Most keyword research stops at 5-10 queries. That gets you 200-400 keywords. The full protocol - 21 queries across 5 categories - consistently produces 1,300-2,500 viable keywords per product line.
We've tested this across 85+ e-commerce accounts onboarded between 2023 and 2025. The difference between 5 queries and 21 isn't incremental. It's the difference between targeting 15-20% of your addressable market and seeing the whole picture.
Here's every query, what it does, what to look for, and how to run it.
Tools needed: Any keyword research platform (DataForSEO, Ahrefs, Semrush, Google Keyword Planner). The specific tool doesn't matter. The process does.
Time estimate: 4-6 hours for the full protocol. Most of that is Queries 14 (Alphabet Explosion) and 19-21 (Shopping Queries).
Purpose: Build your baseline keyword universe. Expected Output: 300-500 keywords.
What to do: List your seed terms - brand name, every product TYPE you sell (not product names - types), your industry category, and 3-5 variations of how customers describe what you sell. Pull search volume, CPC, and competition data for every seed.
What to look for:
Your seeds:
Status: [ ] Complete | Keywords found: ____
What to do: Take your top 10 seeds by volume. Pull semantically related keywords for each.
What to look for:
Status: [ ] Complete | Keywords found: ____
What to do: Same top 10 seeds. Pull long-tail suggestions - the "autocomplete" style queries.
What to look for:
Status: [ ] Complete | Keywords found: ____
What to do: Feed ALL seeds into an ideas generator simultaneously. Don't run them individually.
What to look for: Cross-pollination between product types. Terms that bridge multiple seed concepts. Keywords that wouldn't surface from any single seed alone.
Status: [ ] Complete | Keywords found: ____
What to do: Take every keyword from Queries 1-4 with volume above 500. Pull detailed data: CPC, competition level, SERP features, trend direction.
What to look for:
Status: [ ] Complete | Keywords enriched: ____
Total unique keywords after Category 1: ____
Target: 300-500. If you're under 200, your seeds are too narrow. Go back to Query 1 and add product type variations.
Purpose: Expand with competitive intelligence and advanced metrics. Expected Output: 200-400 additional keywords.
What to do: Take your top 20 commercial keywords. Identify which domains rank for them organically.
Why this matters: Your perceived competitors and your SERP competitors are often different. The domains ranking for your top 20 keywords define who you're actually fighting for visibility.
Your top 5 SERP competitors:
Status: [ ] Complete
What to do: Score every keyword collected so far for difficulty.
Action framework:
| Difficulty | Action |
|---|---|
| 90+ | Flag for review - only keep if volume justifies the fight |
| 60-89 | Keep - standard competition |
| 40-59 | Good balance of competition and opportunity |
| Under 40 | Flag as easy win |
Target portfolio mix: 30-40% easy wins, 40-50% medium difficulty, 10-20% difficult terms.
Status: [ ] Complete | Easy wins flagged: ____
What to do: Pull 12-month trend data for your top 100 keywords.
What to look for:
Seasonal angles found:
Status: [ ] Complete
What to do: Run organic SERP analysis on your top 50 commercial keywords.
What to look for: Shopping carousel presence (shopping viability), paid ad density (commercial intent confirmation), People Also Ask questions (content opportunities for Query 13).
Status: [ ] Complete
What to do: Pull keyword rankings for your top 5 SERP competitors (from Query 6). Limit to 200 keywords per domain.
What to look for: Entire keyword categories your competitors target that you don't. This is typically the biggest source of missed opportunities.
New categories discovered:
Status: [ ] Complete | New keywords from competitors: ____
What to do: Compare your domain authority against your top 5 competitors.
Your DA: ____ | Competitor range: ____ to ____
Implication: If competitors have 3x+ your authority, prioritize longer-tail keywords (Layers 2, 6, 7) where you can compete.
Status: [ ] Complete
What to do: Filter all keyword suggestions for low competition only.
What makes a "gem": Volume 100-500, CPC under $1.00, competition LOW, relevance score 70+.
Why they matter: A pocket of 20-30 gems with combined volume of 5,000-8,000 monthly searches at $0.40 average CPC generates traffic at a fraction of what competitors pay. This is where brands with smaller budgets ($5K-15K/month) compete effectively against larger advertisers.
Gems found: ____
Status: [ ] Complete
Total unique keywords after Category 2: ____
Target: 500-900 cumulative. If you're under 400, run Query 10 against more competitor domains.
Purpose: Extract question-based keywords from People Also Ask. Expected Output: 100-250 question keywords.
What to do: Run SERP analysis with PAA extraction (depth: 2 levels) on your top 30 informational keywords.
Level 1 PAA: Direct questions - "How long does [product] last?" "Is [product] worth the money?"
