amazon-ppc-campaign

Amazon PPC Campaign Optimization 📢

Build profitable PPC campaign structures from scratch, or audit and optimize existing campaigns with data-driven bid adjustments. No API key — works out of the box.

Installation

npx skills add nexscope-ai/Amazon-Skills --skill amazon-ppc -g

Two Modes

Capabilities

• ACoS financial framework: Calculate break-even ACoS, target ACoS, and Max CPC from product margins — the foundation for every bid decision
• Campaign architecture design: Build a structured Auto → Broad → Exact funnel with proper negative keyword isolation between campaigns
• Keyword grouping: Organize keywords into campaign buckets with match types and initial bids based on confidence level
• Bid optimization: Apply ACoS-based bid adjustment rules using industry-standard formulas (cut/increase by percentage based on ACoS range)
• Keyword funnel analysis: Identify migration opportunities (Auto→Broad→Exact) and wasted spend (high-click zero-sale terms)
• Negative keyword management: Generate seed lists (cross-campaign, irrelevant terms, generic waste modifiers) and ongoing additions from search term data
• Search term report analysis: Parse user-provided campaign data to find profitable terms, wasteful terms, and optimization gaps
• Competitor ASIN targeting: Build product targeting campaigns aimed at competitor product pages
• Integration chain: Works with amazon-keyword-research for keyword input and amazon-listing-optimization for pre-launch listing quality checks

Usage Examples

Mode A — Build New Campaigns

I'm launching a portable blender on Amazon US. Price: $39.99. Product cost: $8, shipping: $3, Amazon fees: $7.50. Here are my keywords: portable blender, personal blender, smoothie maker. Build me a PPC campaign structure.

Use amazon-keyword-research to find keywords for "bamboo cutting board", then build a PPC campaign structure. Product costs $6, sells for $29.99. Brand new product launch.

Mode B — Optimize Existing Campaigns

Short Prompts Work Too

How This Skill Collects Information

Users rarely provide everything upfront — and they don't need to. This skill follows a progressive information gathering approach:

Step 1: Extract from the prompt. Parse whatever the user already provided — ASIN, price, ACoS numbers, campaign names, keywords, etc.

Step 2: Auto-discover. If an ASIN is given, run the bundled scripts/fetch-competitor.sh <ASIN> to get price, category, BSR, and competitor context. This script handles Amazon's anti-bot protections. If the user mentions a product type without an ASIN, use web_search to understand the market.

Step 3: Identify gaps. Compare what you have against what's needed (see the Required Information tables in Mode A Step A1 and Mode B Step B1 below). Focus on what's critical to proceed:

• Mode A critical: product costs (to calculate ACoS) + monthly ad budget (to size campaigns) + keywords or competitor ASINs (to build campaigns)
• Mode B critical: current ACoS + profit margin (to know the gap and set targets)

Step 4: One consolidated follow-up. Ask only for missing critical items — in one conversational message, not a questionnaire:

Step 5: Use estimates when stuck. If the user can't provide something (e.g., doesn't know exact fees), use reasonable category-based estimates and clearly note the assumption. Never block progress waiting for perfect data.

Key Concepts

Three formulas drive every recommendation in this skill. They're introduced here and applied in Step A2 (for Mode A) and Step B2 (for Mode B).

Break-even ACoS = Profit margin before ad spend. If your product sells for $40 with $15 in costs after Amazon fees, your margin is $25/$40 = 62.5%. At 62.5% ACoS you spend all profit on ads — break even.

Target ACoS = Break-even ACoS − Desired profit margin. Want 25% profit after ads? Target ACoS = 62.5% − 25% = 37.5%.

Keyword Funnel = The core PPC optimization loop, applied in Steps A4/A6 (building) and B3 (optimizing):

Mode A Workflow — Build Campaign Structure

Step A1: Collect Product Info

The following details are needed. Many can be extracted automatically (see "How This Skill Collects Information" above) — only ask for what's truly missing.

Step A2: Calculate ACoS Targets

Using the formulas from Key Concepts, compute the financial framework that governs all bid decisions:

If user doesn't know their conversion rate, use category benchmarks: 10-15% is average.

Step A3: Collect Keywords

Keywords can come from three sources (use one or combine):

1. From amazon-keyword-research skill (recommended): Run keyword research first, then feed the ranked keyword list into this skill.
2. From competitor ASINs: User provides 1-3 competitor ASINs → run scripts/fetch-competitor.sh <ASIN> for each → extract keywords from their titles and bullet points. The script returns title, brand, bullets, price, category, BSR, and review count.
3. From user's list: User provides their own keywords (e.g., from Helium 10, search term reports, or manual research).

