How to do keyword research for Google Ads

To do keyword research for Google Ads, start from buyer-intent seed terms, expand each into every real phrasing, then score the results for commercial intent and volume. Keep the keywords a ready-to-buy customer would type, group them into tight ad groups of 5 to 20 related terms, and flag the off-intent phrases as negative keywords before you launch. The goal is not the biggest list. It is a structured list where every keyword earns its click.

Paid search punishes sloppy keyword research faster than SEO ever does, because you pay for the mistake on every impression. Below is the process that keeps a Google Ads campaign from spending its budget on people who were never going to buy.

Start with buyer-intent seeds, not category names

The single most expensive mistake in Google Ads keyword research is starting too broad. A term like "software" or "accounting" pulls in students, researchers, job seekers and browsers, and you pay for all of them. Your seeds should sound like a person about to spend money: "invoice OCR software", "convert bank statement to QuickBooks", "PPC agency for SaaS". Write down five to ten of these. They set the intent for every keyword the expansion turns up, so a tight seed list is worth more than a clever tool.

How do I find keywords for Google Ads?

Find Google Ads keywords by combining three sources: seed expansion for the obvious terms, competitor keywords for proven demand, and your own search terms report for the exact phrases real customers already use. Expansion adds the modifiers buyers type (best, pricing, near me, for small business) and the long-tail specifics that cost less per click. Competitor research shows what rivals already bid on. The search terms report, once the campaign is live, is the richest source of all because it is real queries, not estimates.

Google Keyword Planner is the free baseline for volume and forecast data, and most research still starts there. Its limit is that it hands you a flat list with no grouping and no intent read, so the planning work happens after the export. That gap is exactly what a PPC keyword research workflow is built to close.

Score every keyword for intent before volume

In SEO you can chase a high-volume informational term and monetize it slowly. In Google Ads that same term charges you for every click whether the searcher buys or not, so intent has to outrank volume. Sort your expanded list into three buckets: transactional ("buy", "pricing", "get a quote", "software"), commercial investigation ("best", "vs", "alternative", "reviews"), and informational ("what is", "how to", "examples"). Transactional and commercial terms belong in your campaigns. Most informational terms belong in your content plan or on your negative list, not in a paid ad group.

A lower-volume keyword with clear buying intent almost always beats a big head term full of researchers. Ten clicks from people ready to buy are worth more than a hundred from people three years out.

How many keywords should a Google Ads campaign have?

There is no magic total, but the structure matters more than the count. Build tight ad groups of roughly 5 to 20 closely related keywords each, then run several themed ad groups inside each campaign. A single ad group stuffed with 200 loosely related terms drags down Quality Score, because no single ad can be relevant to all of them, and low relevance means a higher cost per click. Fewer keywords per group, more groups, is the pattern that wins.

Think of each ad group as a keyword cluster with one message. When the keywords, the ad copy and the landing page all point at the same intent, Google rewards you with a better Quality Score and cheaper clicks. That clustering step is the heart of the work, and it is the part people most often skip because doing it by hand in a spreadsheet is tedious. A keyword clustering tool groups the terms that share an intent automatically, so the ad-group map is built for you instead of dragged together row by row.

Build the negative keyword list before you launch

Half of good PPC keyword research is deciding what not to bid on. As you expand your seeds you will surface plenty of terms that look related but signal the wrong intent: "free", "jobs", "salary", "cheap", "DIY", "tutorial", competitor brands you cannot serve. Every one of those goes on your negative keyword list. Adding negatives before launch, rather than waiting to discover them in the search terms report a month and several hundred wasted dollars later, is the fastest way to protect a new campaign's budget.

Google Ads supports negative broad, phrase and exact match, so you can block a whole theme or a single precise query. Start broad with the obvious money-wasters, then tighten as real search data comes in.

Set match types deliberately

Match types decide how loosely Google interprets your keywords. Broad match reaches the most queries and needs the most negatives to stay efficient. Phrase match keeps your term's meaning while allowing extra words. Exact match is the tightest, best for your proven money keywords once you know they convert. A common approach is to start newer campaigns on phrase and exact for control, then test broad match on your best performers with a strong negative list in place. This is the one part of the process no research tool handles for you: match types and bids live inside Google Ads, so plan them there.

What is the difference between SEO and PPC keyword research?

Both find the terms your audience searches, but PPC keyword research weights commercial intent and cost far more heavily because every click is paid for. An informational keyword can be a fine SEO target and a money pit in a paid campaign. PPC research also cares about match types and negatives, which SEO ignores entirely, while SEO cares about keyword difficulty and content depth. If you already run organic, do not just copy the SEO list into Ads. Re-score it for intent, keep the buyers, and send the rest to your SEO keyword research plan instead.

Turn the research into a launch-ready structure

The output of good Google Ads keyword research is not a spreadsheet of a thousand rows. It is a clean structure: a handful of campaigns, each holding tight ad groups, each ad group a cluster of buyer-intent keywords with a matching message and a negative list already attached. Once that skeleton exists, the campaign almost builds itself, and the next step is writing an ad for each group, which is far easier when you can generate ad copy and on-brand creative from your product page instead of starting from a blank box.

If you want the research and structuring part done fast, Keywordpilot takes one seed, expands and scores the terms for intent and volume, and clusters them into the ad-group-ready groups described above. It does not manage bids or match types, so it is not a replacement for Google Ads itself. It is the research layer that gets you to a launch-ready keyword map without the spreadsheet afternoon. Watch it work in the free demo with no account, or read the full workflow on the PPC keyword research page.