Google Ads Keyword Research: Tools, Intent and Strategy

by Francis Rozange | Apr 4, 2026 | Google Ads

Category: Google Ads | Reading time: 25 minutes | Last updated: April 2026

Keyword research is the foundation of every profitable Google Ads campaign. Without it, you are flying blind: paying for searches nobody makes, ignoring high-intent queries that convert, and watching competitors capture customers that should have been yours. The platform has changed substantially since 2023, with broad match driven by AI, Performance Max consuming a growing share of new-account spend, and search intent recognition (what Google now calls “query fan-out”) interpreting queries in ways that traditional exact-match thinking does not capture.

This guide walks through the complete 2026 keyword research process: how to use Google Keyword Planner correctly, how the four classic intent categories actually break down across the auction, when to choose Ahrefs over Semrush over SE Ranking, how to build a negative keyword strategy that survives broad match, and how to refresh keywords on a monthly cadence without burning budget. Sources are linked inline: Google’s official documentation, WordStream’s published 2024-2026 benchmarks, Search Engine Land’s coverage of platform changes, Tinuiti’s agency case studies, and the keyword research methodologies published by Ahrefs, Semrush, and Search Engine Journal.

Why keyword research is still the foundation, even in the AI era

A common claim in 2026 is that keyword research is dead, that Performance Max and Smart Bidding interpret intent automatically, and that the advertiser’s job is reduced to feeding the machine assets. The claim is half true and dangerously misleading. The AI does interpret intent, but it interprets the intent of the keywords and audience signals you give it. If those inputs are wrong, the algorithm scales the error rather than correcting it.

Consider two keywords with similar surface form: “project management software” (broad, exploratory, the searcher is comparing options without a shortlist) and “Monday.com vs Asana alternative” (commercial intent, the searcher has narrowed to two vendors and is actively evaluating a third). Both might cost you $4 to $8 per click in the auction, but the second converts at three to five times the rate of the first because it captures genuine buying intent at the bottom of the funnel. WordStream’s 2024 benchmark across B2B SaaS shows the gap is structural, not anecdotal: comparison-stage keywords average 6.1% conversion versus 1.4% for top-of-funnel exploration.

Keyword research is the discipline that surfaces those distinctions before you spend money on the wrong half. Performance Max and Smart Bidding need historical conversion data to optimize against; that data only exists if you fed the algorithm intent-aligned keywords from the start. Skipping research means letting the AI optimize against noise, which produces the canonical “Performance Max wasted my budget for a month” outcome documented across Tinuiti’s 2024 case studies.

Search intent: the four categories that separate waste from profit

Not all searches are equal. Internet Live Stats and Ahrefs’ 2024 SERP analysis classify the search volume distribution roughly as 53% informational, 32% navigational, 14% commercial, and around 1% transactional. Your Google Ads money lives almost entirely in the commercial and transactional buckets, which together account for less than one search in six but produce the bulk of advertiser revenue across every vertical Google reports.

Informational intent: usually not your customer

Informational searches express a desire to learn, research, or troubleshoot. The searcher is not ready to buy. Examples include “how to fix a leaky kitchen faucet”, “what is a CRM”, “best practices for email marketing”, and “how do I choose project management software”.

If you sell plumbing services and bid on “how to fix a leaky faucet”, you pay for clicks from people trying to save $100 by doing it themselves. If you sell CRM software and target “what is a CRM”, you reach researchers in week one of a twelve-week buying cycle, most of whom will not remember your brand by the time they actually shortlist vendors. These keywords belong in your blog, your YouTube channel, or your organic SEO strategy, not your paid Search campaigns. The exception is when the informational query reliably leads to a same-session conversion (a how-to that ends in a service call), and that exception needs to be proven with data, not assumed.

Navigational intent: low ROI unless you own the brand

Navigational searches target a specific company, product, or login page. The user knows where they want to go. Examples: “HubSpot pricing”, “Slack login”, “Shopify app store”, “best running shoes Nike”.

For your own brand, navigational intent is critical. Even if you rank organically for your company name, you should bid defensively on it to prevent competitors from intercepting your traffic with a comparison ad. Search Engine Land’s 2023 brand defense coverage documents the typical pattern: a competitor bids on your brand name, takes the top ad slot, and converts 5 to 15% of your warm traffic at a CPC well below their non-branded rate. Your own branded campaign costs cents per click and protects double-digit percentages of pipeline.

