Most Amazon products that fail don't fail in month six. They fail in the first ninety days — before they have reviews, before they have conversion history, and before Amazon's ranking systems have any reason to trust them. The cold-start window is unforgiving: Amazon evaluates every new ASIN against the performance of established products in the same niche, and early signals (click-through rate, conversion rate, sell-through, review velocity) compound. A listing that converts at 3% in week two gets fewer impressions in week three; one that converts at 15% gets pushed harder and cheaper.
In 2026 the cold-start problem has a second layer. Amazon's AI shopping surfaces — Rufus, AI-powered search, and the recommendation modules fed by models like COSMO — now influence a meaningful share of product discovery, and they build their understanding of a new product from the listing content itself. A new ASIN with thin, generic, or contradictory content doesn't just rank worse in keyword search; it is also under-represented or mischaracterized in AI-driven answers and recommendations. Launching well in 2026 means giving both the keyword ranking system and the AI understanding system something accurate and complete to work with, from day one.
This guide lays out a phase-by-phase plan for the first 90 days: what to prepare before launch, what to verify in the first two weeks, how to earn the conversion signals that drive ranking, and how to scale without wasting budget. It's written for a single new ASIN launch on the US marketplace, but the sequence applies to any Amazon marketplace.
The most expensive mistake in Amazon launches is sending paid traffic to a listing that can't convert. Ads can buy impressions, but they cannot buy a second click from a shopper who bounced. Before inventory even lands, every element of the detail page should pass a quality bar:
| Item | Standard | Why it matters at launch |
|---|---|---|
| Title | Primary keyword in first 80 chars, no keyword stuffing | Drives CTR, the first ranking signal |
| Main image | 85%+ fill, white background, one-glance clarity | CTR competition on the search page |
| Bullets | Benefit-first; answers top competitor review complaints | Conversion + AI summary source |
| A+ content | Live before first sale | Session depth and trust |
| Backend attributes | Every relevant field filled | Filters + AI attribute reasoning |
| Search terms field | Complementary terms not already in visible copy | Covers alternate phrasings without clutter |
| Pricing plan | Launch price mapped to a reference price strategy | Avoids "was" price suppression later |
| Inventory plan | Enough stock for 60–90 days at forecast sales | Stockout in week 3 resets the flywheel |
The first two weeks are about making sure Amazon has correctly understood what your product is and which queries it should surface for. Ranking improvements in this phase come less from spend and more from verification.
Indexing is binary: either Amazon associates your ASIN with a keyword, or it doesn't. A new listing that isn't indexed for its primary keyword is invisible no matter how good the content is. Check indexing for your top 10–20 priority terms in the first days after the ASIN goes live. If a core term isn't indexed, the usual causes are content that never contains the term in any field, categorization errors, or a suppressed listing.
External traffic in week one — email lists, social followers, a landing page, a launch audience — is one of the cleanest ways to generate real, converting sessions without bidding against yourself on expensive keywords. Amazon tracks sessions and conversions regardless of where they originate; a new ASIN with a converting session history starts its ranking life from a better baseline. Keep this traffic honest: real purchases at the listed price, no incentivized orders.
If your brand is enrolled in Brand Registry, Amazon Vine remains the most reliable compliant way to get first reviews. Enroll with a modest unit count in the first week rather than after sales have already stalled. Early Vine reviews also put substantive text on the page — pros, cons, use cases — which both converts shoppers and gives AI surfaces more to quote. Budget for it: Vine enrollment fees and free units are a real cost, and a small percentage of early reviewers will be critical. That's fine; a 4.3 average with twenty reviews beats a blank review section in every measurable way.
Once indexing is verified and the listing has a trickle of real sessions, the middle month is where you buy volume efficiently. The goal of ads in a cold start is not ACOS — it is converting session volume that gives the ranking system evidence your product satisfies intent.
| Campaign | Type | Targeting | Purpose |
|---|---|---|---|
| Exact — core terms | Sponsored Products | 5–10 highest-intent exact keywords | Capture proven demand, control bids on money terms |
| Auto — discovery | Sponsored Products | Auto, all match types, modest bid | Let Amazon surface terms you didn't predict; feed search term reports |
| Broad — long tail | Sponsored Products | Broad/phrase on secondary terms | Fill the middle of the funnel, mine for exact candidates |
Three campaigns is enough. Resist the urge to build fifteen campaigns on day one — fragmented budgets starve every campaign below the learning threshold, and you get no clean read on anything. Review search term reports twice a week, promote converting search terms into the exact campaign, and negate wasting terms. Expect a high ACOS in this phase and don't panic: a 40–60% ACOS on a new ASIN that's building rank is often cheaper than relaunching a dead listing in month four.