Level 2 PAA: Deeper questions from clicking into Level 1 answers. This is where the real insights hide.
What to look for:
Key objections discovered:
Status: [ ] Complete | Questions found: ____
Purpose: Systematic expansion across angles most brands never discover. Expected Output: 275-600 additional keywords.
What to do: Take your top 5 product types. Append every letter of the alphabet. "[Product type] a" through "[Product type] z." That's 130 total sub-queries (26 x 5 product types).
Yes, this is tedious. It's also where some of the highest-converting long-tail keywords hide. We've found entire market verticals from alphabet explosions that didn't surface in any other query type.
Your 5 product types for alphabet explosion:
New verticals/angles discovered:
Status: [ ] Complete | Keywords found: ____
What to do: Filter specifically for low-competition keywords across all product types.
Target: Build a portfolio of 50-100 low-competition keywords. These convert at 2-3x the rate of high-competition terms.
Status: [ ] Complete | Low-comp keywords: ____
What to do: Generate three types of comparison queries:
Why they matter: People actively comparing are close to buying. Most brands don't target these at all.
Status: [ ] Complete | Comparison keywords: ____
What to do: Generate queries in these formats:
These are TOFU and MOFU keywords. Lower CPC, earlier journey stage. Critical for scaling beyond bottom-funnel.
Status: [ ] Complete | Problem-solution keywords: ____
What to do: Generate "[product] for [demographic]" queries. Think age, gender, experience level, profession, lifestyle, condition.
Examples: "running shoes for flat feet," "protein powder for beginners," "standing desk for home office"
Persona segments discovered:
Status: [ ] Complete | Persona keywords: ____
Total unique keywords after Category 4: ____
Target: 900-1,800 cumulative. If you're under 700, revisit Queries 14 and 18 with more product type variations.
Purpose: Shopping-specific intelligence. Expected Output: 150-300 shopping keywords.
What to do: Pull keywords your own website currently ranks for organically. Look for terms you rank for but aren't advertising on.
Why this matters: If you rank organically, you already have page relevance signals. Adding paid gives you dual visibility (organic + paid), typically increasing combined CTR by 25-35%. Quality Score starts higher due to landing page relevance - meaning cheaper clicks.
Organic-only keywords to add to paid: ____
Status: [ ] Complete
What to do: Combine these 10 high-intent modifiers with your top 20 product types:
| Modifier | Intent Signal |
|---|---|
| buy [product] | Direct purchase |
| best [product] | Comparison shopping |
| [product] deals | Price-sensitive buyer |
| [product] free shipping | Conversion-ready |
| [product] near me | Local intent |
| [product] reviews | Evaluation stage |
| cheap [product] | Budget buyer |
| [product] sale | Promotion seeker |
| [product] online | E-commerce intent |
| premium [product] | Quality buyer |
This is 10 modifiers shown here. The full protocol uses 40 modifiers. For each modifier x product type combination, check if search volume exists. Drop combinations with zero volume.
Status: [ ] Complete | Shopping keywords found: ____
What to do: Compare your domain against each top competitor domain. Pull keywords they rank for that you don't.
This is the most consistently productive query in the protocol. In 85+ accounts, Domain Intersection has uncovered 15-25% of the final keyword database - keywords that would never surface from seed-based research.
Process:
Gap keywords found: ____
Status: [ ] Complete
After all 21 queries:
Final keyword count: ____
Target: 1,300-2,500 unique keywords. If you're under 1,000, review which queries you may have run too narrowly.
Use this to assign each keyword after classification:
| Layer | Name | Description | Budget % |
|---|---|---|---|
| 1 | Product-Defined | Core product terms everyone bids on | 30-40% |
| 2 | Attribute-Specific | Color, size, material, use-case modifiers | 15-20% |
| 3 | Conquest | Competitor brand captures and comparisons | 10-15% |
| 4 | Temporal | Seasonal, trending, event-driven | Varies |
| 5 | Question | Informational intent for retargeting pools | 5-10% |
| 6 | Persona | Demographic and use-case segments | 10-15% |
| 7 | Shopping Micro-Intent | High-intent purchase modifiers | 10-15% |
This template gets you from 150 keywords to 1,300-2,500. That's the research phase.
The next step is turning those keywords into a campaign architecture - mapping each keyword to specific campaigns with funnel-appropriate budgets, bid modifiers, landing page assignments, and a staged deployment plan.
The Google Ads AI Agentic System ($4,997+) covers the full system: all 21 queries with worked examples across multiple product categories, the complete 7-Layer Angle System with campaign architecture templates, budget allocation frameworks by layer, landing page mapping protocols, and a staged deployment plan that prevents Google from spreading budget too thin.
The template finds your market. The Agentic System builds the campaigns.