For competitor brands, the math is different. Bidding on “HubSpot pricing” when you sell Pipedrive is expensive (the auction has HubSpot’s own ad in it, plus three competitors, plus you), and the searcher already has a preference. The play works only when your differentiation is concrete enough to flip a meaningful share of those clicks, and most accounts overestimate that share.

Commercial intent: the sweet spot for Google Ads

Commercial intent shows the searcher is considering a purchase but evaluating options. They use modifiers like “best”, “top”, “vs”, “review”, “comparison”, “alternative”, “for [vertical]”. They want to narrow choices before buying. Examples: “best SEO tools for small business”, “HubSpot vs Pipedrive”, “top project management software 2026”, “enterprise CRM software comparison”, “affordable video editing software for Mac”.

These searches are gold for Google Ads. The searcher has budget, has timeline, and has not yet committed to a vendor. Your job is to show why you win that comparison. Commercial keywords typically convert at two to three times the rate of informational searches, and although they make up only 14% of total search volume, they generate a disproportionate share of paid-search conversions across every vertical Google reports. The competition is correspondingly stiff: comparison and “best” keywords are among the most contested in any category.

Transactional intent: immediate purchase or action

Transactional searches show the highest purchase intent. The user is ready to buy, subscribe, sign up, or call now. Examples: “buy running shoes size 12 online”, “download Slack”, “best WordPress hosting to buy today”, “emergency plumber near me open now”, “HubSpot pricing buy now”.

These keywords convert at the highest rates but face the stiffest competition because every vendor in the category wants them. They are worth high CPCs ($8 to $30 or more per click in competitive verticals) because the conversion probability is five to ten times higher than broad informational searches. The discipline is matching transactional ad copy and landing pages to transactional intent: a “buy now” searcher hitting a homepage with marketing copy bounces, while the same searcher hitting a product page with price, stock, and a one-click checkout converts.

Google Keyword Planner: the free foundation, with caveats

Google Keyword Planner is free to anyone with a Google Ads account and lives under Tools and settings > Planning > Keyword Planner. It is the most direct source of Google’s own search data, but the tool has changed substantially over the past three years and the most common workflow advice from 2020 is now wrong.

The two workflows: Discover versus Volume

Discover new keywords takes a seed keyword, a list of seeds, or a website URL and returns a related-keyword list with estimated monthly search volumes, competition levels, and CPC ranges. The URL workflow is underused: pasting a competitor URL into Discover returns the keywords Google associates with that competitor’s content, which is a free competitor research feed.

Get search volume and forecasts takes a list of keywords you already have and returns volume, competition, and CPC for each. This is the workflow for validating a keyword list compiled from other tools, brainstorming sessions, or competitor research.

Reading the data, with the data quality caveat

Keyword Planner returns three primary metrics per keyword:

Average monthly searches: an estimate of searches per month. Accounts with no active spend see ranges (“10K-100K”). Accounts with active campaigns see exact numbers, though Google rounds aggressively at the long-tail.
Competition: Low, Medium, or High, based on how many advertisers are bidding on the keyword. High competition usually means higher CPC and high-intent traffic.
Top of page bid (range): the estimated CPC range to appear in the top auction positions. A range of “$2-$5” means advertisers typically pay $2 to $5 per click for that keyword.

The caveat: Ahrefs published a 2023 study comparing Keyword Planner volumes against their own clickstream data and found that 91% of Keyword Planner volumes were over-reported, with 54% off by more than 10x in either direction. Treat Keyword Planner volumes as directional, not absolute. The relative ranking between two keywords is usually correct (keyword A has more search volume than keyword B), but the absolute numbers should not drive budgeting decisions on their own. Cross-reference with Ahrefs, Semrush, or SE Ranking before committing significant spend.

A worked example: running shoes

Searching “running shoes” in Keyword Planner returns:

Monthly volume: 500K-1M (very broad, very competitive)
Competition: High
Top of page bid: $3-$8 per click
Note: thousands of advertisers competing; conversion rate typically low because the keyword captures everything from “running shoes for kids” to “running shoes vs walking shoes” to “best free running shoes app”

Searching “best running shoes for flat feet size 12” returns:

Monthly volume: 1K-10K (much lower traffic)
Competition: Low to Medium
Top of page bid: $0.80-$2.50 per click
Note: far fewer advertisers; higher conversion potential because intent is specific

The second keyword typically converts at two to three times the rate of the first, costs roughly 70% less per click, and is far easier to rank or position for. This is why long-tail keywords dominate profitable Google Ads accounts.

