Evaluate exact, phrase and broad match through actual search intent. Connect matching decisions to service boundaries, measurement and controlled testing.
Choosing a match type involves more than adding a symbol to a keyword. It affects the needs associated with an advertisement, the variety of searches encountered and the work required to assess them. Replace rules such as “broad is always bad” or “exact gives complete control” with a review of your offer's boundaries and the quality of your measurement.
Separate the keyword from the customer's query
A keyword is the expression configured in the account; a query is what someone actually searched. Google explains that matching considers meaning rather than word order alone. Assessing a campaign solely through its keyword list therefore misses the relationship between observed searches and the service.
Exact match should not be interpreted as a character-for-character restriction either. Google's close-variant documentation describes matching to similar meanings. The practical implication is straightforward: even a narrower match choice does not remove the need to review search intent.
Write down the needs the offer accepts
Imagine a company providing enterprise software integration. A broad reference to “software” could encompass training, employment, packaged products and custom projects. In this hypothetical example, clarify which enquiries the sales team can serve before choosing the match structure.
The service page should communicate those boundaries too. Advertising settings cannot fully compensate for an unexplained offer. Establish what you provide, who it is for and what makes an enquiry suitable before treating keyword configuration as the main problem.
Assess the choice against operational readiness
The following framework does not prescribe one universal option. It identifies the preparation to inspect and the question a test should answer.
| Situation | Area to inspect | Decision question |
|---|---|---|
| Narrow specialist service | Boundaries of search intent | Do similar expressions mean the same need? |
| Exploring additional queries | Measurement and review capacity | Can we distinguish the new intentions? |
| Poor enquiry quality | Goal event and sales feedback | Which definition of success is being learned? |
| Different services in one group | Advertisement and page relationship | Can one message answer all of them? |
| A working existing structure | Controlled expansion test | What commercial contribution does added scope bring? |
Keep the test question clear
Changing match types while introducing a service, page and success definition makes interpretation difficult. Write the hypothesis first: are you seeking additional relevant demand or reducing unsuitable enquiries? Select the outcome measure accordingly.
Do not judge the test only through cost per click. Investigate whether cheaper visits progress to enquiries and useful discussions. Allow for the delay between enquiry and sales assessment. When evidence is limited, retain that uncertainty instead of presenting a handful of records as a decisive result.
Share query learning with content and sales teams
When a relevant need appears but the destination does not answer it, do more than add a keyword. The service explanation or question section may need development. Repeated out-of-scope requests may instead indicate that the advertisement's message is too general.
Leave short decision notes for accepted intent, excluded intent and needs requiring further research. The next account owner should inherit the reasoning as well as the settings. This prevents the same experiments from being repeated without context.