Pay Per Click Report

In 2021 a client's Google Ads campaign kept charging for clicks while the phone stopped ringing. He raised the daily budget, and a few clicks later customers started calling again. Analytics showed visitors arriving and leaving within a second, which said something was wrong, not what.

The server log told us more. It held the same GCLID, Google's click ID, on line after line, so we built a pay-per-click report around it: log to CSV to Google Sheets. This is that report rebuilt for 2026, with one correction. A repeated GCLID is one ad click showing up again, and Google does not charge for it twice. In one Fort Worth service business's log, one click ID came back 13 times, while one address brought 6 new ones in 9 minutes.

Pay per click means Google charges you each time someone clicks your ad. With auto-tagging on, each click arrives with a GCLID on the end of the URL, and that ID is the thread between your bill and your server. Your log shows which of the charged clicks reached a page, and how often each came back.

In 2021 we read the repeats as one visitor clicking the ad over and over. Google's own help describes a milder cause: a visitor shares the ad URL, GCLID included, and everyone who opens it arrives carrying the first visitor's click ID. Google's answer sits on a page now marked legacy, since Universal Analytics was retired:

So watch the other count: many different GCLIDs from one address in one day. Each is a separate ad click, so each ends up either in Google's invalid clicks column or on your bill. Our GCLID investigation guide and blocking click fraud cover what to do next.

Our live log sits in ~/access-logs and gzipped older months in ~/logs; zcat -f reads both, and awk ships on every shared host:

zcat -f ~/logs/example.com-ssl_log-*.gz ~/access-logs/example.com-ssl_log \
  | awk '$7 ~ /[?&]gclid=[A-Za-z0-9_-]/ { match($7, /gclid=[A-Za-z0-9_-]+/);
           print substr($4, 2) "," $1 "," substr($7, RSTART + 6, RLENGTH - 6) "," $9 }' \
  > gclid.csv

The filter reads only $7, the requested URL. A plain grep gclid also matches the Referer, which carries the click ID onto every page the visitor opens next. Then the two counts:

# one click ID seen more than once: reloads, bookmarks, shares
cut -d, -f3 gclid.csv | sort | uniq -c | sort -rn | head

# many click IDs from one address: separate ad clicks
cut -d, -f2,3 gclid.csv | sort -u | cut -d, -f1 | uniq -c | sort -rn | head

Method: on October 13, 2026 we ran this pipeline on two logs. A Fort Worth service business's log, September 30 to October 12, held 57,480 requests and 227 ad-click lines from 128 different clicks; a plain grep gclid counted 416, because 149 lines carried the ID only in the Referer and 40 had an empty value. The 13 visits came from 4 addresses over 10 days, the bookmark pattern; the 6 IDs are six ad clicks to check against the invalid clicks column. Our own log, 171,311 requests since August 31, held only our two test links.

The 2021 version pulled the CSV from a public folder with IMPORTDATA, adding a timestamp to the URL to force a fresh copy. Google refreshes IMPORTDATA once an hour, so that trick still has a job. Google's servers fetch the file, though, so the URL has to open without a login, and your visitors' IP addresses become public.

Take the second row: copy the CSV down with scp, then use File > Import with "Append rows to current sheet", and each week stacks under the one before it.

A pay-per-click strategy comes down to three settings: the daily budget, the keywords you bid on, and the conversions you count. Each week, put Google's click count beside the GCLID lines in your sheet; if Google shows clicks your log lacks, that gap is your first finding. For the full weekly, monthly and quarterly checklist, see our pay-per-click account audit.

On a hosted builder with no log access, the pipeline has nothing to read. Ask your host or ad agency to run the two counts weekly and send the CSV with the address column trimmed, not the raw log. For an owner, the short version is this: ask for the weekly number of ad click IDs that reached your site, and compare it with the clicks on your Google Ads bill.

• Google Ads Help: Google Click Identifier (GCLID)
• Google Analytics Help: Common questions about Google Ads Clicks and Analytics Sessions (legacy)
• Google Docs Editors Help: Set a spreadsheet's location and calculation settings
• Google Docs Editors Help: Import data sets and spreadsheets

Google maintains the Help Centers that define every feature named here.

Sources: Google Click Identifier (GCLID): Definition, [UA] Common questions about Google Ads Clicks and Analytics Sessions, About invalid traffic, IMPORTDATA, Set a spreadsheet's location and calculation settings, Import data sets and spreadsheets. Method for the log report: pipeline run by PCDrama on its own server log on October 13, 2026.