Match types in 2026: what changed and what to actually do

Google has rebuilt match types twice since 2021. The rules in most legacy guides are wrong. The current state is documented in Google’s match types reference but the practical implications are best understood through what each match type now triggers in the auction.

Exact match

Exact match (square brackets, [keyword]) used to mean exactly the keyword. Since 2021 it means the keyword and “close variants” with the same search intent. [women’s running shoes Nike] now triggers for “Nike women’s running shoes”, “running shoes Nike for women”, “Nike women’s running sneakers”, and any reformulation Google interprets as the same intent.

Pros: highest Quality Score, lowest CPC, most control over which auctions you enter. Cons: lowest reach; misses variations Google does not classify as the same intent. Use for: high-intent commercial and transactional keywords where the auction economics demand precision.

Phrase match

Phrase match (quotes, “keyword”) used to require the keyword phrase in order. Since 2021 it includes “the meaning of the phrase”, which means it triggers for searches that contain the same intent regardless of word order. “women’s running shoes” triggers for “best women’s running shoes Nike”, “running shoes for women on sale”, and “Nike running shoes for women”.

Pros: a balance of reach and relevance. Cons: medium CPC, moderate control. Use for: most of your keywords. Phrase match in 2026 is the safe default for commercial-intent keywords where you want reach without the chaos of broad match.

Broad match

Broad match (no operators) was historically the worst-performing match type because it triggered on tangentially related searches. Since 2023, Google has rebuilt broad match around Smart Bidding: when paired with a Maximize Conversions or Target CPA bid strategy, broad match uses the same query-fan-out signals as Performance Max to find converting searches in places exact and phrase miss. Without Smart Bidding, broad match still produces the historical garbage results.

Search Engine Land’s 2024 coverage of the broad-match rebuild is explicit: broad match plus Smart Bidding now outperforms phrase match in roughly 60% of accounts where the conversion volume is sufficient (more than 50 conversions per month at the campaign level). Below that threshold, Smart Bidding has insufficient signal to control broad match, and the old “broad match is a budget bonfire” rule still applies.

Pros: maximum reach, AI-driven discovery of high-intent searches. Cons: requires Smart Bidding, requires conversion volume, requires aggressive negative keyword management. Use for: discovery campaigns once you have a baseline of conversion data, expansion of proven exact and phrase match keywords.

Practical match-type strategy

The pattern that works in most accounts is a three-tier approach:

Exact match on the highest-intent, highest-converting keywords (around 30-40% of budget)
Phrase match on the broader commercial keywords (around 40-50% of budget)
Broad match plus Smart Bidding on a discovery campaign with a tight negative keyword list (around 10-20% of budget)

The discovery campaign feeds new exact and phrase keywords back into the main campaigns as it identifies them, which is the canonical “broad match as discovery engine” pattern documented in Optmyzr’s 2024 broad-match playbook.

Long-tail versus head terms: why specificity beats volume

Long-tail keywords are longer, more specific phrases (typically four or more words) that capture narrow, high-intent searches. They represent less traffic per keyword but far better conversion rates and lower competition. The math is overwhelming.

The ratio that changes everything

Across audited accounts in 2025-2026, long-tail keywords typically drive 2.5 times higher conversion rates than head terms:

Head term example: “CRM software” – 1,000 clicks, 15 conversions (1.5% conversion rate, $5.20 average CPC, $347 cost per conversion)
Long-tail example: “CRM software for B2B sales teams under 20 people” – 120 clicks, 12 conversions (10% conversion rate, $2.10 average CPC, $21 cost per conversion)

In most Google Ads accounts, long-tail keywords represent less than 30% of total impressions but drive 50 to 60% of conversions. This inverts the conventional thinking: the lowest-volume keywords are often the highest-ROI keywords in the account.

Why long-tail keywords convert better

Less competition. “CRM software” attracts hundreds of advertisers. “CRM software for small B2B teams under 20 people” attracts five to ten. Less competition means lower CPC and easier ad positioning.

