Amazon PPC reporting throws dozens of metrics at you, but almost every optimisation decision comes down to just three: whether people see the ad and click it (CTR), what each click costs (CPC), and whether that click turns into a sale (CVR). Diagnosing a problem starts with knowing which of the three is actually broken.
CTR — Click-Through Rate
Formula: Clicks ÷ Impressions × 100
CTR measures whether the ad is compelling enough to earn a click once it’s shown. For Sponsored Products, 0.3–0.5% is a typical average, and anything above 1% is strong.
Low CTR usually means: the main image or price isn’t competitive against what’s around it, or the ad is showing to the wrong audience entirely — a targeting problem, not a bidding problem.
CPC — Cost Per Click
Formula: Total Spend ÷ Total Clicks
Amazon runs a second-price auction: you typically pay just above the second-highest bid, not your own maximum. A high CPC on its own isn’t necessarily bad news — it usually just means real competition for that keyword. CPC only becomes a problem when read against CVR: a high CPC with a strong CVR can still be profitable; a high CPC with a weak CVR is where budget disappears fastest.
CVR — Conversion Rate
Formula: Orders ÷ Clicks × 100
CVR is the number that tells you whether the click actually turned into a sale. 10–15% is generally considered good on Amazon, though it varies a lot by category and price point.
The one rule that matters most here: no amount of bid or keyword optimisation fixes a low CVR. If the click is landing on a weak listing — thin bullet points, a poor main image, missing reviews — PPC will keep paying for traffic that a stronger listing would convert without any extra spend. Listing quality has to come before PPC spend, not after it.
Putting the three together
- Low CTR → fix the ad’s visibility or creative, not the bid
- High CPC + strong CVR → often still profitable, leave it
- High CPC + weak CVR → the expensive click isn’t the problem, what happens after the click is
- Healthy CTR + weak CVR → the listing, not the campaign, needs the work
Treating these as one blended “performance” number hides which lever actually needs pulling. Separating them is usually the fastest way to find where a campaign is actually leaking money.