In the cold-start window your job is to prove conversion, and price is the strongest conversion lever you hold. A launch price slightly under your target margin position — not at a loss you can't sustain, but visibly competitive against the incumbents on page one — moves conversion rate enough to change your trajectory. As reviews accumulate and rank stabilizes, step the price up in small increments and watch the conversion rate; every increment that holds conversion is permanent margin.
Strip the dashboard down to two numbers for this phase: unit session percentage (conversion rate) and click-through rate on your main keyword placements. If CTR is healthy but conversion is weak, the problem is inside the detail page — price, reviews, images, or copy. If CTR is weak, the problem is the search result impression: main image, title, price point against neighbors. Fix the weaker of the two before increasing spend; scaling traffic into a leaking funnel is how launch budgets die.
By the middle of month two you should have enough data for the feedback loop that separates structured launches from hopeful ones.
Brand Registry sellers get Search Query Performance data — impressions, clicks, cart adds, and purchases per search query. This is the single most underused launch tool. It tells you, for your ASIN, which queries you're winning and which you're merely visible on. The gap between "impression share" and "purchase share" per query is your prioritized roadmap: queries where you show up but don't convert need content or price work; queries where you convert well but barely appear need bid support or content reinforcement.
Days 46–90 are when expansion pays: additional exact keywords from your mined search term data, Sponsored Display for a share-of-aisle presence, video ads if you have assets, and — for products with genuine variant potential — expanding the family with new colors or sizes once the parent has traction. Expansion works when it rides an existing base of reviews and rank; done in week one it just multiplies the cold-start problem.
| Checkpoint | Day 30 target | Day 90 target |
|---|---|---|
| Reviews | 10–15 live, 4.0+ average | 30–50 live, 4.2+ average |
| Conversion rate (unit session %) | 8–12% in core terms | 12–18% in core terms |
| Primary keyword rank | Top 3 pages | Page 1 for 2–3 core terms |
| Ad structure | 3 campaigns, terms promoted weekly | Proven terms exact-only; waste negated |
| Inventory | Coverage to day 90 | Reorder cycle established |
| Search Query Performance | Baseline reviewed once | Gap analysis driving the roadmap |
These ranges are directional, not universal — a $12 consumable and a $90 gear item live in different conversion worlds. What matters is the trend: each checkpoint should improve against its own baseline. A listing flat in every metric across eight weeks is telling you something is structurally wrong, usually in the offer itself (price, reviews, or product-market fit) rather than in traffic.
Expect meaningful movement in two to six weeks for moderately competitive niches, and longer for high-competition core terms. The timeline is driven by conversion history and review accumulation, not by calendar alone. A listing that converts well from early sessions can hold page-one positions within a month; a listing that converts below niche average can stay buried regardless of spend.
Realistically, yes — unless you bring substantial owned traffic. Ads are the fastest compliant way to generate the converting sessions that build rank. The structure can be simple: exact terms you know matter, auto for discovery, and disciplined negation. What you should not do is run ads into an unready listing.
For most launches, yes. Vine is the cleanest compliant path to first reviews, and early review text improves both conversion and the material AI shopping surfaces can draw on. Enroll early with a small unit count, and treat critical reviews as free product feedback rather than a threat.
There's no fixed threshold, but conversion measurably improves in the 10–15 review range for most categories, and again around 30–50. The average rating and the recency of reviews matter as much as count. A new listing with 15 genuine reviews outperforms one with zero — the gap is the cold-start tax.
Category-dependent. Many launched products sit at 5–10% initially and climb to 12–18% in their core terms as reviews and content mature. Track unit session percentage weekly and compare it to the category benchmark in your category's product opportunity or niche reports rather than to a universal number.
Only if it's genuinely the same product in a new variant under a valid theme. Adding a new child to an established family gives the new variant the family's reviews and rank as a starting point — which is exactly why Amazon polices variation abuse. Adding a fundamentally different product to borrow reviews risks suppression across the family.
AI shopping surfaces summarize products from listing content — titles, bullets, A+, and attributes — and recommend based on how well that content matches shopper intent. A new listing with specific, structured, complete content gets characterized accurately and recommended for the right questions; a vague listing gets skipped in favor of competitors the AI understands better. In 2026, listing quality is not just a conversion asset, it's how the AI layer learns you exist.
Modestly and strategically, yes. A launch price visibly competitive with page-one incumbents improves conversion when you have no reviews to lean on. Coupons are a visible way to signal the deal without permanently anchoring your price low. Step pricing up as review count and conversion history accumulate.
Ready to pressure-test your listing before the traffic arrives? Run it through the free Amazon Listing Checker to catch the content gaps that quietly kill conversion — then fix them before day one, not after week six.
Check your listing before launch →