The order of operations I use for new-product ads on Amazon
Most sellers I talk to can open campaigns, stuff keywords into a listing, and swap main images. What they don't have is a launch plan. So the ads drift. One week it's about rank, the next it's about ACoS, the next it's about scaling spend, and nothing finishes.
Here's the sequence I use when I launch a new ASIN. It's boring on purpose.
**1. Decide what the ads are actually for**
New-product ads have two jobs: push the organic rank of the keywords you care about, and pick up orders at an ACoS you can live with. Early on, the rank job matters more. Weight is low, there are no reviews, organic position is buried. Judging a cold-start campaign by pure ad profit is misreading the stage. If ads are lifting your keyword positions and feeding the associated placements, short-term losses are often fine.
**2. Answer these before you open the campaign**
- Is this a long-term hero SKU, a seasonal SKU, a test SKU, or filler?
- How long is the launch window?
- Does your first inventory run plus restock cadence actually support sustained ad spend?
- What are the week-1 targets for orders, sessions, keyword rank?
- What's the cap on daily spend, and what's the stop-loss?
- Is phase one about testing creatives and price, or purely about rank?
If any of these are fuzzy, the campaign will flip between goals and never finish any of them.
**3. Learn manual exact before you bolt on anything else**
I know auto campaigns are the default recommendation. I still don't start there. Manual exact with a tight keyword set gives you the cleanest signal: who clicks, who converts, who burns budget. Fewer campaigns, less noise, easier calls. Add auto, product targeting, SB, and SD once the core story is clear.
**4. Structure for week one**
Small ad groups. No more than three exact or long-tail keywords each. Give every group enough daily budget to run uninterrupted for roughly a week. This is the only way to tell which keyword is actually driving the results instead of splitting your budget across 30 terms and seeing nothing.
Pick keywords before you run, not after. Reverse-search competitors, pull a short seed list, and prioritize terms where the first page still has gaps. Precision beats volume.
Match types: start with exact and phrase so you control where traffic lands. After data stabilizes, layer in broad on top, but keep exact as the workhorse and phrase as the scaler. Broad only enters when you need discovery or long-tail coverage. If rank is the goal, exact. If you need safer expansion, phrase.
Auto's role in week one is a sanity check: does Amazon read your listing the way you intended? Product targeting should also go live early, aimed at direct competitor detail pages. Keep those bids high enough to actually show. A low bid on a competitor page is a wasted line item.
**5. Placement strategy**
Top of search is expensive, but it pulls organic rank faster than anything else. Product pages are where you pick up associated traffic and cross-sell your own catalog. If the budget is tight, rest-of-search holds you in position at a lower CPC so the core keywords don't fall off the map.
**6. What to actually watch in week one**
Three numbers:
- CTR. If CTR is weak and the keyword is right, the problem is main image, price, title angle, or review count. Ads can't fix a listing that doesn't compete on the search page.
- Burn rate. Start bids near the suggested midpoint. If budget empties early, CPC is climbing too fast. Cheap effective clicks beat expensive top placements in the first week.
- Organic rank on your target keywords. This is the number. If spend is happening and rank isn't moving, something upstream is wrong: keyword fit, product competitiveness, or listing quality. Fix that before adjusting bids.
**7. Tuning after it's live**
- Don't adjust daily. Run three days, then decide. Most "issues" in the first 48 hours are noise.
- No impressions, no clicks: likely bid. Raise 10-20% and watch another few days. Don't slam the bid up, a new product with a runaway CPC bleeds fast.
- Impressions but no clicks: search the keyword on the front end. If your ad is buried on page three, it's bids. If it's on page one and still ignored, it's main image or price.
- Stop-loss cleanly. Inside a single campaign, once keywords separate into winners and losers, cut the losers. Move that budget into a new campaign testing new terms.
- Shift from dashboard-watching to weekly reviews. Pull the report, pick the keywords that are consistently clicking, converting, and controllable on ACoS, and give them their own campaign, budget, and exact targeting. That's how structure gets cleaner over time.
**8. Four-week rhythm for a competitive category**
Week 1, cold start. Simple structure. Lower price plus a strong coupon (around 20% off), Vine enrolled. Point is to confirm the creative works and that a handful of keywords convert.
Week 2, scale. Ratings start showing up. Raise daily budgets. Pull any winning search terms from auto into their own manual campaigns. Increase product-targeting share to pick up more associated traffic.
Week 3, weight building. Core keywords start lifting organically. Widen into long-tail and related placements. If organic rank is climbing, you can ease off the coupon. Prune underperforming terms.
Week 4, harvest and hold. Ad-order share goes down as organic carries more weight. Concentrate spend on the keywords you've proven. Run a Deal or Lightning Deal to lock the core keywords into the first two search pages.
**9. Common ways new-product ads fail**
- Auto-only. Feels easy, ends up training the system on the wrong signals. You waste the cold-start window.
- Too many manual keywords on day one. Budget is diluted across everything, and no single term gets enough clicks to tell you anything.
- ACoS tunnel vision. If you only look at ad profit, you'll cut the ads that are quietly pushing rank, and the listing sinks.
- Impressions that don't matter. If your ad placement is on page four, those impressions aren't real opportunities. Rank won't move from ghost traffic.
- No phase plan. Running auto, manual, and product targeting with no idea which one is supposed to drive rank, which one is discovery, and which one is defense. Busy dashboard, zero direction.
- Ignoring the listing itself. This is the one that kills launches quietly. If the main image, price, reviews, or A+ can't carry traffic, ads will amplify the problem, not fix it. Ads are a magnifier. Whatever the product is, they make more of it.
**10. If you're about to launch**
Don't run every lever at once. Set the goal, build the structure, read the signals, then tune. One phase at a time.
The winners in new-product advertising aren't the ones spending hardest. They're the ones who can name, out loud, what each campaign is for this week.
Happy to answer questions in the comments. If you're launching something right now and want into the small Discord where a few of us compare numbers, drop a comment and I'll send the invite.