Clearer searcher intent. Someone searching a long-tail phrase has done preliminary research. They know roughly what they want. The specificity signals real need, not exploration.

Easier ad copy alignment. Specific keywords allow specific ad copy. You can highlight “supports 5 to 20 team members” or “$49/month flat pricing” directly in your headline because you know the searcher cares about those attributes. Generic “CRM software” ads cannot do this without alienating half the audience.

Higher Quality Score. Google rewards ads that align tightly with search intent. An ad for “CRM for small B2B teams” shown against the long-tail keyword “CRM for B2B teams under 20” earns higher Quality Scores than a generic CRM ad, which lowers your effective CPC. Optmyzr’s published data on Quality Score impact shows the difference between Quality Score 7 and Quality Score 5 is roughly 33% on cost per click for the same auction.

How to find long-tail keywords

Start in Google Keyword Planner with your seed keyword, then expand:

Look for variations with modifiers: “best”, “cheap”, “for women”, “brands”, “on sale”, “reviews”, “near me”
Focus on keyword ideas with 1K-10K monthly searches (the sweet spot for low competition and real volume)
Expand with Ahrefs, Semrush, or SE Ranking to find long-tail variations Google’s tool misses
Mine the Google Ads search terms report (after launching) to see real searches users make, which are almost always longer than your target keywords
Use AnswerThePublic for question-style long-tails (“how”, “what”, “why”, “when”) that surface informational variants

Third-party tools: when free is not enough

Google Keyword Planner is solid but limited. It does not show competitor keyword lists, historical trends, or keyword difficulty scores, and as noted above its volume estimates have known accuracy problems. Third-party tools fill these gaps. Pricing below is the published rate as of April 2026; verify on each vendor’s site before purchase.

Ahrefs ($129-$449/month)

Ahrefs is the strongest tool for keyword difficulty and SERP analysis. The Lite plan ($129/month) includes Keywords Explorer with search volume, keyword difficulty (0-100 scale), and CPC estimates, plus Site Explorer for competitor analysis. Higher tiers add team seats, expanded data limits, and advanced features.

Strengths: best-in-class keyword difficulty scoring, the largest live backlink index for SEO-PPC alignment, accurate clickstream data that often outperforms Keyword Planner volumes. The “Parent topic” feature surfaces the head term that captures the most volume across long-tail variants, which is invaluable for ad group structure.

Weakness: fewer PPC-specific features than Semrush. Ahrefs is designed for SEO first, with PPC as a secondary use case.

Semrush ($139.95-$499.95/month)

Semrush is the most comprehensive all-in-one platform. The Pro plan ($139.95/month) includes keyword research, competitor analysis (organic and paid), the PPC Keyword Tool specifically for Google Ads campaign setup, position tracking, and historical trend data.

Strengths: deepest competitive intelligence on paid search, including estimated competitor ad spend, ad copy history, and keyword overlap analysis. The Keyword Gap tool shows keywords competitors rank for that you do not, which is the canonical competitor research workflow. Native integration with Google Ads for one-click campaign import.

Weakness: highest entry price, can feel overwhelming for beginners, data updates lag the live SERP by a week or more in some categories.

SE Ranking ($65-$259/month)

SE Ranking is the value play in the comprehensive-tool category. The Essential plan ($65/month) covers keyword research, competitor analysis, position tracking, and on-page SEO at roughly half the price of Ahrefs or Semrush. The data set is smaller but covers most use cases for accounts spending under $20,000/month on ads.

Strengths: best price-to-feature ratio in the category, clean interface, white-label reporting for agencies. Volume data is sourced from Google Keyword Planner directly, which means the same accuracy caveats apply but the relative ranking is consistent with Google’s own.

Weakness: smaller backlink index than Ahrefs, smaller paid-search competitor database than Semrush, less mature historical data.

Keyword Tool ($89-$199/month)

Keyword Tool specializes in long-tail keyword discovery using Google Autocomplete data. It returns thousands of long-tail variants per seed keyword, far more than Keyword Planner surfaces, and supports YouTube, Amazon, eBay, App Store, Bing, and Instagram in addition to Google.

Strengths: unmatched long-tail discovery volume, multi-platform coverage useful for cross-channel campaigns. Pro plan unlocks search volume and CPC data for the keywords it surfaces.

