Google Ads says conversions. Sales says junk. Which keywords cause it?
When your Google Ads conversion is a form fill, every submission counts the same, so Google can't tell you which leads sales rejected. To find the keywords behind bad leads, match each lead in your CRM to the keyword or search term it came from, then compare spend with qualified leads per keyword. Keywords that spent more than twice your cost per qualified lead over 90 days, with no qualified lead, are the first to pause.
Why Google Ads can't see lead quality on its own
A conversion in Google Ads is usually a form submission. A student downloading a template, a job seeker and a real buyer all count the same. Sales then marks most of the first two as unqualified in the CRM, but that verdict never travels back to the keyword report. So the keywords with the most "conversions" can be the ones bringing the most rejected leads.
Syncing lead status back to Google (offline conversions, or the HubSpot and Salesforce integrations) helps Google's bidding learn over time. It doesn't give you a list of keywords and search terms with what they cost and what sales did with each lead.
Step 1: Get the keyword onto every paid lead
You need one thing in your CRM: the keyword or search term on each lead from Google Ads.
- HubSpot: with Google Ads connected, "Original Traffic Source Drill-Down 2" holds the search term for paid search contacts. Open one paid contact to check it's filled.
- Other CRMs and forms: add
utm_term={keyword}to your Google Ads tracking template and pass it into a hidden form field, so new leads carry it. - Check: at least half of your paid leads should have a keyword. Below that, work at campaign level until the tracking is fixed.
Step 2: Export three files for the same 90 days
- Google Ads keyword report: keyword, match type, campaign, cost, clicks, conversions.
- Google Ads search terms report: search term, the keyword it matched, cost.
- Your paid leads from the CRM: created date, keyword or search term, lead status, and the rejection reason if you record one. Leave out names and emails.
Our export guide has the exact menus for Google Ads, HubSpot, Pipedrive and Salesforce.
Step 3: Count qualified leads per keyword
For each keyword, add up spend, leads, and leads sales marked qualified (for example "Qualified", SQL, or a won deal). Then work out your average cost per qualified lead: total spend divided by total qualified leads.
Step 4: Decide keyword by keyword
- Pause
- Spent more than twice your average cost per qualified lead, with no qualified lead.
- Negative
- Words whose searches spent more than three times your average cost per qualified lead and never brought a qualified lead, such as "free", "template" or "jobs". Add them as phrase-match negatives.
- Bid down
- One qualified lead at more than three times your average.
- Add budget
- Your cheapest keywords per qualified lead.
- Watch
- Keywords that haven't spent enough to judge. Don't judge leads under 14 days old that sales hasn't worked yet.
Example (made-up company)
Sample data · made-up company
In our sample account, Northwind HR (made-up data), the keyword "free hr software" spent $3,100 in 91 days and brought 35 leads. Google Ads counted 35 conversions. Sales qualified none; the main reason given was "no budget". With an average of $546 per qualified lead, it passed the pause line ($1,092) almost three times over.
Common mistakes
- Judging on conversions in Google Ads instead of the status in the CRM.
- Pausing a keyword when one search word is the problem; block the word instead.
- Using different date ranges for the three files.
- Judging leads that are only a few days old.
Or let Pipeyield do steps 3 and 4
Pipeyield does steps 3 and 4 for you from the same three exports, with the dollar amount on each change. Our free check emails you the headline: the monthly spend on keywords and searches that brought no qualified leads, and the biggest one.
No logins · Never changes your ad account · Free-check files deleted within 7 days at most