How to Do PPC Keyword Research That Actually Builds a Profitable Google Ads Campaign
- Mar 3
- 12 min read
PPC keyword research is the process of finding the search terms your potential customers type into Google before they buy, and deciding which of those terms are worth paying to show up for. Get this right and your entire campaign runs on a foundation that produces actual customers. Get it wrong and you're paying Google for traffic that was never going to convert in the first place.
This is the full process we use at CRP Marketing when building Google Ads campaigns for small businesses and ecommerce brands. It covers how to find the right keywords, how to evaluate them, how to use Google's AI tools effectively in 2026, and, critically, how to keep your keyword strategy sharp after launch.
What Is PPC Keyword Research?
PPC keyword research is the work of finding, evaluating, and selecting the search terms you'll bid on in a pay-per-click campaign. Unlike SEO keyword research, which weighs long-term ranking difficulty, PPC keyword research has to answer one immediate question: will paying for this click produce a return?
That means evaluating three things for every keyword you consider:
Commercial intent. Is the person searching this term ready to act, or still in research mode?
Cost-per-click. What does it actually cost to win this click at auction?
Expected conversion rate. Based on similar campaigns or past data, how often does a click on this term turn into a sale or a lead?
The goal isn't to find keywords with the most search volume. It's to find keywords where the math works: where the cost per click, divided by your conversion rate, produces a customer acquisition cost your business can actually afford.
A landscaping company spending $8 per click with a 12% landing page conversion rate is paying about $67 per lead. If their average job is worth $450, that's a strong campaign. The same company spending $18 per click on a different keyword with a 4% conversion rate pays $450 per lead. The math falls apart entirely. Same industry, same business, wildly different outcomes based entirely on keyword selection.
Why Keyword Research Shapes 80% of Your Outcome
Most advertisers focus their energy on ad creative and ignore the keyword foundation beneath it. That's backwards.
If you're bidding on the wrong keywords, no amount of polished ad copy will rescue the campaign. You could have the most compelling offer in your industry, but if the ad is showing up for searches where the person has no intention of buying, you're just paying for impressions that go nowhere.
If you're bidding on the right keywords, even straightforward creative can produce strong results. We've watched campaigns built around tight, high-intent keyword lists consistently outperform campaigns built around clever copy but loose targeting.
Keyword research is where the majority of your campaign outcome gets decided. Everything else (bidding strategy, ad copy, landing pages) is optimization work on top of that foundation.
Step 1: Build Your Seed List
The first step doesn't require any tools. It requires you to think like your customer.
Make a list of every product, service, and outcome your business delivers, written the way your customers describe them, not the way you describe them internally. If you run a home services company, your internal language might be "HVAC maintenance contracts." Your customer types "AC tune-up near me."
Walk through each of your service lines and write down the phrases a buyer would use to find you. For a local plumbing company:
Emergency plumber near me
Water heater replacement cost
Burst pipe repair
Drain cleaning service [city name]
Plumber open on weekends
Three things to add before you move on:
"Near me" and city variants. Local-intent searches convert at significantly higher rates than generic searches (often 2-3x). Every local business needs these.
Problem phrases, not just service names. "Toilet won't stop running" has buying intent even though it sounds informational. The person needs a plumber.
Competitor brand names. You can bid on competitors' names as keywords. It works, especially if your offer is competitive on price or speed.
Aim for 30-60 seed keywords before moving to expansion tools. Depth here prevents gaps downstream.
Step 2: Expand With Keyword Tools
Once you have a seed list, tools help you find the adjacent searches, long-tail variants, and seasonal patterns you didn't think of.
Free tools (start here):
Google Keyword Planner is built into every Google Ads account (no spend required to access it). Shows monthly search volume ranges, CPC estimates, and keyword expansion ideas. This is the single most important tool for PPC keyword research.
Google Trends shows seasonality and geographic interest over time. Useful for businesses with demand that shifts by season or region.
Google Search itself: type any keyword and scroll to "People also ask" and "Related searches." Free competitive intelligence on what real users want.
Paid tools (worth it at scale):
Semrush: full PPC and SEO data including competitors' paid keyword strategies and estimated ad spend. Starting around $140/month.