Weakness: narrower feature set than Ahrefs or Semrush; primarily a discovery tool, not a research platform.

AnswerThePublic ($9-$99/month or free with limits)

AnswerThePublic visualizes the questions, prepositions, and comparisons people search around a seed keyword. Entering “Google Ads” returns a wheel of “how does Google Ads work”, “what is Google Ads quality score”, “Google Ads vs Facebook ads”, “Google Ads for plumbers”, and hundreds of other variants.

Strengths: surfaces the actual language prospects use, which is invaluable for ad copy and landing page alignment. Free tier with two searches per day covers light use; paid plans unlock unlimited searches and CSV export.

Weakness: no search volume on the free tier, limited competitor analysis, narrower than the comprehensive tools.

Choosing the right tool for your budget

Solo operator or small business: Keyword Planner (free) plus AnswerThePublic free tier covers 80% of needs at zero cost
Small agency or growing e-commerce: SE Ranking ($65/month) or Keyword Tool ($89/month) for the next layer of depth
SaaS or B2B with significant paid spend: Semrush ($139.95/month) for competitive paid-search intelligence and SDR alignment
SEO-led or content-heavy operation: Ahrefs ($129/month) for the backlink index and keyword difficulty depth
Enterprise or large agency: Ahrefs plus Semrush in combination, plus AnswerThePublic for content angles. The combined cost is roughly $300/month and covers every dimension of keyword research without compromise.

Negative keyword strategy: preventing wasted spend before day one

Negative keywords tell Google what not to show your ads for. They prevent low-intent, non-converting searches from draining your budget. With broad match playing a larger role in 2026, negative keyword strategy has become more important, not less, because broad match without negatives generates the highest variance in search-terms quality of any match type.

Why negative keywords matter at the research stage

Suppose you bid on “plumbing services” in broad match. Google’s algorithm will show your ad for “become a plumber” (job seeker), “plumber salary” (career researcher), “plumbing supply store” (contractor buying parts, not your customer), and “DIY plumbing” (the person trying to avoid hiring you). Every one of these clicks costs money with zero conversion probability. Negative keywords block them before they trigger.

The discipline is to add negative keywords at the research stage, not after wasting budget for two weeks. The patterns are predictable enough that 80% of the negatives you will eventually need can be added on day one from industry templates. The remaining 20% are discovered through search-terms reports during the first 30 days.

Common negative keywords by industry

Service businesses (plumber, dentist, lawyer, electrician):

“jobs”, “job”, “hiring”, “career”, “salary”, “wage” (job seekers, not customers)
“training”, “school”, “course”, “certification”, “license” (career researchers)
“DIY”, “tutorial”, “how to”, “guide” (people trying to avoid hiring you)
“free”, “cheap”, “discount” (unless you compete on price)

E-commerce (shoes, electronics, furniture, apparel):

“free” (unless you offer free shipping as your primary value prop)
“coupon”, “discount”, “promo code” (deal hunters, low margin)
“review”, “reviews” (research stage, not buying; bid on these only with comparison-stage landing pages)
“fake”, “counterfeit”, “knockoff”, “replica” (brand protection)
“used”, “second hand” (if you sell new only)
“wholesale” (if you sell retail only)

SaaS and software:

“free”, “open source” (if you do not compete in those segments)
“tutorial”, “how to”, “guide” (too early in buyer journey)
“download crack”, “torrent” (piracy searches)
“alternative free” (compound modifier worth blocking explicitly)
“jobs”, “careers” (job seekers)

Structural negative keywords: campaign isolation

If you run multiple campaigns targeting different segments, apply exact-match negatives to prevent crossover between them. The canonical case is a brand campaign and a generic campaign:

Brand campaign: keywords are [your company name] and variations, CPCs are low ($0.30-$1), conversion rates are high (15-30%)
Generic campaign: keywords are “project management software” and similar, CPCs are higher ($3-$12), conversion rates are lower (1-3%)

Add as exact-match negatives in the generic campaign: [your company name], [your brand variations], [common misspellings of your brand]. This prevents the generic campaign from competing in the brand auction, which would inflate brand CPC and waste generic budget. Without this isolation, your generic campaign cannibalizes your brand campaign and the data on both becomes uninterpretable.

