the order in which an audit prunes them determines whether the account recovers margin or loses revenue.
In our data across Thai LazMall accounts, Sponsored Search converts at 1.4–1.8× the true ROAS of Sponsored Discovery on comparable budgets. The differential is largest on long-tail and branded variants, where Sponsored Discovery's broad targeting drifts into adjacent and substitute queries that consume budget without converting to seller-margin sales.
Two operational implications follow directly. First, the right pruning order is Sponsored Discovery first — pause placements below the SKU's break-even ROAS, redistribute the saved budget to Sponsored Search top-decile keywords on top-quartile-margin SKUs. Second, broad-match keyword expansion on Sponsored Search is opt-out by default and behaves differently from Sponsored Discovery's audience expansion; treat it as a third audit lever rather than collapsing it into the Search/Discovery split.
The dashboard ROAS problem on Lazada
Lazada's Sponsored Solutions dashboard reports ROAS as ad-attributed revenue divided by ad spend, the same definition Shopee uses. The numerator is gross. Every variable cost that compounds at scale on a Lazada account sits below the line the platform draws: LazMall commission (3–10% by category, vs. 1–4% on the standard store), the ~2% payment processing fee, the seller-funded portion of any Voucher Wallet contribution applied at order close, the seller-share of the Free Shipping Logo subsidy on program-eligible orders, the SKU's COGS, packaging, and a category-specific returns reserve.
Both campaigns: ad spend THB 1,000 · attributed revenue THB 8,000 · platform ROAS 8.0
LazMall, 35% gross margin SKU:
COGS 5,200 + commission 800 + payment 160 + voucher 400 + free-ship 240 + ad 1,000 = 7,800
Net ad profit: +THB 200
True ROAS = (8,000 − 5,200 − 800 − 160 − 400 − 240) / 1,000 = 1.20
LazMall, 22% gross margin SKU:
COGS 6,240 + commission 800 + payment 160 + voucher 400 + free-ship 240 + ad 1,000 = 8,840
Net ad profit: −THB 840
True ROAS = (8,000 − 6,240 − 800 − 160 − 400 − 240) / 1,000 = 0.16
Spread: ~7.5x in true ROAS at the same dashboard signal. The dashboard cannot tell them apart.The platform metric ranks both campaigns identically; the bank account disagrees by THB 1,040. Replacing platform ROAS with true ROAS at the placement level — keyword for Sponsored Search, audience for Sponsored Discovery — produces the per-placement decision list that the campaign-level aggregate cannot.
A two-week audit, applied per placement
The right Lazada ad audit is a two-week measurement window that collects the data needed to make durable decisions, not a one-day pause list. The structure below produces a per-placement decision list with auditable reasoning, applied at production cadence rather than only at quarter close.
1. Export per-order data from Lazada Open Platform.
Required columns: order ID, SKU, ad-attributed source (Sponsored Search keyword, Sponsored Discovery audience, organic), ad-attributed revenue at order close (not click time), commission (program-aware), payment fee, Voucher Wallet split, Free Shipping Logo seller-share, fulfillment, returns flag.
2. Reconstruct true ROAS at the placement resolution.
For each Sponsored Search keyword × SKU and each Sponsored Discovery audience × SKU: contribution profit before ads = revenue − COGS − program-aware commission − payment fee − Voucher Wallet seller-funded portion − Free Shipping Logo seller-share − returns provision. Sum across the 30-day window. Divide by ad spend at the same resolution.
3. Compute break-even ROAS per SKU.
Break-even ROAS = 1 / contribution margin rate. Apply per SKU. A 35% margin SKU breaks even at ROAS 2.9; a 15% margin SKU breaks even at ROAS 6.7. The break-even bar is the audit floor.
4. Flag and tag every placement below break-even.
Tag the underlying cause: bid pressure (auction-inflated CPC), broad-match drift (Sponsored Search keyword expansion), Discovery audience drift (broad targeting), voucher-tier escalation (campaign-window seller-funded voucher above category margin), low-margin SKU, weak conversion. Each flagged placement gets one of four actions: pause, lower bid, narrow targeting, or reassign budget to a higher-margin SKU.
Sensitivity — what changes the audit’s output
The audit's precision depends on input quality and on platform-side conditions that drift week-to-week. The table below shows how the share of ad spend flagged as loss-making shifts under common scenarios.
| Scenario | Share flagged loss-making | Note |
|---|---|---|
| Clean data; per-placement attribution; LazMall commission table refreshed | 20–30% | Reference — typical research-grade audit output |
| Campaign-level aggregation only (no per-placement) | 8–14% | Misses the Sponsored Discovery audience leak entirely |
| LazMall vs. standard-store commission misapplied (flat-rate) | 14–22% | Misallocates 2–4 pp across the catalog |
| Voucher Wallet seller/platform split aggregated | 24–34% | Overestimates seller-funded share — false positives |
| Click-time attribution (not order-close) | 12–18% | Understates ad cost during the 1–7 day lag window |
| Pay Day / 11.11 campaign-window data only | 28–40% | Auction inflation peaks during campaign weeks |
| No category-specific returns reserve | 14–22% | Understates waste on returns-prone categories (apparel, beauty) |
Each row holds all other inputs at the reference baseline. The sensitivity confirms the structural argument: the audit's value depends on data-quality discipline. Bad inputs produce results indistinguishable from the standard "pause low platform ROAS" approach.
Limitations and where this argument breaks
- Account-size lower bound. The two-week per-placement audit assumes operating capacity to attribute ad spend at the keyword × SKU and audience × SKU resolution. Below ~THB 200K monthly revenue, the operational overhead exceeds recovered margin; an account-wide ROAS target plus a small per-SKU break-even table outperforms.
- Open Platform access. Smaller Lazada accounts may rely on Seller Centre CSV exports rather than Open Platform API access — fine for a one-off audit but harder to automate at production cadence.
- Attribution-window mismatch. Lazada's Sponsored ad attribution is at order close, not at click time, with a typical 1–7 day lag. The audit treats the campaign as concluded at attribution; real net-of-returns reconciliation runs ~30 days later.
- Campaign-window seasonality. The 1.4–1.8× Sponsored Search vs. Discovery true-ROAS differential is observed across rolling 30-day windows. During Pay Day, Mega Sale, and 11.11, Sponsored Discovery drift increases as the platform's relevance model surfaces ads on more loosely-related queries; re-run the audit after each major campaign window.
- Internal-data scope. The 20–30% recovery share, the audit-cohort distribution, the typical Sponsored Search vs. Discovery differential are aggregated across the SEA-6 Thai Lazada accounts we model directly. They are not population claims about all Lazada Ads accounts; they exclude operators below the size bound.
Methodology
Public-data citations are taken from the Lazada Sponsored Solutions seller portal (Sponsored Search and Sponsored Discovery documentation), the Lazada Open Platform documentation (commission, Voucher Wallet, Free Shipping Logo, ad-attribution), Alibaba's SEA segment commentary across the 2025–2026 reporting periods, and the Bain e-Conomy SEA 2025 commentary on retail-media inflation in SEA marketplaces.
Internal-data claims — the 20–30% loss-making spend share, the cohort distribution, the Sponsored Search vs. Discovery true-ROAS differential, the 30-day recovery window — are aggregated across the Thai SEA-6 Lazada accounts we model directly. The Lazada Sponsored subset comprises approximately 180 active accounts across the DataGlass research methodology sample frame (Jan 2024 – Apr 2026, 28-month observation window).
Don’t pause Sponsored Search first. Prune Sponsored Discovery to the SKU break-even bar, then tighten Sponsored Search match types, then audit campaign participation.