How I Research Google Ads Keywords Before I Spend a Dollar

Most Google Ads keyword lists start in a tool. Someone types their service into a keyword planner, sorts by search volume, and picks the top twenty. I start somewhere else: the places where the customer already talks about the problem, in their own words.

In 2021 I ran a search campaign for United Spay Alliance, a US animal welfare nonprofit that needed volunteers. The campaign ran on a Google Ad Grants account, the free search advertising Google gives eligible nonprofits. The audience was pet lovers, split into two groups: aspiring vets who wanted hands-on experience, and animal activists. The goal was to get both groups onto the organization’s landing page and signed up for its volunteer network.

I’m Canadian, and this audience wasn’t. The differences in how Americans search for animal welfare topics were subtle, but phrasing is the whole game in search ads, so I didn’t guess at it.

Where I looked first

I did the research on the open web. Reddit communities like petlovers and petparents. Facebook groups. X forums where animal health workers talk to each other. Then the places people leave reviews: pet groomers, pet training books on Amazon, the comment sections of pet blogs.

Reviews and comments are useful because nobody in them is writing for a search engine. People describe what they wanted, what went wrong, and what they’d tell a friend, in the same words they’ll later type into Google.

Turning that into a keyword list

After a few days of reading, patterns showed up. The same phrases repeated across groups that had nothing to do with each other. That repeated language became the first draft of the keyword list.

Narrowing the list with a tool

Then I brought in SEMrush. It showed how often each phrase was searched in the region the campaign would run in, and how much competition sat behind it. I kept the phrases with high search volume and lower competition.

I also asked one question of every keyword: what stage is the person in when they type this? For this campaign the target action was signing up to volunteer, so I favored phrases from people ready to act. Phrases from people still learning about the topic got less weight. The same sorting applies to a local service business. Someone searching “furnace replacement cost” is ready to call, and someone searching how furnaces work is not, which is the same intent split I cover in choosing an ad channel.

Writing ads from the list

With the shortlist set, I built the ads in Google Ads with as many headline and body variations as the format allows. I wrote each variation from the keyword list, aimed at someone deciding whether to help a community organization. Then I let the variations compete. Different copy won for different phrases, and I would not have picked the winners by reading them.

After launch

Then the real work started. I pulled the data at three levels: the campaign overall, each individual ad, and each audience segment. Ads that looked fine in the campaign total turned out to be carrying or dragging specific segments, and I adjusted keywords and copy from there. Conversions on the campaign rose 30% year over year.

The same habit applies to any account. For ECM Heating & Cooling, the Google Ads turnaround came from the same kind of digging: narrowing the whole campaign changed the result. If you are still deciding which campaign type fits your business, start with choosing a Google Ads campaign type. If you want someone to build the keyword research and the account, that is what my Growth and Paid Search work covers, and I run it for businesses in Belleville, Ottawa, and Brockville as well as Kingston.