Google Ads Keyword Research: The Process That Finds Profitable Terms

Most keyword research happens once at campaign launch and never again. The result: accounts running on the same keywords for 18 months while user behavior, search volume, and competitor positioning shift constantly.

Profitable keyword research is ongoing. Here's the process.

The three keyword buckets

For e-commerce, every keyword falls into one of three categories:

Buyer-intent keywords

The user has decided what to buy and is looking for where to buy.

Examples:

• "men's leather chelsea boots"
• "stainless steel water bottle 32oz"
• "buy [specific product]"

These convert well, are competitive, and are the bread-and-butter of e-commerce search.

Comparison/research keywords

The user is researching options before buying.

Examples:

• "best running shoes for flat feet"
• "Allbirds vs Vessi"
• "[product category] comparison"

These convert at lower rates than buyer-intent but higher than informational. Often great targets for retargeting follow-up.

Informational keywords

The user is gathering information, not yet shopping.

Examples:

• "how to clean leather boots"
• "what is single origin coffee"
• "running shoe sizing guide"

Lowest direct conversion rate. Best for content marketing and top-of-funnel paid media — not your primary search budget.

For most accounts, 70% of paid search budget should go to buyer-intent, 25% to comparison, 5% to informational.

Where to find keywords

1. Google Keyword Planner (free)

Inside Google Ads → Tools → Keyword Planner. Lets you:

• Discover keywords from a seed term or URL.
• Get search volume estimates.
• See competition and bid range estimates.

Free and useful for initial research. Volume numbers are bucketed (10-100, 100-1K, etc.) — directionally accurate, not precise.

2. Google Search Console

Your existing organic traffic data. Shows:

• Queries you already rank for (organically).
• Click-through rates by query.
• Position in results.

Queries with high impressions but low position are paid-ads candidates — capture clicks while organic rank improves.

3. Search term reports in Google Ads

The actual queries triggering your existing ads. Categorize:

• Profitable queries → consider as exact-match keywords.
• Unprofitable queries → add as negatives.
• Surprising queries → investigate (maybe a new ad group opportunity).

The most underused free keyword research source.

4. Competitor analysis tools

• SEMrush, Ahrefs, Spyfu. Show what keywords competitors rank for and bid on.
• Auction Insights in Google Ads. Shows competitors in your auctions.

Use to find gaps where competitors are bidding profitably and you're not.

5. AI tools

• ChatGPT, Claude, etc. Generate keyword variations, related terms.
• Use cautiously. AI-generated keyword lists often include irrelevant terms. Validate volume in Keyword Planner.

Keyword research process

A systematic 4-hour audit:

Hour 1: Seed list

Brainstorm 20-30 seed keywords:

• Your products.
• Categories.
• Customer language ("running shoes" vs "trainers" vs "joggers").
• Use cases.
• Competitor brands.

Hour 2: Expand

In Keyword Planner, enter each seed. Pull suggestions. Filter by:

• Search volume above your minimum threshold (varies by category, often 100+/month).
• Bid estimate within your acceptable CPC range.

Build a master list of 200-500 keyword candidates.

Hour 3: Categorize

Sort by intent (buyer, comparison, informational) and theme (product type, audience, use case).

Group thematically — these become your ad groups.

Hour 4: Prioritize

Score each by:

• Search volume.
• Estimated CPC vs your conversion economics.
• Match to your product portfolio.
• Competition (Auction Insights).

Top 50-100 become your initial keyword set.

Long-tail vs head terms

Head terms

High-volume, short, generic.

• "running shoes"
• "coffee beans"
• "office chair"

High volume but high competition and high CPCs. Often unprofitable for direct response unless you're a major brand.

Long-tail

Lower-volume, specific, niche.

• "running shoes for plantar fasciitis women"
• "single origin Ethiopian coffee beans medium roast"
• "ergonomic office chair under $300"

Lower volume but higher conversion rates and lower CPCs. Often the most profitable terms.

For most accounts: focus 60-80% of budget on long-tail and mid-tail. Head terms only when you have the brand strength and margin to compete.

Match types and modern best practice

Today's match types:

• Exact match: brackets [running shoes]. Captures the exact term and very close variants.
• Phrase match: quotes "running shoes". Captures queries containing the phrase in order.
• Broad match: no symbols. Captures wider variations including synonyms and related concepts.

In 2025, with Smart Bidding, broad match has become more usable — Google's algorithm filters out irrelevant queries better than it used to.

Best practice:

• Phrase match as your default for most keywords.
• Exact match for top-volume terms where you want precision.
• Broad match with Smart Bidding for expansion (paired with strict negatives).

Monitor search term reports closely with broad match — it can pull irrelevant traffic without aggressive negative management.

Keyword volume thresholds

Don't add every keyword to your account. Quality over quantity.

Minimum volume thresholds:

• High-stakes accounts: 100+ monthly searches.
• Smaller accounts: 50+ monthly searches.
• Niche accounts: 20+ monthly searches.

Below these levels, keywords don't accumulate enough data for Smart Bidding to optimize.

Refresh cadence

Keyword research isn't one-and-done:

• Monthly: review search term reports, add high-converting queries as keywords, add bad queries as negatives.
• Quarterly: re-audit using fresh Keyword Planner data. New trends, new seasonal terms, new competitor positioning.
• Annually: strategic re-think. Has the category shifted? Are new product categories worth adding?

Common keyword research mistakes

• Keyword stuffing. 500+ keywords per ad group. Smart Bidding can't optimize across that many.
• Ignoring negative keyword research. As important as positive keyword research.
• Copying competitor keyword lists. Their economics differ from yours.
• Over-relying on Keyword Planner volume. Bucketed estimates miss long-tail opportunities.
• Ignoring user intent. A high-volume term with wrong intent burns budget.
• No quarterly refresh. Keyword sets go stale.

Keyword research for Performance Max

PMax doesn't accept keywords directly. But:

• Account-level negative keywords apply.
• Audience signals can include "search themes" (a beta in some regions).
• Custom audiences from search behavior tell PMax what queries map to your ideal customers.

For PMax: the equivalent of keyword research is shaping audience signals and providing high-quality custom audiences. The intent of keyword research still applies; the implementation is different.

A 90-day keyword strategy cycle

• Month 1: Initial research (4-hour deep dive). Launch with top 50-100 keywords.
• Month 2: Search term review. Add winners as new keywords; add losers as negatives. Target 20% improvement in wasted spend.
• Month 3: Strategic review. Are new categories worth testing? Are some keywords being eclipsed by changes in user behavior?

Repeat cycle. Keyword sets should grow and refine continuously.

What "good" looks like

A well-managed keyword strategy:

• 100-300 active keywords for a typical account.
• Monthly search term review with disciplined negatives addition.
• Quarterly fresh research.
• Mix of buyer-intent, comparison, and (sparingly) informational.
• Clear logic for which keywords go in which campaigns/ad groups.

Keyword research is the operator's craft. Tools provide data; judgment turns data into structure. Operators who keep refining keywords quarter after quarter compound their accounts. Operators who set up once and never revisit slowly bleed performance to operators who don't.