Ad Spend

Ad spend is the actual money Amazon charged you for ad clicks in a given period. It is the realised cost — distinct from the budget, which is the cap you set, and distinct from the CPC, which is the per-click price.

Ad spend appears as the Spend column in every Amazon Advertising report. It rolls up cleanly:

Ad Spend = Σ (Clicks × Avg CPC) across all targets in the period

It is also the numerator of every efficiency metric:

ACOS  = Ad Spend / Ad Sales
TACOS = Ad Spend / Total Sales
ROAS  = Ad Sales / Ad Spend

So while it sounds like the most basic number in the account, it is also the number through which every optimisation decision is filtered.

Gross spend vs. net spend

Amazon's reported ad spend is gross — the money charged for clicks. There is no "net of refunds" view because clicks aren't refundable; you pay for the click regardless of whether the resulting order was later returned or cancelled.

For unit-economics analysis, the relevant figure is gross ad spend ÷ net ad sales (after returns), not gross ÷ gross. Most accounts under-report true ACOS by 5–15% by using gross ad sales in the denominator while real revenue is post-returns.

Spend pacing

A campaign's spend pace within a day matters as much as its total spend:

• Even pacing — spend distributed across the day proportional to traffic. Amazon's default behaviour when budget is comfortable.
• Front-loaded pacing — spend concentrated in the morning when CVR is often highest. What happens when the budget is tight and the campaign caps out before noon.
• Dayparted pacing — spend concentrated in specific hours by operator intent. Requires automation tooling against Amazon Marketing Stream.

A campaign that consistently spends 80% of its budget by 2pm is telling you the budget is the binding constraint, not the bid. Lift the budget and observe the unconstrained spend curve before deciding on bid changes.

Ad spend velocity

The rate at which a campaign spends is the leading indicator of any account-level issue:

• Spend velocity spike at flat impression count = CPCs are rising (auction pressure or modifier creep).
• Spend velocity spike at rising impressions = demand surge or auction expansion (healthy if CVR holds).
• Spend velocity collapse at flat budget = impression share dropping (often a CVR or relevance problem).

Spend velocity should be a daily-checked metric. Most accounts only check it weekly in the rollup and miss the early warning by 4–6 days.

Ad spend vs. ad cost

A small vocabulary trap: some platforms use "ad cost" to mean the same thing as ad spend; others use "cost" to include platform fees, tooling, agency retainers. When comparing Amazon ad spend to "total advertising cost" in a CFO conversation, name the components explicitly.

Total advertising cost (the CFO view) = Amazon ad spend + DSP spend (if applicable) + platform fees + tooling subscriptions + agency retainer + internal headcount allocation.

Spend allocation: the 70/20/10 default

A defensible default for a mature account:

• 70% of ad spend to the profit layer (harvested exact, branded exact, refined PAT).
• 20% to discovery (auto, broad, category PAT without refinements).
• 10% to upper-funnel (Sponsored Brands, Sponsored Display reach, conquest).

Accounts heavily skewed to discovery (>40%) typically have weak harvest discipline. Accounts heavily skewed to profit (>85%) typically have under-fed discovery and are stagnating on the same keyword set.

Common mistakes

• Using gross ad sales in the ACOS denominator. Under-reports true ACOS; bid models compound the error.
• Ignoring spend velocity until the weekly rollup. Early-warning signal arrives 4–6 days late.
• Confusing budget and spend. Budget is the cap; spend is reality. A €50 budget that consistently spends €23 has 54% unused capacity — that's a bid problem, not a budget problem.
• No spend-allocation target. Without an explicit 70/20/10 (or similar) discipline, allocation drifts toward whichever campaigns are loudest in the dashboard.