How should we launch a new product with Amazon ads?
Our answer
Ready the listing, stock and Vine, then launch Sponsored Products in layers with bids capped by margin. Expect ads to cost about 25 to 40% of Amazon sales for 1 to 3 months. Set a ceiling and review date.
- Confidence
- Medium
- Strongest proof
- 3 Amazon rules
- 6 practitioners, 3 opinions or claims
- Sources
- 7 sources
- 12 statements
- ACM
SCNAL
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentReadiness comes before traffic. Amazon's launch guide asks you to review the detail page, stock and price before launching. A stockout in week 5 erases the velocity bought in weeks 1 to 4 (Canopy Management, Amazon marketplace agency).
Start with an automatic campaign, then add manual targeting built from what converts (Amazon, Canopy Management). Bids are where advice differs. Amazon starts at the suggested bid, Canopy Management bids above it at first and Setting bids and targets on Amazon starts on bids down only.
Launches cost more in ads. Eightx and Sequence Commerce put launch ad cost at 25 to 40% of total Amazon sales in the first 1 to 3 months. Canopy Management expects the first 30 to 45 days to run at a loss.
Our judgment
- Treat readiness as the first launch step. Every click in week 1 lands on the listing as it stands then, and a stockout erases the velocity paid for.
- Start bids near Amazon's suggested bid, under a margin cap, with down only bidding (Setting bids and targets on Amazon). Raise them toward that cap where impressions are too thin to learn. We weigh Amazon's advice to start at the suggested bid, Canopy Management's bids above it so campaigns win enough auctions to learn, and the down only start that Setting bids and targets on Amazon sets.
- Plan launch ad cost at 25 to 40% of Amazon sales for the first 1 to 3 months, the Eightx and Sequence Commerce range, with a fixed ceiling and review date, and expect it to fall by weeks 6 to 10. Canopy Management's 15 to 20% sits below it.
- Don't plan around a honeymoon period. LandingCube asserts one without evidence, and Canopy Management says Amazon has never confirmed it.
Our judgment, built from Canopy Management, Amazon, Eightx, Sequence Commerce and LandingCube: our reading of these sources, not a finding from any one of them.
What the sources say
12 statements from 7 sourcesAmazon's launch guide says to check product pages, stock and price before ads, run automatic and manual Sponsored Products together starting at suggested bids, then add Sponsored Brands for awareness.
“Using both automatic and manual targeting creates a balanced advertising strategy.”
Amazon suggests starting Sponsored Products with an automatic campaign for about two weeks before adding manual ones, and testing match types for 1 to 2 weeks before pausing keywords.
“We recommend that you let your automatic campaign run for about two weeks before creating a manual campaign.”
Amazon recommends enrolling in Vine with stock in its warehouses at least three weeks before launch. Existing products qualify with under 30 reviews, and you pay nothing if no review arrives in 90 days.
“For best results, we recommend enrolling in Vine and having your inventory available in our fulfillment network at least three weeks before your planned launch date.”
Amazon says brand owners using Sponsored Brands see 23% higher sales after 13 weeks, from its own internal data comparing users with non-users, not a test.
“In fact, brand owners who use Sponsored Brands see 23% higher sales after 13 weeks.”
Nova Analytics, which sells Amazon analytics, says to find organic units by subtracting ad-attributed units from total units per ASIN. New products start at 80-90% ad units; it targets 40% by month 6.
“What matters is the direction of travel: the paid share should fall quarter after quarter as reviews accumulate and organic rank builds.”
Canopy Management, an Amazon agency, launches with auto, tight broad-match (5 to 12 keywords) and competitor campaigns, bids above Amazon's suggested bid at first, and budgets 1.3 to 1.5 times steady-state spend.
“Bid above Amazon’s suggested bid at first, because low bids don’t win enough auctions to produce usable data.”
Canopy Management, an Amazon agency, says to expect launch ACoS above target and 30 to 45 days at a loss, so set a spending ceiling, avoid stockouts and send outside traffic to the listing.
“The most expensive launch mistake is going live half-finished and paying for ad traffic to a listing that can’t convert it.”
Canopy Management, an Amazon agency, says its established products run at 8 to 12% TACoS and launches at 15 to 20%. Rising TACoS with flat ROAS signals a listing or review problem. No account count.
“TACoS rising while ROAS stays flat means you’re becoming more dependent on ads to maintain sales volume.”
Darkroom, an Amazon agency, describes one client launch (DedCool, 90 products) run with video and prospecting first. It targets a new-to-brand order rate above 60% on awareness. No results are given.
“Judging all three against one blended ACoS is the most common allocation error we see in audits.”
Eightx, compiling agency benchmarks, calls ad spend of 10-15% of total Amazon revenue (TACOS) healthy, 25 to 40% normal at launch, and says break-even TACOS equals your margin before ads.
“Break-even TACOS equals your pre-ad gross margin.”
Sequence Commerce, an Amazon agency, benchmarks by stage: TACoS of 25% to 40% at launch, 12% to 20% in growth and 5% to 12% when mature, and says low ACoS at launch usually means underinvestment.
“A low ACoS during launch usually signals underinvestment, not profitability.”
LandingCube, which sells Amazon launch tools, claims new listings get a honeymoon period when ranking is easiest and urges strong PPC from day one. It gives no evidence, and Amazon has never confirmed one.
“You should consider a stronger PPC push from day one.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer