Ad waste is the share of advertising spend that returned no profit once marketplace fees, COGS, seller-funded vouchers, and fulfillment are deducted from the revenue it was credited with. It is what remains when a campaign is judged on true ROAS instead of platform ROAS, and it is a quantity, not a judgement: for any campaign, budget, and cost stack there is a specific number of baht that did not come back. On most marketplace ad accounts that number is a meaningful fraction of the budget, and it is concentrated rather than spread — a minority of keywords and product groups usually account for the bulk of it.
Formula
Ad waste (฿) = Σ over campaigns of max(0, spend − attributed contribution)
attributed contribution = attributed revenue × contribution rate
Waste rate = ad waste / total ad spendExample
Monthly spend ฿180,000 across four campaigns, blended contribution 22%.
campaign spend attributed rev contribution waste
A 60,000 420,000 92,400 0
B 50,000 180,000 39,600 10,400
C 40,000 95,000 20,900 19,100
D 30,000 60,000 13,200 16,800
Ad waste = ฿46,300, or 26% of spend — almost all of it in C and D.Where does ad waste actually come from?
Four structural sources, in rough order of size. Broad-match keyword expansion, which is on by default and buys queries adjacent to the ones you chose, inflating attributed revenue without inflating profit. Mixed-margin campaigns, where a single target forces thin products to be advertised at the rate a fat one can afford. Spend past the saturation point, where marginal ROAS has fallen below break-even while average ROAS still looks respectable. And non-incremental placements — brand terms and retargeting — that harvest orders which were arriving anyway. Only the first of these is visible in a keyword report; the other three require cost data the ad platform does not hold.
How do you find it without a full rebuild?
Rank campaigns by spend, take the top handful that account for most of the budget, and compute the break-even ROAS of the SKU mix each one actually sold rather than the shop average. Any campaign whose platform ROAS sits below its own break-even line is producing waste, and the gap times the spend sizes it in baht. This is deliberately crude — it uses a blended contribution rate per campaign rather than per order — but it identifies the concentration quickly, and concentration is the point. Chasing waste evenly across two hundred keywords is a poor use of a week when three campaigns hold most of it.
Is zero ad waste the goal?
No, and pursuing it destroys value. Some spend has to be exploratory: a new product, a new keyword, or a new placement has no performance history, and the only way to acquire one is to spend money that may not return. A portfolio with literally zero waste is a portfolio that stopped learning, and it will be slowly overtaken as its known-good keywords get more expensive. The useful distinction is between waste that bought information and waste that bought nothing — the second is spend on a mature, well-measured campaign that has been below break-even for weeks.
Why it matters
Ad waste is usually the largest recoverable sum in a marketplace P&L, because it requires no negotiation, no price change, and no new customers to recover. Cutting it also compounds: money withdrawn from a below-break-even campaign is available for one that is still climbing, so the same baht produces contribution instead of consuming it. Sellers who watch only platform ROAS keep funding the campaigns that look profitable on the dashboard and are net losses once full economics are applied.
In DataGlass
DataGlass computes true ROAS per campaign, ad group, and keyword, then quantifies the spend that did not clear break-even as both a percentage of budget and an absolute baht figure. Shopee campaigns below the line surface as recommendations to pause, lower the bid, or reallocate, with the SKU mix responsible shown alongside.