Start with the search terms, not the dashboard
The search terms report shows what people actually typed before an ad click, while keywords show what you intended to target. Those are not always the same. I look for queries that are unrelated, too broad, outside your service area, or unlikely to become a profitable customer. I also separate useful variations from noise so promising language is not accidentally blocked along with poor-fit searches.
The goal is not to add negative keywords blindly. A good review considers intent, location, customer value, and whether a query belongs in a separate campaign or ad group. I organize findings into keep, exclude, investigate, and potentially promote categories. That gives you a repeatable maintenance process instead of a one-time list that becomes outdated.
- Review search terms by intent and business fit
- Check location and service-area relevance
- Identify recurring irrelevant themes for negative keywords
- Look for high-intent queries worth dedicated coverage
Find broad match leaks and campaign structure problems
Broad match can help discover demand, but it can also expand into searches that only loosely relate to your offer. I examine how broad match is being used, which signals and settings influence expansion, and whether the resulting traffic fits the campaign’s purpose. The important question is not whether broad match is good or bad; it is whether the account has enough control and review discipline for the current budget.
I also inspect campaign and ad group organization. Mixed services, locations, or customer intents can make ad relevance and reporting difficult to interpret. When everything shares one budget and one message, you may not know which service is consuming spend or which search deserves a better landing page. I flag structural issues that prevent meaningful comparisons without recommending unnecessary rebuilding.
- Compare match types with actual query quality
- Check whether campaigns separate services and locations sensibly
- Review negative keyword lists and conflicts
- Identify budget priorities hidden by mixed intent
Test whether conversion tracking can be trusted
A campaign cannot be evaluated properly if its conversion actions are incomplete, duplicated, or tracking the wrong behavior. I review the definitions behind calls, forms, bookings, purchases, and other actions to see whether they represent genuine business outcomes. I also look for softer actions being counted alongside primary leads in ways that can make performance appear stronger than it is.
Tracking problems often sit outside the ad interface. A form change, consent setting, thank-you page update, tag issue, or analytics configuration can interrupt measurement without stopping ad delivery. I document what appears to be tracked, what needs verification, and which findings require access to the website or related measurement tools. The result is a clearer boundary between an advertising problem and a data problem.
- Separate primary conversions from secondary actions
- Check for duplicate form or call counting
- Verify that tracked actions match real business goals
- Note website or analytics access needed for confirmation
Match ads and landing pages to the search
A relevant click can still waste money if the destination page is vague, slow, difficult to use, or mismatched with the ad promise. I compare the search intent, ad copy, call to action, and landing page content as one path. For example, a search for a specific service should not land on a generic homepage that makes the visitor hunt for the next step.
I look for practical friction: unclear service areas, buried contact details, weak proof, confusing forms, mobile layout problems, and calls to action that do not fit the visitor’s readiness. I do not treat every conversion issue as a request for a new website. Often the best first move is a focused page edit, clearer offer, or cleaner route from ad to contact.
- Compare the ad promise with the destination page
- Check mobile forms, calls, and primary actions
- Review service-area and offer clarity
- Prioritize changes that can be tested without a rebuild
Turn findings into a budget and testing plan
An audit should end with decisions, not a long export of observations. I rank issues by likely impact, confidence, effort, and risk. Immediate controls may include excluding clearly irrelevant searches or correcting a broken conversion action. Larger recommendations may include restructuring campaigns, creating a service-specific page, or testing a different offer. Each item should have an owner and a reason for doing it.
I also distinguish between fixing waste and learning something new. A negative keyword can prevent a known mismatch, while an ad or landing page test addresses an uncertain question. Keeping those categories separate helps you avoid changing too many variables at once. After implementation, review search terms, conversion quality, and spend by intent on a regular schedule so the account does not drift back toward waste.
- Prioritize urgent tracking and relevance issues first
- Separate cleanup tasks from experiments
- Assign a measurement question to each test
- Schedule recurring search-term and conversion reviews