SpyFu: focused on uncovering what keywords your competitors have been bidding on, and for how long. Starting around $39/month.
Ahrefs: strong keyword volume data, better suited for SEO but useful for cross-referencing PPC lists.
For most small businesses in year one of advertising, Google Keyword Planner alone gets the job done. The paid tools earn their cost once you're spending $5,000/month or more in ad budget, or competing in niches where you need a detailed read on what competitors are doing.
Step 3: Filter Your List Ruthlessly
Raw keyword lists are just inputs. The actual work is cutting them down.
Run every candidate keyword through four filters before it makes the final list:
Filter 1: Commercial intent. Map each keyword to a stage in the buying journey:
"Best [service] near me" → High intent. Bid.
"How much does [service] cost" → Medium intent. Worth testing, but watch conversion rates.
"How to [DIY version of your service]" → Low or no intent. Skip it.
Filter 2: CPC vs. your unit economics. Look up the estimated CPC in Keyword Planner. Divide it by your expected conversion rate. That's your estimated cost per lead. If a keyword has a $20 CPC and your landing page converts at 5%, you're paying $400 per lead. Does your business model support that? Most don't.
Filter 3: Competition reality. High-competition keywords cost more to win. Sometimes paying the premium is worth it because those keywords close at higher rates. Sometimes it's smarter to find a less-competitive long-tail variant. This is a judgment call that gets easier with campaign data.
Filter 4: Volume floor. A keyword searched 10 times a month probably can't carry its own ad group. Either cluster it with similar low-volume keywords or skip it for now. Thin ad groups dilute Quality Score and generate almost no data to optimize against.
A well-filtered final keyword list is typically 30-50% smaller than the raw expanded list. That reduction is the work. Tight lists outperform sprawling ones almost every time.
Step 4: Match Types and the Modern Broad Match Approach
Here's where a lot of beginner guides give outdated advice, so pay attention to this section.
For years, the standard guidance was: start with exact match and phrase match, avoid broad match until you have data. In 2024 and 2025, Google's Smart Bidding dramatically changed this equation. In 2026, broad match paired with a Smart Bidding strategy is often the recommended approach (not the risky one).
Here's why it changed:
What Match Types Do
Exact match: your ad shows only for searches that match your keyword very closely. Precise, but limited volume and often more expensive CPCs because everyone is fighting for the same exact queries.
Phrase match: your ad shows for searches that include your keyword phrase, with words allowed before or after. Middle ground.
Broad match: your ad can show for searches related to your keyword, including synonyms and conceptually similar queries. Historically risky, now much smarter.
Why Broad Match + Smart Bidding Works Now
Smart Bidding (Maximize Conversions, Target CPA, Target ROAS) uses your conversion data to evaluate whether a given search is worth bidding on before showing your ad. Broad match surfaces the traffic; Smart Bidding decides which of that traffic actually gets served your ad.
The practical effect: broad match with Smart Bidding often finds high-converting searches you never would have explicitly added to an exact match list, because you never thought of the exact phrasing a customer used.
The play for most campaigns today:
Start campaigns with Smart Bidding enabled (Maximize Conversions to build data, then shift to tCPA or tROAS once you have 30+ conversions per month)
Use broad match for your core keywords to capture the full range of relevant searches
Keep your negative keyword list tight (more on this in a moment)
Review the search terms report weekly for the first 60-90 days to catch anything broad match is surfacing that shouldn't be there
This isn't "set it and forget it." The broad match + Smart Bidding combo requires active management in the early weeks. But when it's working, it consistently finds conversion volume that narrow match strategies miss.
Step 5: Performance Max and Search Themes: Keyword Research for the AI Campaign Era
Performance Max (PMax) campaigns work differently from standard Search campaigns. Instead of bidding on keywords, you provide Google with asset groups, audience signals, and, importantly, Search Themes.
What Are Search Themes?
Search Themes are up to 25 phrases you give Google inside a PMax campaign to tell the AI what kinds of searches your business should show up for. They're not keywords in the traditional sense: you're not telling Google to bid exactly when someone searches that phrase. You're telling Google's AI "this is the territory of searches that matters to my business."
