Glossary
Glossary
Incremental ACOS (iACOS)
Incremental ACOS (iACOS) is ad spend divided by only the sales that would NOT have happened without the ad. It strips out cannibalised organic orders to reveal the true cost of incremental revenue.
incremental ACOSiACOSincremental ROASiROAStrue ACOS
Incremental ACOS (iACOS) measures the cost of ads in terms of only the orders that advertising actually caused. Reported ACOS counts every ad-attributed sale, including the orders a shopper would have placed anyway. iACOS removes the freeloaders.
iACOS = Ad Spend / Incremental Ad Sales
Incremental Ad Sales = Ad Sales − Cannibalised Organic Sales
A campaign with 12% reported ACOS and a 60% incrementality rate has a true iACOS of 20%. Decisions made on the 12% number will systematically over-invest in low-incrementality placements.
How to measure incrementality
Three methods, in order of rigour:
- Holdout test in Amazon Marketing Cloud. Pause the campaign in selected geographies or for selected ASINs; compare total sales against a matched control. The gap is incremental sales.
- Pause and observe. Cheaper than AMC: pause one branded campaign for two weeks, watch total branded sales. If they drop 100% of the campaign's prior ad sales, incrementality was 0% — every paid click was cannibalising organic. A 30% drop means 30% incrementality.
- GeoLift / pre-post analysis. A weaker statistical proxy when AMC is not available.
Where incrementality is usually low
- Branded keywords for established brands (20–40% incrementality typical)
- Top-of-search Sponsored Products for ASINs already ranking organic #1
- Product targeting on your own ASINs (defence with high cannibalisation risk)
Where incrementality is usually high
- Competitor product targeting
- Generic / non-branded keywords on a new ASIN
- Sponsored Display retargeting of off-Amazon audiences
Common mistakes
- Treating reported ACOS as truth. Without an incrementality estimate, you cannot tell good ads from expensive substitutes for organic.
- Cutting branded campaigns to zero. Even 20% incrementality is real — and the visible ACOS will look brutally bad after a competitor moves into the top slot.
- Running iACOS tests below statistical significance. Two-week tests with low volume produce noise, not signal.
Related terms
ACOS (Advertising Cost of Sales)
ACOS is the share of advertising-attributed revenue spent on ads — calculated as ad spend divided by ad sales. It's the core profitability metric for Amazon PPC campaigns.
TACOS (Total ACOS)
TACOS (Total ACOS) is ad spend divided by total revenue — organic plus advertised. It measures the share of the whole business that advertising consumes, not just the share of ad-attributed sales.
Cannibalisation
Cannibalisation is when paid ad spend replaces sales that would have happened organically — the advertiser pays for traffic that was already converting for free. It is the central question incremental ACOS and AMC holdout testing are designed to answer.
Branded Keyword
A branded keyword is any keyword that contains your brand name (e.g. "amalyze ppc tool", "amalyze pricing"). Branded keywords carry the highest CVR and lowest ACOS in the account; bidding on your own brand is defensive, not optional.
Statistical Significance
Statistical significance is the discipline of not making bid decisions on samples too small to be meaningful. In PPC, the practical threshold is roughly 30+ clicks or 3+ orders before any keyword-level inference is defensible.