Keyword research and match types

Start away from the tools. Open your last ten sales call notes, support emails and won-deal summaries. Highlight the phrases customers used to describe the problem, not the phrases you use for the product. If you record calls, a tool such as Fireflies.ai makes the phrases searchable.

Put them in a Google Sheets file with one phrase per row. Aim for 20 to 30 rows. These are your seeds.

Paste the seeds into Keyword Planner inside Google Ads. It suggests related searches and shows an indicative volume range. Add a second source: the queries already bringing people to your site in Google Search Console. Add a third: the autocomplete suggestions Google shows as you type a seed.

Copy anything relevant into the sheet. Ignore the volume number as a ranking. Low volume with strong intent beats high volume with none.

Add a column and label each row with one of four intents:

• Problem: "late payment reminders".
• Solution: "invoice automation software".
• Comparison: "invoice software alternatives".
• Brand: your name and your product names.

Start with solution and comparison. Problem searches are often early research, and brand searches belong in their own campaign. If you are unsure about a keyword, ask: would someone typing this be ready to talk to a supplier this month?

Delete keywords that are too broad, for example "software", and ones where you do not serve the searcher. Keep 20 to 50 keywords for a first launch, grouped into clusters of 5 to 15 that share one intent. Each cluster becomes an ad group. How to build the account around them is in Campaign structure and account setup.

Match types decide how loosely Google reads your keyword:

• Exact match, written [invoice automation]: the query and close variants of it. Tightest control, least reach.
• Phrase match, written "invoice automation": queries that include the meaning of your phrase, with words before or after.
• Broad match, written invoice automation: anything Google judges related, which can be far from what you meant.

Start with exact and phrase for your best keywords. Add broad match only later, with strong negatives and smart bidding in place. The reasoning is in Match type and negatives are how you control the spend.

Negative keywords stop your ad showing on searches that cannot buy. Start with a list at account level, such as:

• free, template, example, sample, pdf.
• jobs, salary, career, course, training, certification.
• definition, meaning, what is, wikipedia, reddit.
• diy, open source, download.

Check each term against your business first. If you sell training, "course" is not a negative. Add the list as a shared list so every new campaign gets it automatically.

After the first week, open the search terms report. It shows the actual queries that triggered your ads. Move good ones into your keyword list as exact match. Move bad ones into the negative list. Do this twice in the first month, then weekly. The routine is in Search campaign optimisation.

Add a notes column: where each keyword came from and what you decided. In three months you will want to know why you bid on a phrase. You can also hand the sheet to an agency or a colleague without a briefing.

• Starting from a keyword tool instead of from customer language.
• Chasing high-volume terms that bring researchers, not buyers.
• Using broad match on day one with no negative list.
• Putting 80 keywords into one ad group.
• Forgetting that close variants mean exact match is not literal.
• Never reading the search terms report.

In the first two weeks, most of the queries in the search terms report make sense for your business. The share of clicks from irrelevant queries falls each week as you add negatives. Cost per qualified lead falls, and the sales team recognises the language leads use in their forms.

On Monday morning: block 90 minutes, pull ten call notes, and build the first 30 rows of the sheet.