Negative keyword lists at the account level

Build account-level negative keyword lists in Tools and settings > Shared library > Negative keyword lists. A baseline list of 50 to 200 universal negatives applied to every Search campaign saves hours of per-campaign exclusion work. Performance Max also supports account-level negative keyword lists since 2023, and they should be applied there as well.

Branded versus non-branded keywords: the architecture choice

Branded keywords (your company name, product names, and close variants) and non-branded keywords (generic category terms) behave so differently that they should almost always be separated into distinct campaigns with separate budgets, separate bid strategies, and separate ad copy.

Why separation matters

Branded keywords typically have:

CPCs of $0.30 to $2 (low competition, high Quality Score)
Conversion rates of 10 to 30% (warm traffic, high intent)
Cost per acquisition of $5 to $30 (the cheapest acquisitions in the account)

Non-branded keywords typically have:

CPCs of $2 to $15 (high competition, lower Quality Score)
Conversion rates of 1 to 4% (cold traffic, exploratory intent)
Cost per acquisition of $80 to $400 (10 to 20 times the branded cost)

Mixing the two in the same campaign distorts the optimization signal Smart Bidding receives. The algorithm sees the average performance of the campaign and bids accordingly, which produces under-bidding on branded (losing cheap conversions) and over-bidding on non-branded (paying for clicks that will not convert). Separating them lets each campaign optimize against its own performance reality.

Branded campaign structure

Keywords: exact-match variations of your company name, product names, and common misspellings. Match type: predominantly exact match, with phrase match for combinations like “your-company pricing” or “your-company login”.

Bid strategy: Target Impression Share at 95% top-of-page or higher. The goal is to dominate the auction on your own terms, not optimize CPA, because CPA is already low and the marginal click is high-value.

Ad copy: emphasize what makes your brand the right choice over alternatives. The searcher already knows you exist; the ad needs to confirm the decision and route to the right page (pricing, signup, demo).

Non-branded campaign structure

Keywords: category and intent-based, organized by ad group around tight themes. Match type: a mix of exact, phrase, and (with sufficient conversion volume) broad match plus Smart Bidding.

Bid strategy: Maximize Conversions or Target CPA, with a target set to your actual current CPA, not aspirational. Smart Bidding needs a realistic target to optimize against.

Ad copy: emphasize differentiation and the specific value prop that wins comparison-stage decisions. Lead with what makes you different, not what you do (which the keyword already implies).

Vertical-specific considerations

B2B versus B2C

B2B keyword research lives in a longer cycle and a higher LTV regime. The typical B2B SaaS deal cycle is 30 to 180 days, conversions are split across multiple touchpoints (whitepaper download, demo request, trial, contract), and the LTV per customer ranges from $5,000 to $200,000. The keyword strategy follows:

Top of funnel: informational keywords with content-led landing pages, measured by lead capture (whitepaper, newsletter), not direct conversion
Middle of funnel: comparison keywords (“X vs Y”, “best X for [vertical]”) with comparison landing pages, measured by demo request
Bottom of funnel: transactional and product-specific keywords (“X pricing”, “buy X”, “X for [specific use case]”) with high-intent landing pages, measured by trial start or contract

B2C is shorter and more direct. The cycle is hours to days, the conversion is a purchase, and LTV is $50 to $500 in most categories. The keyword strategy collapses funnel stages: most B2C accounts run primarily commercial and transactional keywords, with informational reserved for content marketing rather than paid search.

E-commerce versus services

E-commerce keyword research benefits from product-feed integration. Shopping campaigns and Performance Max with a Merchant Center feed bid on queries based on the product catalog rather than an explicit keyword list, which means the keyword research focus shifts to:

Negative keywords (more important than positive keywords for feed-based campaigns)
Brand and category-level Search campaigns running in parallel with Shopping
Audience signals (Customer Match, similar audiences) feeding Performance Max

Services keyword research is keyword-list driven. Local services especially benefit from “near me” and city-modified keywords (“plumber Austin”, “emergency plumber near me”), with location targeting tightened to the actual service area. WordStream’s 2024 benchmark shows local services keywords with city modifiers convert at 2 to 3 times the rate of generic local keywords because the searcher has self-qualified geographically.

