Wasted Spend

Wasted spend is advertising spend that produced no attributed orders within the campaign's attribution window. It is the single largest controllable inefficiency in most Amazon PPC accounts and the primary target of the weekly negative-keyword cycle.

The working definition for diagnostic purposes:

Wasted spend = Σ Spend on search terms with ≥15 clicks and 0 orders in 30 days

The 15-click threshold is a rough statistical-significance proxy — fewer clicks could be noise, but at 15+ clicks with zero conversion, the search term is genuinely converting at <2% (roughly 1/50 odds of seeing zero orders by chance at a 2% CVR).

Where wasted spend comes from

Three primary sources, in order of typical magnitude:

1. Untracked search terms in broad / auto campaigns. Broad and auto match a long tail of search queries; without a weekly STR review, the unprofitable tail accumulates indefinitely.
2. Decaying keyword performance. A keyword that converted at 8% six months ago may convert at 2% today (new competitors, new SERP layout, listing degradation). The bid is still set for 8% CVR; the spend doesn't generate orders.
3. Stock-outs and Buy Box losses. When the SKU goes out of stock or loses the Buy Box, ads continue to spend until the campaign is paused — producing clicks that cannot convert.

The weekly cleanup workflow

Every Monday, against the trailing 7–30-day Search Term Report:

1. Filter to: clicks ≥ 15, orders = 0, spend ≥ €5.
2. For each row, decide: add as negative exact? Or investigate (is this a new competitor on the term, a listing issue, etc.)?
3. Apply negatives in bulk via Bulk Operations.
4. Total the wasted spend for the period; track week-over-week.

A mature account running this workflow consistently sees wasted spend drift toward 5–10% of total spend. An account skipping it routinely runs at 25–40% wasted spend.

Wasted spend vs. discovery cost

A subtle distinction: not all non-converting spend is wasted. Auto and broad campaigns are designed to spend on terms that don't convert, in exchange for surfacing the ones that do. That spend is discovery cost, not waste.

The way to separate them: wasted spend is spend on terms with sufficient data to prove non-conversion (≥15 clicks, 0 orders). Discovery cost is spend on terms still in the statistical-noise zone (<10 clicks). One is actionable; the other is the price of generating the next harvest cycle.

Wasted spend and ACOS

Wasted spend is the largest single lever for moving ACOS in either direction:

• A campaign with €1,000 spend, €4,000 attributed sales (25% ACOS), and €200 of wasted spend is really running at €800 productive spend and €4,000 sales = 20% ACOS once cleaned up. The 5-percentage-point improvement is sitting in the STR waiting for a negative.
• Conversely, untreated wasted spend grows monotonically. A campaign with stable bids and stable CVR can drift from 25% to 35% ACOS over 6 months purely through unaddressed waste accumulation.

Common mistakes

• No weekly cleanup cadence. Wasted spend compounds invisibly.
• Defining waste as "0 orders" without a click threshold. Single-click rows with 0 orders are noise; they don't justify a negative.
• Negating too aggressively. Negative Phrase on common substrings blocks legitimate matches. Always preview the impact before committing a phrase negative.
• Ignoring stock-out waste. A campaign that's been spending for 3 days while the SKU was out of stock is 100% waste. Always pause campaigns when SKUs go out of stock, even temporarily.
• Treating discovery cost as waste. Aggressive negation of broad/auto eliminates the harvest engine. The discovery layer is supposed to spend on non-converting terms — that's how you find the converting ones.