Amazon PPC

Amazon PPC is the umbrella term for Amazon's pay-per-click advertising ecosystem. Advertisers bid on keywords or product targets, ads appear in search results and on detail pages, and the advertiser pays Amazon only when a shopper clicks — not when an ad is shown. It's a second-price auction (with some modifications), settled per impression, billed per click, and reported through both the Amazon Ads Console and the Ads API.

PPC is the lever that most directly translates budget into rank and sales velocity on Amazon. It is also the lever that most often eats margin invisibly when it isn't structured properly.

The three core ad types

• Sponsored Products (SP) — keyword- or product-targeted ads that look almost identical to organic search results. The same image, title, price, and Prime badge, with a small "Sponsored" label. SP is the workhorse of Amazon PPC and typically absorbs 70–90% of total spend.
• Sponsored Brands (SB) — banner-style ads with a brand logo and multiple products, shown at the top of search results, in the middle of results, and in Stores. SB also includes Sponsored Brands Video (in-search auto-play video ads) which has very different CPC and CVR economics.
• Sponsored Display (SD) — programmatic ads targeted by audience (lookalike, in-market, viewed-but-didn't-buy) or by product (specific ASINs, categories, or competitor PDPs). SD runs both on and off Amazon and includes a view-through attribution model that inflates reported conversions vs. SP.

Beyond the three "Sponsored" formats sit Amazon DSP (programmatic display and video off-Amazon), Amazon Marketing Cloud (AMC) for cross-format measurement, and Amazon Marketing Stream (AMS) for near-real-time data — all distinct from but adjacent to PPC.

Targeting modes for Sponsored Products

• Automatic targeting — Amazon picks keywords and ASINs based on the product listing. Inside an auto campaign, four targeting groups exist: close match, loose match, substitutes, and complements. Each can be bid separately. Autos are excellent for keyword discovery and should always be running on every active ASIN.
• Manual targeting — keywords — three match types:
  • Broad — matches the keyword in any order plus close variants, related searches, and (in 2024+) "Broad Match Plus" expansion. Highest reach, lowest precision.
  • Phrase — matches the keyword in the given order with words before or after.
  • Exact — matches the exact phrase plus minor variants (plural/singular).
• Manual targeting — products — ASIN-level targeting (your own listing for halo, competitor listings for conquest) and category-level targeting (with optional brand, price, and review filters).

A mature account runs all of these in parallel: autos and broad campaigns harvest new search terms, harvested terms are promoted into structured exact campaigns with tight bid control, and product targeting both defends owned PDPs and attacks weak competitor PDPs.

How the Amazon ad auction actually works

The textbook model says Amazon runs a second-price auction: the highest bidder wins, but pays one cent above the second-highest bid. Reality is more complicated:

• Ad Rank = Bid × Relevance Score. Relevance is a function of expected CTR, expected CVR, and historical performance of the ad and the landing ASIN. A more relevant ad with a lower bid can beat a less relevant ad with a higher bid.
• Placement modifiers apply on top: a campaign can bid +50% for top-of-search placement and +25% for product pages, which changes the effective bid by surface.
• Bidding strategies modify behavior at click time:
  • Fixed bids — your bid is your bid.
  • Dynamic – down only — Amazon lowers the bid when conversion is unlikely.
  • Dynamic – up and down — Amazon raises bids up to 100% for top-of-search and 50% for other placements when conversion looks likely.

The right strategy is rarely "always dynamic up and down." It depends on whether the goal is harvest, scale, or efficiency.

Why PPC is inseparable from organic ranking

The single most important truth about Amazon PPC is that it is not an isolated channel. Paid clicks and conversions feed back into the organic ranking signal: a product that ranks page 5 organically can be lifted to page 1 by sustained PPC traffic, and conversely, a stalled PPC campaign often signals a deeper listing, pricing, or review problem.

This is why brand-new ASIN launches lean heavily into PPC for 4–8 weeks even at unsustainable ACOS — the spend is buying rank, not just sales. It's also why suddenly cutting a "high-ACOS" campaign on a mature ASIN can crater organic sales a week later. ACOS at the campaign level is a poor proxy for the campaign's true contribution; TACOS at the brand level is the metric that matters.

See ACOS and Target ACOS for the profitability framework that bounds PPC bidding.

Campaign architecture: structure before tactics

The biggest determinant of PPC success is not bid management — it's how the campaigns are structured. A defensible structure typically separates:

• By funnel stage — autos and broad campaigns at the top (harvest), phrase in the middle (validate), exact at the bottom (optimize at known-converting terms).
• By target intent — branded (own brand terms), category (generic terms), competitor (conquest of specific brands).
• By Ad type — SP and SB on the same keyword set should live in separate campaigns so their performance can be read independently.
• By margin tier — high-margin SKUs warrant higher target ACOS and higher bids than low-margin ones; mixing them in one campaign averages away the difference.

Structure determines what data you can read; data determines what you can optimize; what you can optimize determines what ACOS you can hit.

What changes with Amazon Marketing Stream

Historical PPC operated on a 24-hour reporting cadence — yesterday's data, available this morning, acted on by tomorrow. Amazon Marketing Stream collapses that loop to roughly one hour. With hourly data, advertisers can:

• Bid by hour of day instead of by 24h average.
• Pace budgets to avoid early-afternoon sellouts.
• Detect anomalies (a search-term suddenly stops converting) within 1–2 hours instead of 24+.
• Read placement performance at hourly granularity to separate top-of-search performance from rest-of-search.

AMS is the data plumbing that lets disciplined PPC actually run as a 24/7 trading operation rather than a once-a-day cleanup exercise.

Common PPC mistakes

• Treating SP, SB, and SD as one budget. They have different funnels, different attribution, and different roles.
• Optimizing the campaign average instead of the search-term distribution. The average always hides the work.
• Letting auto campaigns run with no negative keyword discipline. Autos will happily spend on terms that will never convert.
• Cutting "expensive" campaigns without checking TACOS. Often the campaign was carrying the organic line.
• Bidding the same on broad and exact. Broad needs much lower bids than exact because broad targets are by definition lower-precision.
• Ignoring placement modifiers. Top-of-search converts dramatically better than rest-of-search; bidding flat across placements leaves CVR on the table.