Additionally, expand keywords using Amazon autocomplete: curl -s "https://completion.amazon.com/api/2017/suggestions?mid=ATVPDKIKX0DER&alias=aps&prefix=<URL-ENCODED-KEYWORD>" | python3 -c "import sys,json; [print(s['value']) for s in json.load(sys.stdin).get('suggestions',[])]"

Step A4: Build Campaign Structure and Group Keywords

Default: 4 campaigns. This is the standard structure for a new product launch:

If budget is tight: Launch Priority 1+2 first (Auto + Exact). Add Priority 3 after one week of data. Add Priority 4 when you have competitor ASINs identified.

Organize keywords into these campaign buckets:

See the Mode A Output template below for the exact format of keyword groupings per campaign.

Step A5: Set Initial Bids

Max CPC (from Step A2) is your profitability ceiling — not your actual bid. Actual competitive bids depend on the category and keyword competition.

How to recommend bids:

1. Calculate Max CPC as the financial guardrail (what you can afford)
2. For actual starting bids, tell the user to check Amazon's suggested bid range when creating the campaign in Seller Central — this reflects real auction data
3. If Amazon's suggested bid > Max CPC, flag the gap and explain: either accept a loss (ranking launch), raise product price, or skip that keyword

When you don't have suggested bid data, use these category-relative starting points:

Important: These are starting points. The real optimization happens after 1-2 weeks of data — adjust based on actual ACoS per keyword.

Step A6: Build Negative Keyword Seed List

Generate an initial negative keyword list before launch. Three types:

1. Cross-campaign negatives: Add all Exact campaign keywords as negatives in Broad and Auto campaigns (prevents internal competition — this is the Keyword Funnel isolation from Key Concepts).
2. Irrelevant term negatives: Terms that share words with your product but are wrong category/intent. Example for "bamboo cutting board":
   • Wrong material: "plastic cutting board", "glass cutting board"
   • Wrong product: "cutting board oil", "cutting board stand"
   • Wrong intent: "how to clean cutting board", "cutting board DIY"
3. Generic waste negatives: Common low-intent modifiers: "free", "cheap", "used", "DIY", "review", "reddit", "how to"

Step A7: Generate Campaign Blueprint

Compile everything from Steps A1-A6 into the final deliverable. Follow the Mode A Output template in the Output Formats section below.

Mode B Workflow — Optimize Existing Campaigns

Step B1: Collect Campaign Data

The following details are needed. Follow the same progressive gathering approach — extract from the user's prompt first, then ask for missing critical items in one follow-up (see "How This Skill Collects Information" above).

Step B2: Performance Audit

Using the ACoS formulas from Key Concepts, analyze across five dimensions: (1) Financial Health — break-even vs current ACoS, monthly profit/loss; (2) Campaign Efficiency — per-campaign ACoS with 🔴🟡🟢 status; (3) Keyword Performance — group keywords by profitable/marginal/unprofitable/zero-sales; (4) Budget Allocation — is spend proportional to revenue? recommend shifts; (5) Missed Opportunities — converting terms not in Manual, high-spend zero-sale terms without negatives, underfunded winners.

Step B3: Keyword Funnel Analysis

Apply the Keyword Funnel from Key Concepts to the user's actual data. Three actions:

• Migrate up (2+ orders): Auto → Exact or Broad → Exact. Add as negative in source campaign.
• Add negatives (10+ clicks, 0 sales): Add as negative exact or phrase in the source campaign.
• Watch list (<20 clicks): Not enough data yet — flag for next review cycle.

Step B4: Bid Adjustments

Apply ACoS-based bid adjustments to keywords with 20+ clicks (minimum for statistical significance):

• ACoS > 200%: cut bid 30-50%
• ACoS 100-199%: cut bid 20%
• ACoS target+10% to 99%: cut bid 10-15%
• ACoS at target (±10%): no change
• ACoS below target: increase bid 10-20%
• 10+ clicks with 0 sales: pause keyword

Output a table: Keyword | Current Bid | Current ACoS | New Bid | Reason

Step B5: Generate Optimization Action Plan

Compile everything from Steps B1-B4 into a prioritized action plan. Follow the Mode B Output template in the Output Formats section below.

Output Formats

The primary deliverable is always an actionable campaign plan the seller can implement directly in Seller Central.

Mode A Output — New Campaign Blueprint

Mode B Output — Optimization Report

Ongoing Management & Integration

After setup, offer weekly reminders (cron/heartbeat): search term analysis + bid adjustments + monthly full audit. Recommended skill chain: amazon-keyword-research → amazon-listing-optimization → amazon-ppc. Always check listing quality before spending on ads. More skills: Amazon-Skills | eCommerce-Skills

Limitations

This skill uses publicly available data and user-provided campaign reports. It cannot access Seller Central directly, pull real-time bid landscapes, or automate campaign changes via API. For deeper PPC analytics with automated bid management, check out Nexscope — Your AI Assistant for smarter E-commerce decisions.

Built by Nexscope — research, validate, and act on e-commerce opportunities with AI.