Think of Search Themes as the PMax equivalent of keyword research. You still need to think carefully about:
What problems your customers are searching for solutions to
What service/product language they use vs. what you use internally
What geographic and intent modifiers matter (near me, cost, best, reviews)
For lead generation campaigns: Search Themes should reflect the high-intent, decision-stage phrases your best leads would type. Don't waste themes on awareness-stage searches.
For ecommerce advertisers: Your product feed does a lot of the targeting work in PMax, but Search Themes help ensure the AI is anchoring on the right product categories and buyer language.
PMax doesn't replace Search campaigns. Most well-structured Google Ads accounts run both. But understanding Search Themes means your keyword research process now produces outputs for two campaign types, not one.
Step 6: AI Max for Search: What It Changes About Keywords
AI Max for Search is Google's newer search campaign type that applies more aggressive AI-driven matching alongside expanded creative options. It's built to run alongside Smart Bidding and takes the broad match logic further.
With AI Max enabled on a Search campaign, Google can:
Show your ads for queries beyond your listed keywords if it believes there's conversion intent
Dynamically generate headlines and descriptions matched to the specific search query
Optimize landing page selection automatically based on predicted conversion likelihood
What this means for keyword research: your keyword list becomes a directional input, not a rigid filter. The explicit keywords you add tell Google's AI where to start and what territory matters to you. The AI then explores intelligently from there.
This doesn't mean keyword research matters less. It means:
Your seed keywords need to accurately represent your real business. If your keyword inputs are off, the AI explores in the wrong direction.
Negative keywords become even more important (see the next section): they're now one of the primary ways you constrain what the AI can do.
The search terms report becomes your primary ongoing keyword research tool. Each week, you're reviewing what the AI surfaced and either validating it or blocking it.
AI Max is still relatively new, and how much you open it up depends on your campaign goals, budget, and comfort level with AI-driven matching. Our paid media services team uses a staged approach: start with standard Smart Bidding + broad match, add AI Max features as conversion data accumulates.
Step 7: Negative Keywords: Still the Most Underrated Work in PPC
Despite all the progress in AI matching, negative keywords remain one of the highest-impact activities in any Google Ads account. The accounts we audit most often don't have a strategy problem. They have a negative keyword problem.
Negative keywords are the search terms you explicitly tell Google not to show your ad for. Every dollar wasted on irrelevant clicks is a dollar that could have gone toward a query that converts.
Negative keyword categories to build from day one:
Research intent: "how to," "what is," "DIY," "tutorial," "guide," "examples"
Job seekers: "[service] jobs," "[industry] career," "salary," "hiring"
Price signals that don't match your offer: "free," "cheap" (if you're a premium provider), or "premium," "high-end" (if you're targeting budget buyers)
Irrelevant product or service variants: If you sell residential services, exclude commercial variants. If you offer one service, exclude adjacent services you don't provide.
Competitor name protection: Add your competitors' brand names as negatives if you don't want to pay for comparison shopping traffic.
Start every new campaign with 25-50 negatives already in place before it goes live. Add to the list weekly based on the actual search terms report.
The reason negative keywords matter even more in the AI Max era: when Google's AI has more freedom to match broadly, it will occasionally surface irrelevant searches. Your negative keyword list is the fence that keeps the AI working in the right territory. Without it, broad match and AI Max campaigns can bleed budget fast.
One concrete example: a home services client we took over was spending 22% of their ad budget on searches related to home improvement TV shows and competitor brand names. Adding a targeted negative keyword list recaptured that budget within two weeks, without changing bids, creative, or any other variable. That's the leverage negatives offer.
Step 8: First-Party Data and Customer Match as a Keyword Strategy Amplifier
Modern PPC keyword strategy doesn't exist in isolation from your customer data. First-party data (email lists, CRM data, past purchase records) can sharpen how Google's AI uses your keywords.
Customer Match lets you upload your customer or lead lists into Google Ads. Once uploaded, Google uses that data to:
Identify patterns in how your best customers search
Adjust bids when someone who matches your customer profile searches for your keywords
Help Smart Bidding understand what a "high-value user" looks like before they've ever clicked your ad
For ecommerce advertisers, this means your keyword strategy for existing customers can be different from your keyword strategy for new customer acquisition. You might bid more aggressively on broad match keywords when your Customer Match signals show a high-value existing customer is searching, because that click is worth more.