Common keyword research mistakes

Volume worship. The keyword with 50,000 monthly searches looks irresistible until the math is run. At a 0.5% conversion rate and $5 CPC, capturing 1% of the search volume costs $25,000 to generate 250 conversions, or $100 per conversion. The keyword with 500 monthly searches at a 5% conversion rate and $2 CPC captures 30% of the volume for $300 to generate 7.5 conversions, or $40 per conversion. The low-volume keyword is 2.5 times more efficient. Most accounts chase the high-volume keyword anyway because the absolute conversion count is bigger; this is a budgeting error.

Ignoring search intent in match-type selection. Bidding [buy running shoes] on exact match makes sense; bidding [running shoes] on exact match wastes the match-type’s precision because the keyword is too broad to benefit from exact-match restriction. Match type and keyword specificity should rise together.

Treating Keyword Planner volumes as ground truth. The Ahrefs accuracy study cited above is one of several showing Keyword Planner volumes are systematically inaccurate at the long-tail. Cross-reference with at least one other tool before making budgeting decisions on volume.

Skipping the search terms report. The keywords you bid on are not the keywords users actually search. The search terms report shows the actual queries that triggered your ads, and reviewing it weekly is the single highest-leverage keyword research activity in the account. Most underperforming accounts have not opened the search terms report in months.

One-time research. Keyword research treated as a launch activity rather than an ongoing discipline produces a static keyword list that decays. Search behavior shifts seasonally, competitors enter and exit auctions, new modifier patterns emerge. Accounts that refresh keyword research monthly outperform accounts that do it once per quarter, which outperform accounts that do it once per launch.

Letting Performance Max handle keywords entirely. Performance Max is keyword-less from the advertiser’s perspective, but it still needs intent-aligned audience signals and a clean negative keyword list to function. Skipping keyword research because “Performance Max handles it” produces the canonical waste pattern documented across Tinuiti’s case studies: branded queries cannibalized at 5 to 10x the dedicated branded CPC, irrelevant search-terms triggering display impressions, and Smart Bidding optimizing against signals that do not align with business value.

Workflow: monthly keyword refresh

The accounts that compound performance over years run a monthly keyword refresh. The cadence is fixed, the steps are repeatable, and the time investment is roughly four to six hours per account per month. Below the cadence, keyword strategy decays; above it, the marginal time produces diminishing returns.

Week one: search terms review

Pull the search terms report for the past 30 days at the campaign level. Sort by impressions descending. For each search term:

If it converted at the campaign target rate or better, add it as a new keyword (exact or phrase match)
If it generated clicks but no conversions, evaluate intent: if the intent was wrong, add as a negative keyword; if the intent was right but the landing page or ad copy missed, fix those
If it triggered impressions but few clicks, the ad copy is misaligned with the search term; consider adjusting or restricting match type

Week two: competitor scan

Use Semrush, Ahrefs, or SpyFu to pull the top 50 keywords your top three competitors are bidding on. Cross-reference against your own keyword list. For each competitor keyword you do not bid on:

Validate volume and CPC in Keyword Planner
Search the keyword in incognito to see the live SERP and competitor ads
If the keyword aligns with your value prop and the auction economics work, add it to a test ad group

Week three: keyword expansion

For your top 10 converting keywords, generate long-tail variants in Keyword Tool, AnswerThePublic, or the Keyword Planner Discover workflow. Add the variants that match your audience and intent. Remove the variants that do not.

Week four: negative keyword pruning and consolidation

Review the account-level negative keyword list. Remove negatives that have been in place for over six months and may be blocking now-relevant queries (search behavior shifts). Add negatives surfaced from the week-one search terms review. Consolidate ad groups where keyword themes have drifted; split ad groups where conversion data shows two distinct intents have been mixed.

Conclusion: keyword research as ongoing discipline

Keyword research is not a one-time task before launch. The best Google Ads accounts treat it as ongoing discipline: weekly reviews of search terms reports, monthly competitor scans, quarterly refreshes of negative keyword lists, annual rebuilds of the keyword architecture as the business and the platform evolve.

The keyword that converts today might not convert in six months because of seasonal shifts, new competitors, or platform changes that reweight the auction. Conversely, a keyword you dismissed as “too low volume” might become your highest-ROI keyword after the market shifts. The discipline is to test continuously, follow the data rather than intuition, and treat the keyword list as a living asset rather than a launch artifact.

The tools and frameworks in this guide are the foundation. The accounts that win are the ones that apply them consistently over months and years, not the ones that read the guide once and never return.