For lead generation campaigns, Customer Match helps Smart Bidding qualify traffic earlier in the funnel. Instead of bidding the same on every impression, the system can weight bids toward people who look like your best historical leads.
This isn't particularly complex. Setting up Customer Match takes a few hours if you have a clean email list. But it's the kind of foundational infrastructure that separates accounts producing customers from accounts just producing clicks.
Step 9: Turn Ongoing Search Term Reports Into Your Living Keyword Strategy
The biggest keyword research mistake we see consistently: advertisers do the upfront work, launch the campaign, and never go back to the search terms report.
Inside Google Ads, the Search Terms report shows every actual query that triggered your ads. This is the most valuable keyword research data available to you. It's generated from your own campaigns, for your specific audience, in your specific market.
Check the search terms report on a weekly cadence for the first 90 days of any new campaign. You will find:
Wasteful spend you didn't anticipate: queries that look related but never convert (add to negatives immediately)
New high-intent keywords to add explicitly: queries that converted well that you never would have thought to add manually
Match type insights: patterns in what the AI is matching to your keywords, helping you decide whether to tighten or expand
After 90 days, most established campaigns can move to monthly reviews. But new campaigns need weekly attention. The cost of not doing this (in wasted spend and missed opportunity) compounds quickly.
What a strong search terms review looks like:
Filter by spend: find the highest-cost queries first
For each high-spend query, ask: did this convert? At what rate? At what cost per conversion?
Non-converting high-spend queries → add to negatives
High-converting queries you're not explicitly bidding on → add as keywords in appropriate match types
Repeat weekly until the account stabilizes
This discipline is the difference between a campaign that improves over time and one that stays stuck at mediocre performance.
How to Structure Keyword Groups for Maximum Quality Score
Keyword organization directly affects campaign performance because it determines how relevant your ads are to each search, which Google measures through Quality Score and rewards with lower CPCs.
The principle: group keywords by shared search intent so each ad group can have ad copy and a landing page that specifically matches what the searcher wants.
Example for a dental practice:
Ad Group | Keywords | Landing Page |
Emergency dental | "emergency dentist near me," "dentist open now," "broken tooth" | Emergency page with same-day booking |
Cosmetic dental | "teeth whitening cost," "veneers near me," "smile makeover" | Cosmetic page with before/after photos |
New patients | "dentist accepting new patients," "family dentist [city]" | New patient welcome page |
Each ad group gets:
Its own ad copy directly referencing the search intent
Its own landing page matching the specific need
Its own subset of closely-related keywords
This structure produces higher Quality Scores, which means Google charges you less per click for the same ad position. It's not uncommon to see CPCs drop 20-35% after restructuring a poorly organized account into tighter ad groups, with no change to bids or budgets.
5 PPC Keyword Research Mistakes to Avoid
1. Broad match with no negative keyword list. This combination guarantees wasted spend. Broad match's power is inseparable from negative keyword discipline.
2. Choosing keywords based on volume instead of intent. A keyword searched 50,000 times a month is worthless if it attracts researchers, not buyers. Intent always beats volume.
3. Using SEO difficulty scores to evaluate PPC keywords. Organic ranking difficulty is irrelevant to paid search. What matters in PPC is CPC, expected conversion rate, and your unit economics.
4. Ignoring local intent. If you serve a specific area, "near me" variants and city-name modifiers often outperform generic service keywords by a wide margin. They signal geographic intent and tend to attract buyers with higher urgency.
5. Treating keyword research as a launch task. Your keyword list at launch is a starting hypothesis, not a finished strategy. The best-performing keyword strategies are the ones refined continuously based on actual search term data.
Want a real read on your paid media?
If you're running Google Ads and you're not sure your keyword strategy is actually sound, or you're starting from scratch and don't know where to begin, that's exactly the conversation we have on a strategy call.
We'll look at your account, your search terms report, and the keywords you're paying for. Then we'll tell you straight: here's what's working, here's what's wasting budget, here's what we'd fix first.